2ND INTERNATIONAL

Digital Human Modeling Symposium Presented by The University of Michigan Transportation Research Institute and the OPEN Design Lab at Penn State, in association with the International Ergonomics Association Technical Committee on Human Simulation and Virtual Environments

Paper Abstracts

Hosted by

11-13 June, 2013 Ann Arbor, Michigan

PAPER ABSTRACTS

2nd International Digital Human Modeling Symposium

presented by The University of Michigan Transportation Research Institute and The Pennsylvania State University in association with The International Ergonomics Association Technical Committee on Human Simulation and Virtual Environments Ann Arbor, Michigan, USA

11-13 June 2013

Full papers are available at http://www.dhm2013.org/proceedings/

Paper 3 Delphi Survey: Digital Ergonomics 2025 S. WISCHNIEWSKI Federal Institute for Occupational Safety and Health (BAuA), Dortmund, Germany Research on how to support product and production design processes by computer-aided tools of ergonomics is one of the scientific focal points within the interdisciplinary field of ergonomics. Digital human models are an important aspect within this research context. Assuring a widespread use of these and other tools gives great potential for increasing product usability and designing safe, healthy and competitive work systems: The identification and redesign processes for dangerous or unhealthy product design and work system parameters can be moved into early product and production process design phases. In order to analyze specific future technological and organizational trends, an expert- and web-based, three-round Delphi survey on "Digital Ergonomics" was carried out from November 2011 until May 2012. A panel of 60 German experts was asked to answer nine lead questions in a first round. 26 statements were then built resulting from a qualitative data analysis of the received 886 answers. Each statement was evaluated according to its impact as well as its most probable date of occurrence in the second round. In the last round, the results of the second run were presented to the experts according to the Delphi technique and they were asked to assess the statements once again. During the whole Delphi survey the response rate was about 55%. Based on the results a ranking was built and the trends were sorted into seven categories according to their importance and most probable date of occurrence. Furthermore, named advantages, disadvantages and requirements of small and medium sized enterprises on digital human modeling were assessed. The results of the Delphi survey give an expert-based roadmap on trends in digital human modeling until 2025, which are presented in detail in this paper.

Paper 4 Digital Human Modeling and Simulation for Ergonomics Workspace Design: two Brazilian cases D. BRAATZ*†, L.TONIN†, A. FONTES‡, N. MENEGON† † Department of Production Engineering (Federal University of São Carlos), São Carlos, Brazil ‡Department of Production Engineering – Campus Sorocaba (Federal University of São Carlos), Sorocaba, Brazil In Brazil the technology of Digital Human Modeling and Simulation (DHMS) on workspace design is still at the initial phase of diffusion. On this context this paper investigates a computational tool for Digital Human Modeling and Simulation contextualized by Ergonomics Analysis of Work (EAW) and the analysis of the likely future activities can assist in the design processes of workplaces. The aim of this study is to show how Digital Human Modeling and Simulation and contextualized by Ergonomics can help on the process of workplace design and which of the theoretical references used as support. Two case studies in which the DHMS was employed with the software Jack are examined. The first case presents the design of an attendance counter in a public postal service. The second shows the development of a workstation for the supply of surgical needles in a company manufacturing products of health and hygiene. From the results of case studies are explained the contributions and challenges of using this technology in design aiming to solve the issues of health and productivity. The results presented show that using this technology on a structured and conceptually based way, can be considered an auxiliary instrument on the search of consideration of the possible future activity (Future Activity Approach) and identifying possible constraints on projects of productive situations. On the other hand, such tool is proper from human factors, of a highly technicality matter, with emphasis on biomechanical factors, anthropometric and with big focus on quantitative analyses. The use of integrated DHMS with ergonomics intervention helped improve anticipation of likely future activities of new situations of work and helped the integration and communication of the actors involved in the social process.

Paper 5 Complementarities of DHM and ergonomics analysis of work into workplace design: the task of manual packing A. R. M. FONTES*†, A. C. PARISE‡, G. F. FILIPPI†, D. BRAATZ‡, N. L. MENEGON‡ †Department of Production Engineering – Campus Sorocaba (Federal University of São Carlos), Sorocaba, Brazil ‡Department of Production Engineering (Federal University of São Carlos), São Carlos, Brazil The practice of Ergonomics centered in analysis of activity allows the design to contemplate the real task to be performed at the future workstation, influencing the conception of an effective working situation compatible with the health of workers. Thereby, the modeling and simulation on workplace design is related with an important role in this process. The aim of this study is to show how Digital Human Models (DHM) and Ergonomic Analysis of Work (EAW) can be integrated in a design process of a new workplace. To reach this goal this paper examines the application of human simulation technology during an ergonomic intervention, which demanded design of a new workplace conception with new technical artifacts. The simulation was utilized to test some design variables and to help the communication among the actors. In all simulations and analysis were used a male manikin percentile 95, as a subject of extreme maximum limit, and a male manikin percentile 05, as the minimum limit. Thus, the aims were to project a workstation able to fit most part of the population under study. The environments were designed in AutoCAD 2010 (Autodesk) and then imported by Delmia V5R18 (Dassault Systemes) to simulated static tests (variables as: posture and movements, extreme situations of range and handling of big dimensioned materials and others) and dynamic tests (animations of postures, movements and displacements in the possible future work). The main results indicate that the use of digital human simulation to support workplace design is an important tool to anticipate the future work activity. But only the application of simulation techniques, by itself, unaccompanied by understanding the activity performed, provides a fragmented view that does not incorporate the variability present in the work situation, focusing only tasks (prescriptions). It was noticed in the case studied, that the integration between EAW and simulation gave workers the possibility of participation/interaction as directly responsible for planning work.

Paper 6 Ingress/egress motion strategies of the elderly people for the rear seat of minivans with sliding doors J.-M. LU*†, M. TADA‡, Y. ENDO‡, M. MOCHIMARU‡ † Gerontechnology Research Center, Yuan Ze University,128 Yuan-Tung Road, Chungli, Taoyuan 32073, Taiwan, R.O.C. ‡ Digital Human Research Center, National Institute of Advanced Industrial Science and Technology, 2-326, Aomi, Koto-ku, Tokyo 135-0064, Japan This study aims to investigate the elderly people’s motion strategies when entering and exiting the rear seat of minivans with sliding doors. In order to represent the varying geometries of minivans, a minivan mockup was constructed with four adjustable parameters at three levels each, including roof height, door width, floor height, and seat height. Based on the survey of seventeen minivans produced by seven manufacturers, only nine reasonable combinations were taken into consideration. Besides, the optical motion capturing system was employed to acquire the 3D motion data. Ten elderly male subjects (66.8 ± 3.8 years old) and ten young male subjects (31.5 ± 6.6 years old) participated in this study. Each subject enters and exits the minivan mock-up for five times under each condition, freely deciding whether to put the hands on the center pillar or the front seat to support the body weight. Three dimensional joint angles and positions of joint centers at fifteen joints are adopted to represent the kinematics of motion. Finally, all the motion data were clustered into few categories as different motion strategies. For ingress motion, seven strategies were identified. Most young people tend to move the inside leg onto the floor first, sit down, and then move the outside leg in. This “inward” strategy is also adopted by the elderly, but it is not the most common case. Most elderly passengers prefer standing on the floor with both

feet first and then sitting down. They all put both hands on the center pillar or the front seat to support the body weight. Interestingly, this “stand-and-sit” strategy is performed only by the elderly in this study and was rarely observed in the previous studies of sedans. For egress motion, there are also seven strategies observed. Both young and elderly passengers take the “outward” strategy most, which starts with moving the outside leg out, then standing up, and finally moving the inside leg out. The major difference is that elderly passengers require hand support on the center pillar or the front seat more than young passengers do. The findings can help automobile manufacturers determine appropriate vehicle dimensions and locations of hand grips for the better design of minivans, which will contribute to the higher satisfaction of elderly passengers.

Paper 7 Integration of Motion Capture with CAD for Camera Placement J. WAN*†, K. KOZAK, N. WANG,, G. GOMEZ-LEVI † Research and Advanced Engineering, Ford Motor Company, USA Positioning motion capture cameras for a study using a vehicle mockup is complicated due to the large amount of obstructions caused by the vehicle, the test subject’s body, as well as the potential complexity of the subject’s motion. In the past, a trial and error method has been used for camera positioning. Cameras are positioned, calibrated and then the motion of a fully markered subject is observed. Data is collected and reviewed to ensure that markers are visible throughout the motion. Adjustments are then made to camera positions and the process is repeated. A tool for streamlining this process has been developed at Ford. It uses a hardware-in-the-loop system including a programmable vehicle buck called the Human Occupant Package Simulator (HOPS) and its virtual representation in a CATIA CAD model (VHOPS). All Vicon motion capture cameras integrated with HOPS are virtually created in VHOPS and matched to their physical setting. With this method, the camera placement is accurately assessed and potential obstructions can be easily detected. This tool makes it possible to achieve optimal camera positioning for various studies and therefore improve marker visibility while greatly decreasing the time it takes to set up motion capture cameras.

Paper 8 Motion Capture and Its Application for Vehicle Ingress/Egress N. WANG*†, G. GOMEZ-LEVI, K. KOZAK, J. WAN † Research and Advanced Engineering, Ford Motor Company, USA Understanding how drivers and passengers move and interact with respect to a vehicle is very important to vehicle packaging design. Ingress/Egress is a typical example of such motion and interaction. Being able to assess a vehicle’s ingress/egress performance early in a design process is therefore crucial to successful vehicle development. Conventional methods rely on clinic research with physical bucks, which is an expensive, time consuming and subjective process. Motion capture techniques have been used in academia for biomechanical research, in the movie industry for animated films, and in the manufacturing industry for simulating and planning operators' paths in the work environment. However, no existing systems provide a systematic approach to incorporate motion capture tests and analysis in an integrated environment for vehicle design. This paper presents a system that uses motion capture techniques and digital human models for analyzing human motions during vehicle ingress/egress. The system was developed on top of the CATIA platform using CATIA’s digital manikins, though other digital manikins could be used.

Paper 9 “Abita 4T”: experimental criteria and software implementation for the prediction and comfort evaluation of tractors driver posture with Jack. V. ANTONUCCI, C. ARDUINO†, R. PIZZONI† † CRF, Centro Ricerche Fiat, Strada Torino 50, 10043 Orbassano (To), Italy The purpose of this work is to investigate the factors involved in the posture selection by tractor drivers in order to be able to simulate and predict comfort driver posture using a new tool developed in the Jack environment: “Abita4T”.

The ergonomic criteria used by Abita4T were obtained through a dedicated experimentation with 27 tractor users, segmented by gender, age, stature, and mission of the vehicle used. The experimental session had two main objectives: the first one was to obtain the range of comfort for primary human articulations involved in the comfort perception; the second one was to obtain the “logic” the driver uses to decide which posture is the best one for him/her so to reproduce the same logic in the Jack environment. Two postures have been analyzed and implemented: the driving posture and the working posture. The model can be used during vehicle design to predict and define the H point position the driver will select, the internal vehicle layout for primary tractor controls and seat and to analyze competitors vehicles according to ergonomic criteria. The tool allows to analyze the posture of a database of 41 manikins(or a subset of them) in few minutes, so it could be very useful during the first development phases when many solutions have to be explored. The analysis done with this type of model includes the evaluation of the posture overall comfort, steering wheel comfort zone, pedals usability (positions and strokes of clutch, accelerator and brake) and the reachability of some tractor controls as gear shift or multifunctional handle. All the results, for each seat step and for each MFH possible regulations, are memorized in a relational database and can be saved for future analysis. The Sw allows also users to select a position of the seat and of the MFH range regulations and reproduce the relevant layout with the manikin correctly positioned in Jack, exactly as was during the step by step analysis.

Paper 12 Assessing shift exposure to material handling tasks using the Jack human simulation tool and a cumulative low back analysis CHRISTINA CORT†, VINCE. RACCO‡, ULRICH. RASCHKE† † Siemens, 2600 Green Rd, Suite 100, Ann Arbor MI, 48105 ‡ Lear Corporation, 21557 Telegraph Road, Southfield, MI 48033 Musculoskeletal disorders (MSDs) associated with manual material handling (MMH) continue to be a leading cause of lost work time and workers’ compensation costs in industry. Human simulation tools can be used during the design phase to mitigate the risk of designing jobs that may contribute to overexertion injuries. These tools allow detailed visualization of work cell layouts as well as required human postures and motions, and quantitative assessment of musculoskeletal stressors. Despite these benefits, complex material handling processes can be difficult to analyze since they often encompass a variety of different elements, can be repetitive and occur over the course of an entire work shift. This paper reviews a method used in tandem with the Siemens Jack human simulation tool to analyze cumulative low back loads, which allows the analyst to consider exposure occurring over an entire work shift. The analysis method has proven successful at producing jobs with lowered MSD risk factors and increasingly is being adopted among the broader user community for both pro-active as well as re-active ergonomic intervention.

Paper 13 A method of human factors evaluation based on key frame posture of maintenance operation in virtual reality environment SHIGUANG QIU, XIUMIN FAN, QICHANG HE CIM Research Institute, School of Mechanical Engineering, Shanghai JiaoTong University, P.R.China Human factor analysis is an important component for product maintainability task assessment. Quantitative assessment of human factors in product design phase could be very useful for saving maintenance costs and time. Current assessment methods are mainly realized through maintenance operation postures predefined in a virtual environment, and then evaluating operation comfort and accessibility, visibility and other human factors. Such assessment methods rely on engineers to manually adjust joint angles to simulate the operation posture. It’s time-consuming and difficult to reflect the real maintenance operation accurately. In addition, the real operation is a continuous process, so a few limited operation postures are not adequate to fully reflect human factors of a product. With virtual reality interactive devices, a real operator can drive a virtual human in real-time for maintenance operation, by this means, it is easy to take into account operator’s subjective factors and

obtain accurate operation postures. In this study, a comprehensive evaluation method of human factors based on real-time maintenance operation data is put forward. Firstly, a repair operation sequence is constructed according to maintenance postures and this true repair process is mapped to virtual repair sequence database. The maintenance posture is defined through a mathematics description based on maintenance feature elements, which include a virtual operator and its attributes, product component, tool, environment model in the virtual maintenance environment. Secondly, a novel algorithm is designed to extract key frame posture of maintenance operation (KPM) from the database automatically according to the operation-related characterized elements. Thirdly, human factors are analyzed quantitatively through evaluating virtual human operation posture in each KPM. All analysis results, including eye-vision, comfort level and accessibility are weighted to establish a comprehensive human factors evaluation model. The above method is realized as a human factors assessment module and integrated into software VRME (Virtual Reality Maintenance Environment) which is developed by our VR Lab. The software provides an environment for interactive product maintenance operation and maintenance process planning, and contains libraries of virtual human, environment model, maintenance tool, parts and subassemblies of product with geometrical constraints information. With this human factors assessment module, maintenance operation datum are recorded to database in real time, and KPMs are calculated for human factors analysis, and final results are shown in a graphic form and a composite score. All these make it easy for operators to choose the best maintenance plan, to know which part is hard to repair, and to check its motion trajectories so as to find the possible existing ergonomic problem. With this human factors analysis evaluation method and software module developed for complex product maintenance planning in early design stage, a case study about a mobile radar equipment repair is given. A real operator wears a HMD and his body movement is captured by NDI tracker and hands’ joints motion captured by CyberGlove. The whole repair process is stored in database and its KPMs are extracted according to the above method. The assessment result shows the method is simple and feasible.

Paper 14 Dynamic Digital Human Model for ergonomic assessment based on human-like behaviour and requiring a reduced set of data for a simulation GIOVANNI DE MAGISTRISA,∗, ALAIN MICAELLIA, JONATHAN SAVINB, CLARISSE GAUDEZB, JACQUES MARSOTB aCEA, LIST, Interactive Simulation Laboratory, 18 route du panorama, BP6, Fontenay aux Roses, F92265 France bInstitut national de recherche et de sécurité (INRS), rue du Morvan, CS 60027, Vandœuvre-lès-Nancy, F54519 France Ergonomic methods for biomechanical risk factors assessment during work activity are usually based on human operator’s postures and forces while performing the work task. A basic analysis of the task can rely on questionnaires, interviews and video analysis, but a more accurate and comprehensive analysis requires collection of exertion (force sensor and/or electromyography) and posture data (e.g. motion capture technique). Such an analysis entails complex and expensive instrumentation that may hamper task performance. In recent years, an alternative solution appeared with the use of digital human models (DHM) for ergonomics analysis. Actually, using industrial DHM software packages available for ergonomic assessment is usually a complex and time-consuming task. For instance, simulation may be built up like a cartoon through interactive manual positioning of the DHM with mouse, menus and keyboard, which entails expert skills in ergonomics and human motion. One can also use a tracking system or motion capture to get realistic simulations but this requires extensive instrumentation, a full scale mock-up of the future workstation, or a similar one, and tricky motion capture data processing. Simulations can eventually be based on pre-defined human motion libraries (reach, grasp...), which usually look quite unnatural. Besides postures, ergonomics assessment also needs at least an estimate of exertions and task execution time. For these reasons, using industrial DHM software for biomechanical risk factors assessment based on simulations of industrial or experimental work task situations may lead to significantly misleading stress estimation. A challenging aim therefore consists in developing a virtual human model capable of computing automatic, dynamic, realistic movements and internal characteristics (position, velocities, accelerations and torques) in quasi-real time, based on a simple description of the future work task, to achieve realistic ergonomics assessments of various work task scenarii at an early stage of the design process. We have developed such a dynamic DHM automatically controlled in force and acceleration [1], inspired by human motor control [2] and based on robotics and physics simulation (in our simulation framework, the DHM motion is dictated by real-world Newtonian physical and mechanical simulation, along with automatic control of applied forces

and torques). Our controller performs multiple tasks simultaneously (balance, contacts, manipulation) in real time with human-like feedforward force and impedance [3]. An experimental insert-fitting activity has been simulated and assessed based on OCRA ergonomic index. In comparison with a real human data-based assessment, consistent results have been obtained. The most interesting property of our DHM is that it requires minimal information for a simulation: a starting point, an intermediate point for obstacle avoidance and an end point, along with the applied force for insert clipping. Moreover, changing the subject’s anthropometry and the scenario does not require new trajectory specification nor additional tuning. Joint torques, DHM movements and trajectory and their related OCRA assessment are realistic and consistent with human-like behaviour and performance.

Paper 15 A finite element model of seat cushion indentation with a soft tissue human occupant model G. PAUL*†, J. MILLER‡, J. PENDLEBURY†† † Queensland University of Technology, School of Public Health and Social Work, Brisbane, Qld, Australia ‡ Futuris Automotive Interiors Pty Ltd, Port Melbourne, Vic, Australia †† Ford Motor Company Australia, Melbourne, Vic, Australia Effective digital human model (DHM) simulation of automotive driver packaging ergonomics, safety and comfort depends on accurate modelling of occupant posture, which is strongly related to the mechanical interaction between human body soft tissue and flexible seat components. This paper presents a finiteelement study simulating the deflection of seat cushion foam and supportive seat structures, as well as human buttock and thigh soft tissue when seated. The three-dimensional data used for modelling thigh and buttock geometry were taken on one 95th percentile male subject, representing the bivariate percentiles of the combined hip breadth (seated) and buttock-to-knee length distributions of a selected Australian and US population. A thigh-buttock surface shell based on this data was generated for the analytic model. A 6mm neoprene layer was offset from the shell to account for the compression of body tissue expected through sitting in a seat. The thigh-buttock model is therefore made of two layers, covering thin to moderate thigh and buttock proportions, but not more fleshy sizes. To replicate the effects of skin and fat, the neoprene rubber layer was modelled as a hyperelastic material with viscoelastic behaviour in a NeoHookean material model. Finite element (FE) analysis was performed in ANSYS V13 WB (Canonsburg, USA). It is hypothesized that the presented FE simulation delivers a valid result, compared to a standard SAE physical test and the real phenomenon of human-seat indentation. The analytical model is based on the CAD assembly of a Ford Territory seat. The optimized seat frame, suspension and foam pad CAD data were transformed and meshed into FE models and indented by the two layer, soft surface human FE model. Converging results with the least computational effort were achieved for a bonded connection between cushion and seat base as well as cushion and suspension, no separation between neoprene and indenter shell and a frictional connection between cushion pad and neoprene. The result is compared to a previous simulation of an indentation with a hard shell human finite-element model of equal geometry, and to the physical indentation result, which is approached with very high fidelity. We conclude that (a) SAE composite buttock form indentation of a suspended seat cushion can be validly simulated in a FE model of merely similar geometry, but using a two-layer hard/soft structure. (b) Human-seat indentation of a suspended seat cushion can be validly simulated with a simplified human buttock-thigh model for a selected anthropomorphism.

Paper 16 Postural and muscular adaptations to repetitive simulated assembly line work S. EBATA†, S. SELBIE‡, C. TSE†, P.J. KEIR*† † Department of Kinesiology, McMaster University, Hamilton, Ontario, Canada ‡ C-Motion, Inc., Rockville, Maryland, USA & HAS-Motion Inc., Kingston, Ontario, Canada Few studies have shown the process of adaptation in joint kinematics during prolonged repetitive work involving complex tasks. Fifteen healthy men performed 61 cycles that were one minute in duration of automotive-related tasks, which included a finger pull, knob turn, drill press and hose connector push. Kinematics of the upper extremity and trunk were collected at 60 Hz using a 3D motion capture system. Marker data were imported and modeled using Visual3D software (C-Motion, Inc., Rockville, Maryland, USA). Data were analyzed at 12-minute intervals, with data divided into the work cycle (start to finish of 4 sub-tasks) and drill task. The mean peak, minimum and coefficient of variation (COV) were computed for all 3D joint angles. Results showed the time to complete the work cycle decreased by 6.3 s over the trials. Peak shoulder flexion decreased and peak elbow flexion increased during the work cycle. This study found adaptations to highly repetitive but light work in only one hour. This study is one of the first to examine adaptations to a highly repetitive simulated assembly work and has provided new insights into the evaluation of complex repetitive jobs as a whole.

Paper 17 Validation Experiments of Ergonomic Evaluation methods Effectiveness of Complicated Human Control System Characteristics ZHIQIANG TIAN1*, TING JIANG1, CHUNHUI WANG1, YIJING ZHANG1 1National Key Laboratory of Human Factors Engineering, China Astronaut Research and Training Center, Beijing 100094, China The objective is to validate the effectiveness of evaluation methods basing on the manual rendezvous and docking (MRVD) integrative performance model. According to established ergonomic evaluation methods, On the MRVD ergonomic experiment platform, researcher selected a kind of MRVD’s human computer interface as a complicated human control system and three experiment schemes which difficulty is gradually increased were designed. 17 subjects who have the MRVD skills were arranged to fulfill the experiments. Subjects manipulated docking two times in every scheme and all performance data were recorded during the MRVD. 1. Experiments data show that subjects are inclined to control relative attitude biases rather than position biases at the beginning of MRVD. Approximate 20m position before docking is watershed of racking control stage and accurate control stage. 2. With the increasing of experiment schemes difficulty, the final docking time and fuel wastage has the trend of increase, but the main effect is not significant (F=1.554, P>0.5 and F=2.906, P>0.5 ). 3. The main effect of docking accuracy of three experiment schemes is also not significant (F=2.797, P>0.5). 4. At tracking control stage and accurate control stage, the integrative performance of three experiment schemes have not notable difference (F=1.319, P>0.5 and F=0.107, P>0.5). 5. Copper-Harper scale score and NASA-TLX scale score show the same result with integrative performance data. 1. Experiments testify the course of MRVD are composed of tracking control stage and accurate control stage, subjects contribute more mental resources to translation distance deviation during accurate control stage. 2. Control performance model and subjective evaluation can ensure the effectiveness of evaluation results of complicated human-display-control-system. 3. Entropy can be used to compute the weight coefficient of all freedoms performance indexes of docking time, and factor analysis can be used to estimate the integrative performance of the indexes of control process. The weight coefficient of docking accuracy and the indexes of control process are analyzed by expert assign method. 4. The method of increasing the task difficulty gradually can effectively estimate the redundancy of complicated human control system ergonomic characteristics.

Paper 18 A 3-D dynamics analysis of driver's Ingress-Egress Motion T. ROBERT*†, J. CAUSSE†‡, LISA DENNINGER‡, XUGUANG WANG† † LBMC, Université de Lyon, IFSTTAR – UCBL, F-69675 Bron ‡ PSA Peugeot-Citroën, Vélizy-Villacoublay, France Ingress-egress (I/E) is a major ergonomic challenge for car manufacturers. I/E motions have been largely studied in kinematics, but very rarely in dynamics. However, the joint loads developed during car I/E motions, and in particular the motor torques (torques acting along the degrees of mobility of the joints), might be an important parameter to explain the difficulties perceived by drivers while getting in and out of cars. In order to fill this gap, we propose in this study: 1/ to estimate and describe the joint torques developed during ingress-egress motions 2/ to analyze the influence of the car geometry and subject anthropometry on the torques developed. An experiment was performed to record the kinematic and contact loads of subjects getting in and out a car mock-up. Three groups of subject, based on their anthropometry, and three vehicle configurations were tested. Net joint loads were then extracted using an inverse dynamics analysis. Motor torques were obtained by projecting these net joint torques along the degrees of mobility. Motor torques were stereotypical across subjects and conditions. Both ingress and egress motions were primarily characterized by large extension torques corresponding to the seating/rising phase. These torques are used to elevate (egress) or control the lowering (ingress) of the center of mass. They are similar to those observed in Sit-to-Stand maneuvers. On overall, the taller a subject and the lower the seat of the vehicle is, the larger the peak torques are. Moreover, peak torques were higher for egress than ingress. These results could be related to previous studies showing that getting in and out is easier for minivans than for smaller cars, and that egress is more difficult than ingress.

Paper 19 The Effect of High Pass Filtering and Non-Linear Normalization on the EMG-Force Relationship during Sub-Maximal Finger Exertions A.C. MCDONALD, K. SANEI, P.J. KEIR* Occupational Biomechanics Laboratory, Department of Kinesiology, McMaster University, Hamilton, ON, Canada, L8S 4K1 Muscle force estimates are important for full understanding of the musculoskeletal system and EMG is a commonly used method to model muscle force. The purpose of this investigation was to examine the effect of high pass filtering (HPF) and non-linear normalization on the EMG-force relationship of sub-maximal finger exertions and to examine the effect of high pass filtering on the linearity of the relationship. Participants performed isometric ramp exertions while seated with their right index finger in an adjustable padded ring attached to a force transducer. Bipolar surface EMG was recorded on the second compartments of the extensor digitorum (ED2) and the flexor digitorum superficialis (FDS2). There were three force application conditions: extension at the mid-proximal phalanx (EXT), flexion at the proximal phalanx (FLXp) and flexion at the distal phalanx (FLXd). These data were analyzed using 30 high pass filter designs and compared to a standardized processing procedure. Non-linear normalization was completed using an exponential fit equation. In the distal flexion (FLXd) condition, HPF significantly (p<.05) reduced the %RMS error between the EMG and force and increased the peak cross correlation when compared to standardized processing. In the FLXp and EXT conditions, there were trends present suggesting possible improvements to the relationship (p>.05). Non-linear normalization improved the EMG-force relationship, demonstrated by reduced %RMS error and increased correlations. The optimal equation constant to maximize the relationship depended on the HPF design and the exertion condition. The different optimal processing parameters suggest that high pass filter design and linearity are dependent on contraction mode and the muscle analyzed.

Paper 20 Introducing an Ergonomics Watch Tool in Early Phases of Products’ Design V. CALATORU*†, S. BASSETTO†, D. BROUILLETTE‡, J. CHARLAND‡ † Département de mathématiques et de génie industriel (École Polytechnique de Montréal), Montréal, Canada ‡ Virtual Ergonomics (Dassault Systemes), Montreal, Canada One of the major limitations of actual Computer Aided Design strategies is that many of the constraints are not considered until very late in the cycle of development. The delay until the moment when ergonomics, among other constraints are checked is a major source of rework, therefore of increased costs and frustration for designers and companies alike. Many CAD systems offer components that make evaluations from the point of view of ergonomics, but they require complex abilities and come usually later in the design process, usually when the design concept is set (Burns and Vicente 2000). The main objective of the research presented in this paper is to analyze the ways the ergonomic issues are dealt with now and to propose a methodology that could facilitate the assessment of the ergonomic issues in the early stages of concept elaboration. In order to support this methodology, we introduce several concepts that were used to define and build an experimental tool to guide designers early in the concept elaboration, such as the use of virtual envelopes and a scale for evaluation of the ergonomic risks for products. A brief example of use of this tool is presented.

Paper 21 The use of humanoid glyphs in graphs for representing human variability in the spatial design of products C. GARNEAU*†, M. PARKINSON‡ †The Pennsylvania State University, currently with the U.S. Army Research Laboratory Human Research and Engineering Directorate ‡OPEN Design Lab, The Pennsylvania State University From simple color-coding to more intricate forms like the “Chernoff face”, glyphs are a means of visualizing multidimensional data. The spatial design of products for their users presents a unique opportunity for using glyphs as a visual metaphor for encoding data, chiefly due to the multivariate nature of many commonly encountered problems. While relatively intricate glyphs called “Digital Human Models” (DHMs) have been studied for the evaluation of user fit in the design of products for decades, they typically only exist as an illustration within the physical environment. However, the current paper explores a novel strategy of positioning simple humanoid figures in relational graphs (e.g., scatter plots) to facilitate decision making in multidimensional design scenarios. This technique is particularly useful when used in combination with a quantitative optimization technique called “virtual population fitting”. Results from a web-based survey with 94 participants compare the intuitiveness of two types of humanoid representations and the comprehension of glyphs to solve problems involving human variability.

Paper 22 Simulation of Flexible Part Installation DIANA WEGNER, WEI ZHOU, NIGUL ILVES, SIMON HALL, MATT REED, JOHAN S. CARLSON, ROBERT TILOVE, THOMAS ARMSTRONG Exciting new infotainment, communication and alternative power systems are being developed by the automobile industry in response to customer demand. Examples include vehicle sync with smart phones and personal music devices like iPods, as well as electric and hybrid power systems. These systems are highly successful and are establishing a customer expectation for more easy to use vehicle electronics. Manufacturing engineering processes must accommodate the additional electrical product content and associated manufacturing challenges. For example, high voltage electrical cables can be stiff and difficult to install. Models that are used to simulate this new assembly work can help to ensure high quality products and efficient assembly operations. However, current product lifecycle management systems (PLMS) have digital human models but have limited capability for bio-mechanic human posture prediction and dynamics. PLMS also lack flexible parts modeling and assembly force prediction to simulate electrical wire harness installation. Application of these capabilities early in the product development process will reduce

reliance on prototype and integration builds and could save significant cost by resolving issues early, efficiently and accurately, addressing the challenges facing industries who manufacture products with electrification content. We present a proof of concept system and case studies which demonstrate vehicle wire harness installation. Siemens Classic Jack, the University of Michigan HUMOSIM Framework and the Fraunhofer Chalmers Center (FCC) Industrial Path Solutions (IPS) software was used to demonstrate wire harness manipulation and installation tasks. Forces required to manipulate and install the wire harness were calculated as inputs to bio-mechanic digital human posture prediction for accurate modeling of wire harness installation assembly tasks.

Paper 23 The influence of a driving task on movement times of goal directed hand-arm movements F. KREMSER*†, M. GEBHART†, M. STECHER†, K. BENGLER† † Institute of Ergonomics, Technische Universität München, Boltzmannstr. 15, 85747 Garching, Germany A model based approach for the prediction of interaction times is needed for the evaluation of display and control concepts during the early virtual phase of the vehicle design process. The main objective of the described experiment is to assess the influence of a driving task on movement times of goal directed handarm movements. Another objective was to create a dataset to empirically model motion times and to surrport the adaptability of Fitts’ law for goal directed hand-arm movements in vehicle conditions. The experiments were conducted with 31 subjects in a fixed based driving simulator. In Scenario A the subjects performed a pointing only task, and in Scenario B, a driving task and the pointing task were performed together. Motions were captured using a VICON MX system. The variables scenario F(1, 27)=14.25, p<.001 and target distance F(4.97, 134.21)=174.962, p=.001 had a significant influence on movement time. Movements were performed 6% faster in Scenario B than in Scenario A. A stepwise linear regression showed that the distance from the origin of the movement to the target is the best predictor for the calculation of movement time.

Paper 24 Comparative analysis of human modeling tools EMILIE POIRSON & MATTHIEU DELANGLE LUNAM, IRCCYN, Ecole Centrale de Nantes, France Digital Human Modeling tools simulate a task performed by a human in a virtual environment and provide useful indicators for ergonomic, universal design and representation of product in situation. The latest developments in this field are in terms of appearance, behaviour and movement. With the considerable increase of power computers, some of these programs incorporate a number of key details that make the result closer and closer to a real situation. With the differences in terms of performance, qualities, limitations, the choice of the tool becomes complicated in this wide range of possibilities. In this context, we propose to study and compare the most human modelling software available on the market, and thus provide an aided decision tool to help the designer to get the most adaptable tool.

Paper 25 Occupant Packaging of a Six-Seat Sports Car Using RAMSIS YUBIN XI Clemson University International Center for Automotive Research (CU- ICAR) Deep Orange is a vehicle prototype project in the Clemson University International Center for Automotive Research. The third generation of the Deep Orange features a hybrid all-wheel-drive six-seat sports car designed for Generation Y, the NBT (“Next Big Thing”). The objective of this paper is to illustrate how RAMSIS was involved in the occupant packaging process of the NBT. Among all design aspects, vehicle conceptualization, seating layout, driver’s reachability and visibility are going to be discussed. The SAE recommended design practice was attempted at the beginning stage and it will be compared with the design practice by using RAMSIS. It is proved that RAMSIS helps designers to visualize the occupant packaging process dynamically.

Paper 26 Generation of a Midsize-Male Headform by Statistical Analysis of Shape Data MATTHEW P. REED *†, BRIAN D. CORNER‡ † University of Michigan Transportation Research Institute ‡ US Army Natick Soldier Research, Development, and Engineering Center Human surrogates for safety applications, such as crash dummies, are often constructed to represent the mean or average of individuals within a population. Generating accurate mean headforms is particularly challenging, even with three-dimensional (3D) surface measurement technology, because hair limits the ability to accurately represent the shape of the head. We present a new midsize-male headform generated using statistical analysis of head dimensions and landmark data, combined with a reference shape obtained by averaging the head shapes of two men close in size to the target stature and body weight. The locations of 26 head and face landmarks from 1747 men were extracted from the 1988 U.S. Army Anthropometric Survey (ANSUR) and registered using a Procrustes superimposition. A linear regression predicted landmark locations as a function of stature and body mass index, using target values of 1755 mm and 27.3 kg/m2. To construct a reference head shape, two bald men whose head and face dimensions were within 2% of the target mean values were extracted from a large database of head scans. The scans were visually aligned using landmark locations, then resampled cylindrically and averaged. The reference head shape was then morphed to match the target landmark locations using a radial-basis-function (RBF) morphing method with a multiquadric kernel. A second morphing step was performed to match the desired head length, head breadth, and the distance from the tragion landmarks to the top of the head. In contrast with headforms obtained solely by averaging individuals close to the desired body size, the current headform is generated from a statistical model that represents a large range of head size and shape, and hence a broad set of statistically consistent headforms can be generated programmatically without additional analysis. Approved for public release.

Paper 27 Human Shape Modeling and Characterization at Multiple Scales Z. CHENG*†, J. CAMP‡,, K. ROBINETTE** † Infoscitex Corporation, Dayton, Ohio, USA ‡ 711th Human Performance Wing, Air Force Research Laboratory, Dayton, Ohio, USA **College of Human Sciences, Oklahoma State University, Stillwater, Oklahoma, USA A new method for building static human shape models from three dimensional body scan data is introduced. The method is based on body segmentation and contour lines with the major steps including joint center calculation, skeleton model building, segmentation, slicing, discretization and registration. Based on this method, wavelet analysis is used for human shape parameterization and characterization. Two dimensional wavelet analysis is performed on the matrix formed by the discretized contour lines of a segment, and the segment shape is then described by a matrix of wavelet coefficients. The assembly of the wavelet coefficients of all segments is used as the shape descriptor for the entire body. Since wavelet analysis is a multi-scaling analysis by its nature, the characterization of shape variation can be readily conducted at multiple scales. Due to the perfect reconstruction provided by wavelet analysis, the body shape can be fully reconstructed from wavelet coefficients. Alternatively, the body shape can be reconstructed at different scales, which provides the approximations to the original shape at different levels.

Paper 28 Comparing RAMSIS driving posture predictions with experimental observations for defining optimum task constraints JEANNE BULLE†, GIANCARLO CACCIA DOMINIONI‡, XUGUANG WANG†*, SABINE COMPIGNE‡ † Université de Lyon, F-69622, Lyon, France, IFSTTAR, UMR_T9406, LBMC, Université Lyon1 ‡ Toyota Motor Europe, Belgium In this study, driving postures predicted by RAMSIS were compared with experimental observations from 4 different vehicles (small to minivan) and 3 different stature groups of 34 males and females. Four configurations were tested for each vehicle allowing an estimation of intra-individual variability of driving

postures. Globally, average differences between RAMSIS and real observations for the hip location were -14 and 3 mm in longitudinal (x) and vertical (z) directions. For the eye, their differences were higher, -35 and 16 mm respectively in x and z directions. Though global average differences were quite small, high dispersions according to subject group and vehicle were observed. If we refer to the range of intraindividual variation in hip and eye locations for assessing RAMSIS predictions, only 30% and 18% of the predictions can be considered as good. Moreover, differences between RAMSIS predictions and observations were significantly affected by subject group and vehicle. RAMSIS tended to predict a lower hip location for short persons and a higher hip position for tall ones. RAMSIS prediction performance also depended on vehicle. This suggests different sets of constraints and/or different posture models should be used according to stature group and vehicle.

Paper 29 3D Digital Platform development to analyze 3D digital human models. A case of study of Jiu Jitsu combat sport C. P. GUIMARÃES †, F. RIBEIRO † G.L. CID, P. STREIT‡, J. OLIVEIRA†, M.C .ZAMBERLAN†, A.G.PARANHOS†, F. PASTURA† † Instituto Nacional de Tecnologia ‡ Universidade Estadual do Rio de Janeiro The purpose of the study is to present a 3D Digital platform that has being developed in a game engine software to work with motion capture analysis using data captured by Xsens system, IPI studio/kinetics System or any other mocap system and has being applied to the study of Jiu-jitsu. This platform has being developed considering the need to analyze data from the same athletes movements being repeated in different time or even to compare and supplement the same athletes movements being captured with different mocap systems. The main 3D Digital Platform advantage is its flexibility to handle motion capture data from different mocap systems in order to facilitate kinematic analysis by users of motion capture system. The 3D platform development follows some specific steps, which make it possible not only to visualize the performed motion but also turn possible the interaction between the user and the 3D character. The first step consist on the automatically reconstruction of the 3D character body segments based on motion capture data. The visual representation has as benefit the reduced noise that may be generated in the process of retargeting the motion capture data to a specific rig and character that differs from the actual bone structure original data. The visual representation is generated based on Xsens data. This makes the representation being a precise copy of original bone position and structure of the specific actor's movements being captured. The second step is linking each bone segment of the procedurally by generating 3D model with a collision area that is necessary for future interaction with user. After those steps, the user can select : a specific body segment to track and generate data; to play/pause motion; to build a graph of segmental angles, joint angles and angular velocity. These functionality is still under development and tests. The first application of this 3D digital platform was analyzed the Jiu-Jitsu athletes’ movements. This analysis allowed studies and improvements of athletes’ performance. In the future the integration between a 3D scanned athlete’s model and a virtual environment will permit development a virtual simulator that can be applied to education, training and entertainment.

Paper 33 Morphological Analysis of Changes in the Thoracic Skeleton with Age and Gender ASHLEY A. WEAVER*†‡, CALLISTUS M. NGUYEN†‡, SAMANTHA M. SCHOELL†‡,, JOEL D. STITZEL†‡ † Virginia Tech-Wake Forest University Center for Injury Biomechanics ‡ Wake Forest University School of Medicine Age and gender-specific variations in the geometry and mechanics of the thoracic skeleton are expected to relate to thoracic injury. The goal of the study was to improve the understanding of thoracic injury by quantifying age and gender-specific variations in the thoracic skeletal morphology for males and females of ages 0-100. Age and gender-specific size and shape variations in the ribs and sternum were quantified using the following steps: 1) Radiological Scan Collection, 2) Image Segmentation, 3) Atlas Development, 4) Image Registration, and 5) the Generalized Procrustes Analysis (GPA). The image segmentation and registration algorithm developed allowed for the collection of extensive homologous landmark data from the ribs and sternum. Significant changes in size and shape for all 24 ribs and the sternum were found to occur with age in both genders. Significant changes in shape alone were found to occur with age for the female sternum and the left and right ribs 1-11 in both genders. The variation functions describe the size

and shape changes in the thoracic skeleton and will be used in future work to generate a parametric finite element model of the thorax. The results of this study provide an improved understanding of the complex relationship between thoracic anthropometry, age, gender, and injury risk.

Paper 35 A Model to Predict Carpal Tunnel Syndrome Risk in the Workplace P.J. KEIR*, J. INMAN, J.A. WERESCH Occupational Biomechanics Laboratory, Department of Kinesiology, McMaster University, Hamilton, ON, Canada, L8S 4K1 CTS remains an issue in the workplace. Increased carpal tunnel pressure (CTP) may lead to the aggravation or development of CTS. CTP of 30 mmHg or higher is associated with CTS symptoms; thus 30 mmHg has been used as a threshold limit value for CTS risk. Deviation from a neutral wrist, neutral forearm, and relaxed fingers results in an increase in CTP. Fingertip loading has also been shown to increase CTP independently of posture. The purpose of this study was to develop an ergonomic tool to predict carpal tunnel syndrome (CTS) risk based on the carpal tunnel pressure (CTP) in healthy wrists. A tool was developed to predict CTS-risk based on CTP. The tool was evaluated by comparing the output of the program (CTS risk) to incidence of CTS in a manufacturing environment. No differences were found for CTS risk between jobs with no incidence of CTS versus jobs with an incidence of CTS. While the tool predicted CTS risk based on CTP, too few CTS claims existed to develop a strong correlation. Further refinement and investigation is needed to include combined postures and mechanical compression, and to further validate tool.

Paper 36 An Unconstrained Kinematic Model of the Hand D.M. COCCHIARELLA, A.M. KOCIOLEK, P.J. KEIR* Department of Kinesiology, McMaster University, Hamilton, ON, Canada, L8S 4K1 A hand model that represents its diverse capabilities is needed to assess injury risk. We improved upon an existing model with grouped metacarpals by adding the trapeziometacarpal joint of the thumb and defining individual metacarpals to represent the metacarpal arches of the hand with motion capture. Eight participants performed 3 static postures (neutral pose, cylinder grip, cap grip) to evaluate kinematic performance of the models (aggregate versus segmented metacarpals). Joint rotations and translations served as the outcome measures. Transverse metacarpal arch angles of the segmented model were between 18.9° ± 3.0° (neutral pose) and 29.7° ± 3.8° (cap grip). Abduction arch angles were also large with values ranging from 35.8° ± 0.9° (neutral pose) to 43.5° ± 1.3° (cylinder grip). Representing the metacarpal arches with individual metacarpals also influenced finger joint kinematics. The 4th and 5th metacarpophalangeal joints displayed lower supination angles in the segmented versus aggregated model with mean differences ranging from 9.3° (neutral pose) to 17.0° (cap grip) in the 4th finger and 16.3° (neutral pose) to 33.0° (cylinder grip) in the 5th finger. The 2nd metacarpophalangeal joint also showed higher supination angles in the segmented (versus aggregate) model during gripping (mean differences: cylinder grip ‒ 12.2°; cap grip ‒ 9.5°). Furthermore, metacarpophalangeal abduction/adduction angles generally decreased (closer to 0°). Representing individual metacarpals and the trapeziometacarpal joint enhanced anatomic fidelity of the hand model and also improved kinematic performance compared to previous radiographic studies. Ultimately, this model will better serve future studies of physical exposures in the workplace.

Paper 37 Reassessing Flexor Tendon Excursion Models with Colour Doppler Ultrasonography A.M. KOCIOLEK†, P.J. KEIR*† † Department of Kinesiology, McMaster University, Hamilton, ON, Canada, L8S 4K1 Patients with carpal tunnel syndrome present with fibrosis of the subsynovial connective tissue (SSCT) next to the flexor tendons, suggesting that shear forces play a role in injury development. Researchers are increasingly utilizing diagnostic imaging to assess musculoskeletal dynamics in vivo. Recent ultrasound studies found that flexor tendon excursions did not match traditional cadaveric models of tendon-joint interaction, which are often used to assess occupational injury risks. To better link ergonomic factors with

pathomechanics, ultrasound was used to measure relative motion of the long finger flexor digitorum superficialis (FDS) and adjacent SSCT in the carpal tunnel. Sixteen participants (8 ♀, 8 ♂) completed three cyclical finger movements (metacarpophalangeal, proximal and distal interphalangeal joints, full finger flexion/extension) in three wrist postures (30° flexion, 0°, 30° extension) with motion analysis of the wrist and long finger. Full finger movements produced larger tendon excursions than metacarpophalangeal and interphalangeal motions (full finger: 21.7 ± 1.0 mm; metacarpophalangeal: 14.0 ± 0.7 mm; interphalangeal: 11.9 ± 0.6 mm). Relative displacements between the FDS and SSCT increased linearly by 32.3% from 30° wrist extension to 30° wrist flexion (F2,30 = 16.2, p < 0.01). This relationship was primarily driven by excursion changes of the SSCT, which decreased by 22.1% with wrist flexion (F2,30 = 9.9, p < 0.01). This study shows that the wrist may be more prone to shear injuries in flexion. We are currently integrating the data into an analytical model that predicts FDS and SSCT motion as well as injury risk in the workplace.

Paper 39 Biofidelic Hybrid Modeling of Skeletal Muscles HYUNG YUN CHOI*1 , MASATO ITO2, SADAYUKI UJIHASHI2, CHRISTIAN MARCA3 1Hongik University, Korea 2Tokyo Institute of Technology, Japan 3ESI Group, France An active joint motion of human body is normally produced by non-isometric skeletal muscle contractions through the motor control. These non-isometric skeletal muscle contractions induce substantial shape change of the muscle segment, e.g., bulged belly of concentrically contracting muscles and consequently the interaction with the tight clothing or footwear becomes an important design issue. In order to simulate both contractile muscular forces and associated volumetric changes using digital human body models, the FE based hybrid muscle modeling approach in which three-dimension solid elements are reinforced with onedimensional Hill type bar elements respectively representing muscle substrates and fibers is introduced. The lower leg muscles and their actions on ankle joint motions (dorsi/plantar-flexions) are accordingly modeled, simulated and quantitatively validated against a live subject experiment.

Paper 41 Integrating Biomechanical Manikins into a CAD Environment M. JUNG†, M. DAMSGAARD†, M.S. ANDERSEN‡, J. RASMUSSEN*‡ † AnyBody Technology A/S, Niels Jernes Vej 10, DK-9220 Aalborg East, Denmark ‡ M-Tech, Aalborg University, Fibigerstræde 16, DK-9220 Aalborg East, Denmark This work investigates the opportunity to combine the best features of digital manikins and detailed musculoskeletal models. An interface between a makeshift digital manikin developed in SolidWorks and a detailed musculoskeletal model in the AnyBody Modeling System is developed. The interface reveals that the detailed musculoskeletal model solves some of the inherent problems in digital manikins and allows for a detailed analysis of the kinetics of the system, including the ergonomic influence of muscle and joint forces.

Paper 42 Computational Modeling of Feature and Conjunction Visual Search Tasks Using Queuing Network-Model Human Processor (QN-MHP) F. FENG*†, Y. LIU† † Department of Industrial and Operations Engineering, University of Michigan, Ann Arbor, MI Human visual search generally refers to the task in which a person looks for a target object among distractors. It is regarded as one of the most basic human cognitive behaviors. This paper aims at building a cognitive-architecture-based visual search model based on Queuing Network-Model Human Processor (QN-MHP) that incorporates the recent findings from the visual search studies in both the psychological and neuroscience fields. Simulations were conducted for the feature and conjunction search, and the model outputs were found to be in accordance with the experiment data. This work added support to a set of assumptions that are borrowed from some other visual search findings, and it could be used to enhance the visual perceptual subnetwork for the current QN-MHP.

Paper 43 Markerless landmark localization on body shape scans by non-rigid model fitting S. YAMAZAKI∗†, M. KOUCHI†,, M. MOCHIMARU† †Digital Human Research Center (National Institute of Advanced Industrial Science and Technology), Tokyo, Japan We present a method of localizing anatomical landmarks on the human body by fitting a template mesh model to individual scans. The template is created from a corpus of body scans, and then registered onto individual scans using a non-rigid variant of iterative closest point (ICP) techniques. The proposed algorithm is robust against such defects as incomplete body parts and presence of background objects, and capable of generating a high-resolution mesh model without large geometric distortion. To evaluate the performance of the proposed system, 1) the template models of Japanese female and male were created from the AIST/HQL Anthropometric Database 2003, 2) the average error in the estimated landmark positions were examined using a different dataset, and 3) the accuracy and reliability of the landmark positions were evaluated according to the criteria specified in ISO 20685. The experiments confirm that the accuracy depends on landmarks, and fifty-six landmarks out of sixty-four can be localized within allowable error variations.

Paper 46 An iterative tasks-priority optimization-based method for predicting driving posture JUNFENG PENG†‡, XUGUANG WANG*‡, L. DENNINGER‡ † Universitéde Lyon, F-69622, Lyon; IFSTTAR, LBMC, UMR_T9406, Bron; UniversitéLyon 1, Villeurbanne, ‡ PSA Peugeot-Citroën, Sochaux, France An accurate prediction of driving posture is essential for the ergonomic analysis of vehicle interior design. The objective of this study is to present an iterative tasks-priority optimization-based method for predicting driving posture. It is based on the task-priority inverse kinematics (IK) architecture and implemented in RPx, a motion reconstruction, analysis and simulation package developed at IFSTTAR. This study was motivated by considering that among the multi-task constraints for driving, some may have higher priority and must be satisfied at first, such as a good reach of critical controls (e.g. clutch pedal) and a good road vision. The algorithm was tested with three scenarios, showing that the algorithm can take into account the defined priority order of multi-constraints imposed by driving tasks. The weighting and priority strategies were also compared. Simulations showed that the weighting and priority strategies resulted in nearly the same prediction when all constraints can be satisfied at the same time. Once conflicts between tasks arose, the priority strategy tried to satisfy the constraints of higher priority at first and then to find a solution tending to satisfy the constraints of lower priority as much as possible without affecting the fulfillment of the constraints of higher priority. The weighting strategy found a compromise among tasks that satisfied none of the tasks exactly. Besides, simulations also show that prediction is dependent on reference posture. Further investigations on the definition of task constraints and their priority as well reference posture are required.

Paper 47 Comparison and evaluation of biomechanical parameters of motion capture systems P. STREIT*A-B, A.S. MONATB, M.C.P.L. ZAMBERLANA,, C.P. GUIMARÃESA , F.C. RIBEIROA, J.L. OLIVEIRAA a Instituto Nacional de Tecnologia – INT/MCTI b Escola Superior de Desenho Industrial - Universidade do Estado do Rio de Janeiro ESDI/UERJ The present study is an under development research project, which’s goal is to evaluate kinematic and anthropometric data from Microsoft Kinect. MS Kinect is already being studied since its release (2010), once it is an inexpensive alternative to the available motion capture (MOCAP) systems found in the market. The validation of its biomechanical characteristics will allow its use in gathering information to complement Ergonomic Work Analysis and other user research that provide product and work environment diagnosis through postural and dynamic assessment.

MS Kinect was used through iPi Soft, since it can acquire depth data from two consoles at the same time and due to the fact that its biomechanical model has a similar joint configuration, as opposed to other applications and motion recognition software. MS Kinect’s data were compared with MVN Xsens inertial MOCAP system and 2D data based on video recording in two laboratorial experiments and a study-case based in work environments of oil and gas laboratories in the Ergonomic Work Analysis (EWA) context. A 3D Digital Platform has been developed in order to democratize information acquired from the different 3D systems without changing original data, allowing them be compared with the same output. This paper presents methodology and results from the laboratorial experiments through two static comparisons and one dynamic. The static comparisons were: joint angles in selected static postures, where 2D data were used in the sagittal plane; and segments dimensional data, where the first goal of comparison was to ascertain if MS Kinect through iPiSoft considered the segments as rigid bodies, and the second was held using the volunteer’s real dimensions, extracted with anthropometers in order to validate anthropometric data. Since each 3D system has its own local coordinate system, the dynamic comparison was based on global orientation as set in the 3D Digital Platform. Differences between the biomechanical models were considered and only joints that represented the same anatomical points were compared. Results have shown similar values and variations between the systems. Therefore, it is a valid MOCAP alternative in the means of kinematic and anthropometric analysis.

Paper 48 Introducing Stability of Forces to the Automatic Creation of Digital Human Postures DELFS N*† , BOHLIN R† , HANSON L§, HÖGBERG D‡, CARLSON J.S.† † Fraunhofer-Chalmers Research Centre, Gothenburg, Sweden ‡ Högskolan i Skövde, Skövde, Sweden § Industrial Development, Scania CV, Södertälje, Sweden Although the degree of automation is increasing in manufacturing industries, many assembly operations are performed manually. To avoid injuries and to reach sustainable production of high quality, comfortable environments for the operators are vital. Poor station layouts, poor product designs or badly chosen assembly sequences are common sources leading to unfavorable poses and motions. To keep costs low, preventive actions should be taken early in a project, raising the need for feasibility and ergonomics studies in virtual environments long before physical prototypes are available. Today, in the automotive industries, such studies are conducted to some extent. The full potential, however, is far from reached due to limited software support in terms of capability for realistic pose prediction, motion generation and collision avoidance. As a consequence, ergonomics studies are time consuming and are mostly done for static poses, not for full assembly motions. Furthermore, these ergonomic studies, even though performed by a small group of highly specialized simulation engineers, show low reproducibility within the group. Effective simulation of manual assembly operations considering ergonomic load and clearance demands requires detailed modeling of human body kinematics and motions as well as a fast and robust inverse kinematics solver. In this paper we introduce a stability measure rewarding poses insensitive to variations in contact points and contact forces. Normally this has been neglected and only the balance of moment and forces has been taken into account. The manikin used in this work has 162 degrees of freedom and uses an exterior root. To describe operations and facilitate motion generation, the manikin is equipped with coordinate frames attached to end-effectors like hands and feet. The inverse kinematic problem is to find joint values such that the position and orientation of hands and feet matches certain target frames during an assembly motion. This inverse problem leads to an underdetermined system of equations since the number of joints exceeds the endeffectors’ constraints. Due to this redundancy there exist a set of solutions, allowing us to pick a solution that maximizes a scalar valued comfort function. Many objectives are included in the comfort function, for example in terms of joint angles, joint moments and solid objects’ distance to the manikin. The proposed stability measure complements the earlier balance criterion and is combined into the comfort function. By increasing the importance of this function the digital human model will reposition to a more stable pose. The digital human model will be tested on a set of challenging assembly operations taken from the automotive industry to show the effect of the stability measure.

Paper 49 Introducing a Formal High-Level Language for Instructing Automated Manikins MÅRDBERG P*†, CARLSON J.S†, BOHLIN R†, DELFS N†, GUSTAFSSON S†, KEYVANI A‡, HANSON L§↑ † Fraunhofer-Chalmers Research Centre for Industrial Mathematics, Gothenburg, Sweden ‡ University West, Department of Engineering Science, Trollhättan, Sweden § Department of Product and Production Development, Chalmers University of Technology, Gothenburg, Sweden ↑ Industrial Development, Scania CV, Sweden Digital Human Modeling programs are important tools in virtual manufacturing that allow simulation of manual assembly work, long before any physical product has been built. By investigating the logistics, ergonomics and the interaction among the workers at an early stage, it may be possible to reduce the cost of design changes, increase the quality and to decrease the ramp-up time of a manufacturing process. However, far from all assembly operations are simulated, even if all the necessary data is available. One reason is the tedious work required to setup and to define all the motions needed by a manikin to perform a simulation. In each simulation, the manikin must be adjusted into the desired posture and the user must ensure that balance is held and that it avoids collision with objects in the environment. Thus, even a small case may be time consuming to simulate. This shows that there is a need of an easier way of instructing the manikins. In this work we propose a new formal high-level language for controlling an automated manikin. The language instructions are structured by a grammar, which defines a hierarchical tree for the manikin instructions. The low-level instructions contains basic functionality for maneuvering the manikin, such as Move, Position and Grasp, and higher levels contain more abstract instructions such as Get and Assemble. Thus, the high-level instructions define sequences of other instructions, whereas a low-level instruction corresponds to a direct instruction of the manikin. The set of available instructions that the manikin may perform during a simulation depends on the current state of the manikin and the objects at the assembly station. Thus, properties of objects, such as grasping and mating points, also help define the set of available instructions for the manikin. The order in which the different parts in the assembly operations have to be connected may also be considered when constructing instruction sequences for the manikin. Furthermore, the instruction sequences may be formally verified to ensure that the manikin only performs valid instructions. The language have been implemented in the manikin simulation software IMMA and tested on elementary cases with relevance to industrial applications. The results show that fewer steps are needed to perform a realistic simulation when the language is used compared to manually instructing the manikin. Furthermore, it is also shown that the instructions generated by the language are formally correct.

Paper 50 Automatic Modeling of Customized Human Avatars using Contour Information D. VAN DEUN*†, V. VERHAERT†, T. WILLEMEN†, K. BUYS‡, B. HAEX†,, J. VANDER SLOTEN† † KU Leuven, Mechanical Engineering Department, Biomechanics Section, Belgium ‡ KU Leuven, Mechanical Engineering Department, Robotics, Belgium This paper focuses on automatic modeling of customized human avatars for use in sleep ergonomics. The main objective is to generate a human avatar having realistic body contours (both sagittal and coronal) since body contours interfacing the sleep system determine the ergonomic support. Information used in the modeling procedure includes coronal and sagittal body contours, combined with gender and Body Mass Index (BMI) input. A generic human avatar (MakeHuman base mesh) is used that can be personalized by morphing different regions of the body. The modeling framework consists of consecutive steps in which age, gender, BMI, body length, lateral contour and back contour are iteratively morphed. Body contours of 25 subjects were extracted using an optical measurement system. Results of this project show a mean Root Mean Squared (RMS) value between real and modeled back contour of 7.97 ± 1.15 mm. Lateral body contours of the trunk show a mean RMS value of 6.20 ± 1.84 mm. Comparison of the models with 3D full body scans showed a mean unsigned distance of 21.1 mm ± 24.6 mm. Mainly body extremities are responsible for large deviations due to differences in bode pose.

Paper 51 Dynamic Assessment of Finger Joint Loads Using Kinetic and Kinematic Measurements N. VIGNAIS, D.M. COCCHIARELLA, A.M. KOCIOLEK, P.J. KEIR* Department of Kinesiology, McMaster University, Hamilton, ON, Canada, L8S 4K1 Assessing finger joint loading is essential for the prevention of musculoskeletal disorders of the hand, wrist and forearm. Due to the technical and invasive nature of direct measurement, biomechanical modeling is necessary to evaluate finger joint forces. Most existing finger models have used maximum grip strength in order to quantify joint loads, although it is unlikely that these forces are routinely experienced in typical daily tasks. The purpose of this investigation was to assess finger joint forces continuously during submaximal tasks using an inverse dynamics approach. Eight participants performed a series of finger movements while pressing on a six-degree of freedom force transducer with the index finger. Participants were asked to maintain a 10 N vertical force with the distal phalanx of the finger during the movement while receiving visual feedback. Simultaneously, kinematic data were obtained using an optoelectronic motion capture system at 60 Hz. The index finger (digit 2) was instrumented with 20 reflective markers (4 mm in diameter). The data were used to model the metacarpals and phalanges based on an innovative segment definition technique. Forces were applied to the distal segment of the finger model to calculate joint reaction forces. The finger movements used in this study included isolated flexion/extension of the distal interphalangeal, proximal interphalangeal and metacarpophalangeal joints of the index finger. Results provided by the current model were promising in comparison with the literature. Using an inverse dynamics approach, joint reaction forces were determined continuously during each finger movement providing joint force profiles for each task which were normalized to the external fingertip force. While the current detailed methodology is limited to the laboratory, a refined and simplified model could be used for ergonomic analysis of manual tasks in the workplace. The current model will be integrated with musculotendinous structures to better assess musculoskeletal implications of finger movements and to further understand the link between fingertip loading and hand pathologies.

Paper 52 Hand gesture recognition for surgical control based on a depth image I. FAMAEY1, K. BUYS2, D. VAN DEUN1, T. DE LAET2, J. VANDER SLOTEN1,, J. DE SCHUTTER2 1Dep. of Mechanical Engineering, Biomechanics section, KU Leuven, Belgium 2Dep. of Mechanical Engineering, Robotics research group, KU Leuven, Belgium The introduction of hand gestures as an alternative to existing interface techniques could result in groundbreaking changes in health-care and in every day life. This research area is confronted with many challenges: variable illumination conditions, cluttered backgrounds, etc. Most recognition techniques for static hand gestures can be divided into three categories: methods based on low-level features, appearancebased approaches, and methods based on high-level features. This article presents three different methods for hand gesture recognition, one of each aforementioned category. The first method is based on information about the convexity defects of the 2D hand contour, the second one compares the 2D hand contour with a database, and the third one performs 3D template registration between the hand and pre- designed templates. Hand detection is performed using depth information from Microsoft Kinect. These methods are evaluated and compared. The recognition rates are 97%, 73.9%, and 50% respectively. The first method was currently chosen for the implementation of a gesture-based tool for the control of visualization displays in the operation room. The developed system can recognize eleven static gestures that are intuitive, easily distinguishable, and minimally tiring for the surgeon.

Paper 53 Surgeon pose detection with a depth camera during laparoscopic surgeries K. BUYS1, D. VAN DEUN2, B. VAN CLEYNENBREUGEL3, T. TUYTELAARS4,, J. DE SCHUTTER1 1Dep. of Mechanical Engineering, Robotics research group, KU Leuven, Belgium 2Dep. of Mechanical Engineering, Biomechanics section, KU Leuven, Belgium 3University Hospital Leuven, KU Leuven, Belgium 4Dep. of Electrical Engineering, KU Leuven, Belgium This paper presents a method to measure the surgeon’s body pose during a laparoscopic operation. The method handles the extreme situations of a darkened operation room in which people are all similarly clothed. This is achieved by enhancing the existing random decision forest approach with a new energy function in the region growing step. A switching behavior between the old energy function and the new one was defined. A proof-of-concept experiment provided successful early results.

Paper 54 Virtual data generation based on a human model for machine learning applications K. BUYS1, J. HAUQUIER2, C. CAGNIART3, T. TUYTELAARS4,, J. DE SCHUTTER1 1Dep. of Mechanical Engineering, KU Leuven, Belgium 2eJibe.net, Belgium 3Technical University of Munich, Germany 4Dep. of Electrical Engineering, KU Leuven, Belgium Most computer vision algorithms for human detection are grounded on a intensively data driven machine learning pipeline. Creating this pipeline is a time and computationally intensive step, so is collecting all the input data for this. Often manually annotated real life images are used as input data, this poses two drawbacks, first only a limited number of datasets are available, secondly this is time intensive or expensive to acquire. This paper presents a work flow to generate input data for human pose detection machine learning algorithms that is grounded on real human motions but is generated in a virtual environment with an accurate sensor model.

Paper 55 Larynx Region Surgical Education and Simulation Using A Three-dimensional virtual model Based on the Chinese Visible Human KAIJUN LIU, LIWEN TAN*, SHAOXIANG ZHANG* Department of Anatomy, Third Military Medical University, Chongqing, 400038, People’s Republic of China Anatomical knowledge of the larynx region is critical for understanding laryngeal disease and performing needed interventions. Virtual reality is a useful method for surgical education and simulation. However, medical students are confronted with few cadaver-based human anatomy courses because of the increasing student enrolments and the decreasing cadavers. The objective of this study was to build the virtual model and the larynx region based on Chinese visible human (CVH), which can be used for anatomical education and surgical training. The segmented cross-section slices of the larynx region were from Chinese visible human dataset. The laryngeal structures were precisely segmented manually as 2D images, then reconstructed and displayed as 3D images in the virtual reality Dextrobeam system. Using visualization and interaction with the Virtual Reality Maker Language (VRML) model, digital laryngeal anatomy instruction was constructed using HTML and JavaScript languages. The volume larynx models thus display an arbitrary section of the model and provide the virtual dissection function. This networked digital laryngeal anatomy teaching system can be read remotely, displayed locally, and manipulated interactively.

Paper 56 The validity of maximum force predictions based on single-joint torque measurements F. GUENZKOFER*, H. BUBB, K. BENGLER Institute of Ergonomics, Technische Universität München, Boltzmannstr.15, 85747 Garching, Germany Sophisticated digital human models (DHMs) allow for strength prediction in arbitrary postures. In most cases the maximum force is calculated based on joint- and posture-specific joint torques. This approach is also used in multi-joint cases, while it has not been assessed so far if this procedure is justified or could lead to biased results. Differences between single- and multi-joint cases should not be dismissed without further investigations based on bi-articular muscles, multifunctional muscles, co-contractions, and inhibition effects. This paper is dedicated to this open question and examines the differences between real maximum forces and maximum force predictions based on single-joint torque measurements. Therefore, 19 young males performed right arm pulling and pushing maximum force measurements in two different postures. Maximum elbow and shoulder joint torques were known for each posture and each subject. Consequently, the corresponding maximum forces could be predicted using an optimization study in AnyBody. A reconstruction of the multi-joint pulling and pushing trials in AnyBody helped to discuss arising differences in terms of muscle activations. The results show that maximum forces were on average overestimated by approximately 20%. This new knowledge should be taken into account in strength models of DHMs.

Paper 57 Method to analyze the body movement for the mirror view and test method to verify the simulated results A. ZAINDL*†, K. BENGLER*†, T. ERBER*†, A. ZIMMERMANN‡ † Technische Universität München, Institute of Ergonomics, Munich, Germany ‡ MAN Truck & Bus AG, Munich, Germany RAMSIS cognitive is a powerful tool to analyze sight requirements for the driver in the cab of a vehicle. Although RAMSIS offers an extensive amount of features, it does not at this point allow the simulation of driver’s mirror glancing behavior. Drivers move their body to increase their field of view and to see objects in the periphery. Especially in trucks, this behavior can be observed in situations where the trailer covers most of the field of view. For creating new systems for indirect vision like camera-monitor-systems, it’s necessary to know the dynamic ranges of the mirrors field of view. A method is developed in order to incorporate mirror glancing behavior in to the RAMSIS model. A road scenario containing a RAMSIS-manikin sitting in the truck cabin with mirrors and an object to observe is built with CATIA. The object is placed out of the normal field of view at a driver ́s point of interest (e.g. the end of a trailer by turning around). From this object a geometrical model showing the area where the object is visible in the mirror. Essentially, a cone is constructed from the object to the mirror and then reflected away from the mirror again. Besides the normal truck restrictions to calculate the position and posture of the manikin (feet to the pedals, H-Point to the seating field, etc.), a new restriction placing the eyepoint into this cone (from the mirror to the cabin) is made. Combined body and eye movements are simulated this way in the RAMSIS environment. To prove the results of the model an experiment is made in a real truck setting, where body and eye movements are measured. In this research, six Vicon cameras were mounted inside the truck cabin to capture the body movement of the driver. Markers were placed on to specific points of the participants. The eye movements were captured by the head mounted Dikablis-System. Vicon and Dikablis were synchronized so that the correlation between the two parameters (eye and body movement) can be made. A comparison between the modeled posture and the real behavior can be done afterwards. Drafting a generic experiment combining the body movement of a person and their eye movements can make the simulated behavior in human models more realistic. By knowing the real correlation between eye and body movements and to implement them in a digital human model, the understanding of drivers’ behavior is facilitated in the preliminary phase of creating a new cabin design.

Paper 58 Conditional Regression Model for Prediction of Anthropometric Variables ERIK BROLIN*†‡, LARS HANSON ‡§, DAN HÖGBERG †, ROLAND ÖRTENGREN ‡ † Virtual Systems Research Centre, University of Skövde, Skövde, Sweden ‡ Department of Product and Production Development, Chalmers University of Technology, Gothenburg, Sweden § Industrial Development, Scania CV, Södertälje, Sweden In digital human modelling (DHM) systems consideration of anthropometry is central. Important functionality in DHM tools is the regression model, i.e. the possibility to predict a complete set of measurements based on a number of defined independent anthropometric variables. The accuracy of a regression model is measured by how well the model predicts dependent variables based on independent variables, i.e. known key anthropometric measurements. In literature, existing regression models often use stature and/or body weight as independent variables in so-called flat regressions models which can produce estimations with large errors when there are low correlations between the independent and dependent variables. This paper suggests a conditional regression model that utilise all known measurements as independent variables when predicting each unknown dependent variable. The conditional regression model is compared to a flat regression model, using stature and weight as independent variables, and a hierarchical regression model that uses geometric and statistical relationships between body measurements to create specific linear regression equations in a hierarchical structure. The accuracy of the models is assessed by evaluating the coefficient of determination, R2 and the root-mean-square deviation (RMSD). The results from the study show that using a conditional regression model that makes use of all known variables to predict the values of unknown measurements is advantageous compared to the flat and hierarchical regression models. Both the conditional linear regression model and the hierarchical regression model have the advantage that when more measurements are included the models will give a better prediction of the unknown measurements compared to the flat regression model based on stature and weight. A conditional linear regression model has the additional advantage that any measurement can be used as independent variable. This gives the possibility to only include measurements that have a direct connection to the design dimensions being sought. Utilising the conditional regression model would create digital manikins with enhanced accuracy that would produce more realistic and accurate simulations and evaluations when using DHM tools for the design of products and workplaces.

Paper 60 Reproducing the Ankle Kinematics by Forwards Dynamics to Evaluate the Effect of Ankle-Foot Orthoses P. FERREIRA†‡, F. G. FLORES‡, M. SIEBLER*, P. FLORES†, A. KECSKEMÉTHY‡ † Mechanical Engineering Department, University of Minho, Campus de Azurém, 4800-058 Guimarães, Portugal ‡ Chair of Mechanics and Robotics, University Duisburg-Essen, Lotharstraße 1, 47057 Duisburg, Germany * Department of Neurology, Heinrich-Heine University Düsseldorf, Moorenstraße 5, P.O. Box 10 10 07, D40001, Düsseldorf, Germany Biomechanics is the scientific domain which deals with the study of biological systems, such as the human body, using physical concepts and mechanical engineering methodologies. It allows for the development of new medical devices and provides a quantitative analysis of the subject being studied. In the present work, the effect of an ankle-foot orthosis (AFO) was studied on a healthy male subject. For this purpose, a biomechanical MB 2D-model was developed in object-oriented multibody (MB) simulation library MOBILE. The model consists of a pair of legs, each one made up of 4 rigid links (thigh, shank, foot and toes), and an additional body to represent the upper-body, totalizing 9 rigid bodies constrained by 9 frictionless hinged joints. Three additional degrees of freedom (DOFs) were attached to the hip so the model can move freely in the plane. Thus, the model has a total of 12 DOFs. Kinematic data acquired in a gait laboratory, with and without AFOs, and was used as time functions to drive the joints. Moreover, a foot contact model was designed based on three Hunt-Crossley’s sphere-plane contacts. The measured ankle kinematics was successfully reproduced using forward dynamics (FD) principles for the stance phase period. In a first approach, barefoot kinematics was reproduced to define the foot properties by adjusting manually the foot parameters and fitting the ankle angle. The ankle moment obtained in the gait lab was used to power the ankle joint. Then, the AFO was added as a linear torsional spring element acting at the

ankle joint and the moment powering the ankle joint obtained in the free case was scaled down by a manual optimization until the simulated ankle angle fitted the measured one with orthoses. It was found that the AFO scales down the muscle moment developed at the ankle by 15% when an AFO with equivalent torsional spring stiffness of k !"# = 50.0  N. m/rad  is used.

Paper 62 Using a Game Engine as a interactive tool in design of productive situations W. SANTOS*†, P. TOLEDO†, D. BRAATZ†, M. A. B. COSTA†, N. L. MENEGON† †Department of Production Engineering (Federal University of São Carlos), São Carlos, Brazil The designs of productive situations process are characterized by technical and social dimensions that demand the involvement of actors in the explanation of the current problems and in making decisions about the future situation. This requires that the designers have instruments for communication and context of the situation in the design process for the different rationalities present in the project. The use of graphical computational tools, such as game engines, configured as an alternative tool due the graphic quality of this technology to create scenarios with a high degree of realism and integration of the human element in the design. This paper presents the development of a new plant for a Brazilian company of metal frames, conducted by a specific methodology for development of plant layout and modeling a virtual scenario using Game Engine CryEngine 2 of digital game Crysis (Crytek). This scenario was used in meetings with stakeholders during the design process to analyze the future plant. During the meeting the representatives of the company were able to navigate the virtual scenario by manipulating the Digital Human Model of Crysis. The results show the potential use of game engine as an alternative tool for simulating human, with high quality graphics, to design productive situations, not restricted the presentation of the final results of the project.

Paper 63 A suite of multivariate test cases for anthropometry-based design G. NADADUR, M.B. PARKINSON OPEN Design Lab, The Pennsylvania State University This work presents a test suite of three design problems involving spatial fit-based interactions between users and the product. The test suite allows for multivariate analyses of fit-based accommodation using anthropometric data. This process of calculating and comparing accommodation percentages is proposed as a method for a three main purposes: assessing the relative merits of different anthropometry synthesis techniques, verifying the accuracy of synthesized datasets of body dimensions, and comparing datasets of multiple populations of interest. The test suite is demonstrated in the context of an actual and synthesized target user population which are compared for an assessment of the multidimensional fidelity of an anthropometry synthesis method.

Paper 65 Anthropometric Change in the U.S. Marine Corps: Implications for Design C. GORDON*†, B. BRADTMILLER‡, S. PAQUETTE†, C. BLACKWELL†,, B. CORNER† † U.S. Army Natick Soldier Center, Natick, MA ‡ Anthrotech, Yellow Springs, OH A major anthropometric survey of the US Marine Corps (USMC) was conducted in 1966 when 70 standard dimensions were measured on 2,008 active duty male Marines (White & Churchill, 1977). That database was subsequently used to design everything from uniforms and protective equipment to aircraft and amphibious vehicles. Later, concerns were raised that the 1966 database was no longer representative of the current U.S. Marines Corps. In 1994, scientists from the U.S. Army Natick Soldier Center measured 12 dimensions on a carefully selected sample of 493 male and 470 female Marines to assess anthropometric change since 1966. The new measurements were also designed to validate a statistically matched database drawn from the U.S. Army 1988 anthropometric survey (ANSUR) which would serve as a design database for the U.S. Marines until a full anthropometric survey could be conducted (Donelson & Gordon 1996). That full survey was subsequently conducted in 2010, when 1,301 Marine males and 620 Marine females were

measured for 94 clothing, workstation and modeling measurements. In addition, 3D scans of the whole body, the head and the foot were taken. Of twelve comparably measured dimensions, all were statistically significantly different between 1966 males and 2010 males. Of twelve comparably measured dimensions between 1994 males and 2010 males, seven were significantly different. The comparisons between 1994 females and 2010 females showed that six of twelve comparably measured dimensions were statistically different. All statistical tests were made with the Bonferroni correction for multiple comparisons. USMC heights have not changed much in the last two decades. However, body weight and related body circumferences have increased significantly in both males and females, suggesting that a review of USMC clothing & equipment sizing and design criteria is needed.

Paper 66 A Comparative Study of Human Motion Capture and Computational Analysis Tools SEUNG-KOOK JUN†, XIAOBO ZHOU†, DANIEL K. RAMSEY‡, VENKAT N. KROVI† † Mechanical & Aerospace. Engineering, SUNY at Buffalo, Buffalo, NY 14260 USA ‡ Exercise & Nutrition Science, SUNY at Buffalo, Buffalo, NY 14260 USA Human motion-capture and computational analysis tools have played a significant role in a variety of product-design and ergonomics settings for over a quarter-century. In moving beyond traditional kinematic (and its dual-static) settings, advances in biomechanics and multibody dynamics have led up to computational analysis tools that can provide significant insights into the functional performance. Such tools now offer the ability to perform numerous what-if type analyses to help virtually-evaluate scenarios, thereby providing enormous cost- and time-savings. However, there exist significant differences in the capabilities and ease-of-use between these tools, necessitating a careful evaluation. Hence, in this paper, we perform comparative analysis of motion data from two alternate human motion-capture systems (Vicon vs Kinect) processed using state-of-the-art computational-analysis systems (AnyBody Modeling System/Visual-3D). The quantitative evaluation of a clinically relevant task (squatting) facilitates an objective evaluation of functional performance including effects of motion capture fidelity (from various sources) and the role of pre- and post-processing (calibration, latent dynamics estimation).

Paper 67 Considerations for aggregation of motion-captured files in structured databases for DHM applications A.KEYVANI*†§, D. LÄMKULL‡, G.BOLMSJÖ§, R.ÖRTENGREN† † Product and Production Development Department, Chalmers University of Technology, Gothenburg, Sweden ‡ Virtual Methods & IT, Volvo Car Corporation, Dept. 81620/PVÖS32, Gothenburg, Sweden § Department of Engineering Science, University West, Trollhättan, Sweden One way of enhancing motion simulation in digital human modeling (DHM) tools is to use data-driven methods which are based on real motion data. In spite of the availability of motion captured datasets which are offered for different purposes by commercial and research institutes, aggregation and integration of these motions in a unified and structured database system is not straight forward. Lack of this integration, limits the availability of existing data and causes DHM tools not to be able to use the data efficiently. Also for the researchers, comparison and analysis of data get very hard if not impossible. When searching for a specific motion pattern, it is optimal if the stored data in the database can be directly compared, analyzed and then retrieved if necessary. This study highlights several sources of incompatibility among motion capture files which shall be considered when implementing a comprehensive data management system for manipulating motion captured data. Subsequently, these incompatibilities are analyzed in more detail and necessary considerations and possible solutions are proposed in order to overcome the integration obstacles.

Paper 68 Performance Evaluation of Glove Design through 3-D Motion Capture A. PETROVA, S. PEKSOZ, M. BALASUBRAMANIAN,, K.M. ROBINETTE Oklahoma State University, Stillwater, OK, U.S.A. The purpose of this study was to explore the possibility of using 3-D motion capture to evaluate and compare firefighter glove designs. Three gloves, which differed in their thumb design, were subject to range-of-motion thumb abduction/adduction test. Thumb movement of bare and gloved hand was recorded for 3 subjects using passive optical 3-D motion capture system. Angle peak value of motion angle was found to be significantly different (at the .05 level) between barehand and each gloved condition. Normalized angle peak values for each pair of glove designs were found to be significantly different (at the .05 level with Tukey HSD test), demonstrating the ability of the proposed motion capture test to differentiate between glove designs.

Paper 70 Joint angles calculation through augmented reality F. C. H. PASTURA*1,2, G. L. DE ALMEIDA1,3, G. CUNHA1,4 1Coordenação dos Programas de Pós-Graduação em Engenharia (COPPE/UFRJ), Rio de Janeiro, Brazil 2Instituto Nacional de Tecnologia (INT/MCTI), Rio de Janeiro, Brazil 3Pontifícia Universidade Católica do Rio de Janeiro (PUC/Rio), Rio de Janeiro, Brazil 4Laboratório de Métodos Computacionais em Engenharia (LAMCE/COPPE/UFRJ), Rio de Janeiro, Brazil The aim of this study was to develop a computational system that assists in the analysis of working postures through joint angles calculation of the human body. This is done through tracking markers positioned on specific anatomical points of the body. Each marker is positioned on anatomical landmarks that define the body segments of a biomechanical model in the sagittal plane. By means of computer vision using a webcam, the Processing programming language and the NyARToolkit and the GSVideo libraries, the markers are recognized in sequence. From markers recognition the system recognizes the top left corner of each marker, draws line segments (vectors) between these vertices and calculates the joint angles defined by two vectors that have a vertex in common. The system also executes postural analysis taking into account postural angles defined in the REBA (Rapid Entire Body Assessment) tool. The main aspect of this study is that with a low cost infrastructure, comprised of a webcam and paper printed markers, it was possible to develop a real time joint angles calculation and assessment tool.

Paper 71 A Collision Avoidance Algorithm for Optimization-based Human Motion Prediction Based on Perceived Risk BRADLEY HOWARD, JAMES YANG* Human-Centric Design Research Lab Department of Mechanical Engineering, Texas Tech University Lubbock, TX 79409, USA Humans are able to plan movements around obstacles in a very efficient manner that almost seems effortless to the person. In digital human modeling (DHM), specifically optimization-based posture and motion prediction, the ability to computationally determine joint angle histories that effectively create motion around obstacles has been developed and employed widely. However, these methods are largely based on detecting spatial interference between objects and finding the optimal path with respect to a biomechanical cost function in an optimization problem to avoid the spatial interference. Although these methods result in a feasible path, they often times do not result in a realistic path, due to the lack of modeling for the cognitive bias that exists in real humans’ path planning in order to minimize a perceived risk. This cognitive bias is the tendency of people to grossly overestimate the low-probability of a collision between their hand and the obstacle. In turn, it results in serious sub-optimal performance, which has been documented by experimental studies in the literature. Therefore, a question arises: is it possible to account for the intentional sub-optimal human performance in a prediction model that optimizes performance? The world of cognitive science has been able to show that many biological tasks can be formulated in statistical terms, including path planning, which can account for cognitive bias. The perceived risk is formulated in terms of an economic loss or gain. This paper formulates the optimization based-motion prediction for a

simple reach task around an obstacle in terms of statistical decision theory, where expected economic gains are optimized as well as biomechanical cost. The prediction models are developed in Matlab and solved using the optimization toolbox. The models are compared against motion capture data collected by a Motion Analysis System in which 6 subjects completed the simple reach task. Each subject completed ten trials in order to determine any effects due to learning as the subject became more experienced with the task.

Paper 72 Ergonomic Risk Assessment of a Manikin’s Wrist Movements - a Test Study in Manual Assembly A. KEYVANI*†, D.HÖGBERG‡, L.HANSON§, D.LÄMKULL√, N.DELFS!!, I.M.RHEN‡,, R.ÖRTENGREN† † Product and Production Development Dep., Chalmers University of Technology, Gothenburg, Sweden ‡ Virtual Systems Research Centre, University of Skövde, Box 408, SE-541 28 Skövde, Sweden § Industrial Development, Scania CV, Scania AB (publ), SE-151 87 Södertälje, Sweden √ Virtual Methods & IT, Volvo Car Corporation, Dept. 81620/PVÖS32, SE-405 31 Gothenburg, Sweden !! Fraunhofer Chalmers Research Centre, Gothenburg, Sweden Use of digital human modeling (DHM) tools enables early assessment of ergonomic risks in the production development process. This early risk assessment can indicate needs for preventive actions in order to decrease risks for work-related musculoskeletal disorders. A method for the assessment of physical workload has been developed, for the analysis of wrist joint data produced by a DHM tool. The method is tested in a simulation model of an actual manual assembly station in industry where it is assumed that physical workload problems exist. The results show that the method can be successfully applied on an industrial case. In addition, presumptive risks and possible diagnoses are predicted based on the similarity of the simulated task’s motions with motions from other known work class profiles stored in a database based on epidemiological research.

Paper 73 Developing Parametric Human Finite Element Models for Injury Assessment Focusing on Various Vulnerable Populations J. HU*†, X. SHI†‡, K.F. KLEIN†, Z. LI†♯, J.D. RUPP†,, M.P. REED† † University of Michigan Transportation Research Institute, Ann Arbor, MI, USA ‡ State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body, Hunan University, Changsha, Hunan, China ♯ State Key Laboratory of Automotive Safety and Energy, Tsinghua University, Beijing, China Children, small female, elderly, and obese occupants are vulnerable populations and may sustain increased risk of death and serious injury in motor-vehicle crashes compared with mid-size young male occupants. Unfortunately, current injury assessment tools do not account for immature and growing body structures for children, nor the body shape and composition changes that are thought make female/aging/obese adults more vulnerable. The greatest opportunity to broaden crash protection to encompass all vehicle occupants lies in improved, parametric human FE models that can represent a wide range of human attributes. In this study, a novel approach to develop such models is proposed. The method includes 1) developing and integrating statistical skeleton and human body surface contour models based on medical image, body scan, and/or sitting posture data using a combination of principal component analysis and regression model, and 2) linking the statistical geometry model to a baseline human FE model through an automated mesh morphing algorithm using radial basis functions, so that the FE model can represent population variability. Examples of using this approach to develop parametric pediatric head model, adult thorax and lower extremity models, and whole-body human models representing various populations were represented. The method proposed in this study enables future safety design optimizations targeting at various vulnerable populations that cannot be considered with current injury assessment tools.

Paper 74 Successful Collaboration to Expand the Use of Digital Human Models A. STEPHENS†1, D. WEGNER†2, V. RACCO3, C. RAE†4 †USCAR - 1000 Town Center Drive, Suite 300, Southfield, MI 48075 1Ford Motor Company 2General Motors 3Lear Corporation 4Sandalwood Engineering & Ergonomics Many companies today are using Digital Human Models (DHMs) within their engineering process to help make critical decisions about product and process design. Digital human models are used to substitute for real people when evaluating the virtual workplace. This transformation, away from reliance on physical prototypes to the virtual environment, enables much earlier, faster and broader ergonomic evaluation at every design stage. However, this shift also represents challenges for the technology because the better the model, the better the output, the more accurate the decisions and the more complex it can be to use. To address this challenge, the users of the Siemens PLM Jack human simulation tool within the United States Council for Automotive Research LLC (USCAR), the collaborative automotive technology company comprised of Chrysler Group, Ford and General Motors, decided to pool resources to accelerate the evolution of digital human models. As users we collaborated on defining future needs to improve engineering decisions, voted on the highest shared priorities, and scoped projects in collaboration with Siemens. The team had years of experience working with the University of Michigan HUMOSIM Program, which produced findings that would bring value to common DHM use if made available in production tools. This collaborative effort is now in the 2nd year. The focus of the first year project was to prove out the collaboration process while making progress towards the implementation of the HUMOSIM research findings into the Siemens DHM. Later, additional workshops were held with users from a broad spectrum of industries, revealing additional common digital human modeling enhancement needs. Projects were identified to accelerate feature development and incorporate the latest research findings in the following areas: the advancement of human performance tools, posture prediction, motion capture flexibility and simulation creation. These projects are targeted to be completed in June 2013. The group is continuing to collaborate to identify the additional enhancement needs to improve the ergonomic analysis capabilities and overall performance. This paper reviews our process of working as a Digital Human Modeling User team to drive model enhancement. It also explains how the current members use DHMs, and the current needs being discussed surrounding this collaboration effort. It will also explain how we use DHM and the needs of the DHM users for this project and future projects.

Paper 76 Lateral Regional Impact Validation of a Full Body Finite Element Model for Crash Injury Prediction N.A. VAVALLE†‡, D.P. MORENO†‡, A.R. HAYES†‡, J.D. STITZEL†‡,, F.S. GAYZIK*†‡ † Wake Forest University School of Medicine, Medical Center Blvd, Winston-Salem, NC 27157 ‡ Virginia Tech – Wake Forest University Center for Injury Biomechanics, 575 N. Patterson Avenue Suite 120, Winston-Salem, NC 27101 With increasing computing power, finite element models (FEMs) of the human body are becoming more anatomically detailed, and have the potential to serve as powerful tools for injury mitigation research. These models must be validated in specific loading modes to quantify their accuracy. The purpose of this study is to examine lateral hub impact validations of a human body finite element model and to examine the robustness of the model a sensitivity study. A full body FEM of the 50th percentile male (M50) was developed from the medical images of a living subject. This study focuses on the validation of M50 in a shoulder impact and a pelvic impact. The shoulder impact applied a cylindrical impactor, weighing 23.4 kg, at a velocity of 4.5 m/s while the pelvic impact used a square- faced block impactor, weighing 16 kg, at a velocity of 10.0 m/s. Transient force traces from model outputs were compared to literature data using Correlation and Analysis (CORA) software. Pelvic fractures were predicted in the model via element deletion at a strain of 0.03 and compared to experimental outcomes. A sensitivity study using the pelvic impact was also performed to determine model robustness. The model showed a good match in the pelvic impact with CORA values of 0.97 (Corridor), 0.98 (Shape), 0.48 (Size), and 1.0 (Phase). CORA values in the shoulder impact were: 0.51 (Corridor), 0.85 (Shape), 0.62

(Size), and 0.09 (Phase). The model output had peak pelvis and shoulder forces of 13.0 kN and 4.6 kN compared to 10.7±2.8 kN and 2.4±0.8 kN in the experiments, respectively. In the pelvic impact the model predicted an ilio-pubic ramus fracture whereas experiments showed a range from no fractures to four fractures, with rami fractures also observed. Results from the sensitivity study indicate that perturbations to the impactor positioning changed peak force by no more than 3% and deflection by no more than 6%. The results of the shoulder impact show that while the energy absorbed values matched well, the peak force in the model was larger than the experimental average. In the pelvis case, both the energy absorbed and the peak force were larger in the model than in the experiments. The sensitivity study determined good model robustness since small input perturbations did not change the outputs greatly. This study represents one component of a range of validation efforts currently focused on M50. This model will become an important tool used by engineers in studying and advancing occupant safety.

Paper 77 Integrating motion capture data into knowledge based ingress motion simulation tool H.-J. WIRSCHING*†, R. BICHLER‡ † Human Solutions, Europaallee 10, 67657 Kaiserslautern, Germany ‡ BMW Group, Knorrstraße 147, 80788 Munich, Germany The present feasibility study investigates the integration of available ingress motion capture data into a knowledge based motion prediction tool to check the productive use in automotive design. The motion capture data were taken on the ingress motions of seven male and female subjects into three different vehicle ingress concepts provided by an adjustable mock-up. The measured motions were modeled in a 3D digital human model system using estimated joint positions and analyzed with respect to characteristic strategies, phases and key events. The analysis results were integrated into a database structure, which is used by a two level motion prediction algorithm. For a given 3D manikin and vehicle design this method extracts the best fitting motion from the database in the first stage. This referential motion is adjusted in the second stage to the vehicle configuration. Hereby the motion is adapted to geometrical constraints at the key frames and reproduced in-between according to the motion acceleration. The constraints contain contact as well as collision avoidance criteria. The study finally evaluated the entire process from motion capture processing to the prediction of motions in new environments from a practicable and accuracy point of view. As a result of the study the knowledge based simulation approach proofs to work properly in principle, especially the calculation performance is satisfying. The motion simulation control is flexible, but requires a deep knowledge of the algorithms behind. The quality of the motion capture modeling is crucial for the simulation quality and modeling decays cannot be compensated in the simulation anymore. The identification of similar strategies and key events separates the motion pool in different groups, which should contain sets of representative motions. In some cases the control of the key frames are not sufficient to predict proper motions and an additional control between key frames is required. Finally the approach cannot provide continuous results for continuous input parameter variations. The investigated motion simulation approach is very promising with respect to quality and usability in automotive ergonomics. Nevertheless there are some conceptual shortcomings which need to be analyzed and overcome. In addition the user interface has to be simplified and stabilized in order to bring benefits to a wider range of design engineers.

Paper 79 An Investigation of Hand Force Distribution, Hand Posture and Surface Orientation R. FIGUEROA†, T. ARMSTRONG † † Department of Industrial and Operations Engineering, University of Michigan, Ann Arbor, MI This work aims to obtain information and develop models for predicting hand postures that can be incorporated into computerized human manikins. This work is concerned with how finger, wrist, elbow and shoulder postures are affected depending the orientation and magnitude of hand force exerted. The hypotheses included: 1) Object orientation affects upper extremities postures and maximum hand force exertions; 2) Higher force exertions have similar postures per orientation. Shoulder-to-Elbow, Elbow-toWrist, Wrist-to-MCP, MCP-to-DP3 vectors were used to determine posture for each condition. Participants performed three exertions (100%, 30% and 10% MVC) perpendicular to an aluminum plate in 45 ̊, 0 ̊, 45 ̊ and 90 ̊ pitch at elbow height. Results show that posture changes depending the force exerted and the

threshold for significant changes is at force levels below 50%MVC. Comparing exertions over 50%MVC with exertions of 50%MVC or lower, elbow and wrist angles were significantly greater, and forearm was more perpendicular to the surface. As the force required decreases, the distance between hand and shoulder increases. If lower levels of force are required, a higher clearance is needed to decrease the possibility of acquiring awkward postures, one of the common physical hazards leading to leads to develop MSDs. Project deliverables (hand position, posture, resultant force, contact force distribution, joint moments) can be used to predict upper extremity postures as function of magnitude and orientation of force exerted, these factors can be implemented into computerized human manikins.

Paper 80 Anthropometric data analysis for morphological features in Korean YOUNG-SUK LEE † Chonnam National University, Dep.of Clothing Science ‡ Yongbongdong, 300, Gwangju, Korea This paper deals with secular trend of height spanning 1979 to 2010 and also morphological growth patterns for Korean. It also investigates the comparison of the growth pattern between male and female based on the anthropometric data samples of KATS taken from the 1979 to 2010 surveys. The great rates of change in height occurred during 1979 - 1997 with 2-3 cm increase in both sexes at each surveys. After 1997 the change rate of height while increasing is remained within less than 1cm till 2010 survey. This pattern is consistent with recent Korean cultural emphasis on health and physical fitness according to the economic development. The causes of the observed trends, insofar as they have been identified, are related to cultural processes.

Paper 81 Statistical surface shape adjustment for a posable human hand model N. MIYATA*†, D. NAKAMURA‡, Y. ENDO†,, Y. MAEDA‡ † Digital Human Research Center (AIST), Tokyo, JAPAN ‡Yokohama National University, Kanagawa, JAPAN This paper proposes a method to reproduce a human hand surface shape precisely by statistically adjusting the shape by the skeletal subspace deformation (SSD) method. Plaster molds of the hand in variable postures were fabricated for one subject and their 3D shapes were obtained by computed tomography (CT) scanning. An SSD model was prepared for the subject to compute difference from the actual human hand. Landmarks were manually selected from each plaster mold model and from the SSD model. Each landmark position on a plaster mold model was compared with corresponding one on an SSD model posed to resemble the plaster mold model. The collected difference from all the plaster molds was analyzed with postures to estimate adjustment vector for landmarks on an SSD model in arbitrary posture. The rest of the vertices were corrected using adjustment vectors of the three nearest landmarks.

Paper 82 Quantifying Hand Force Distribution Relative to Hand Posture Selection W. ZHOU*†, T. ARMSTRONG†, D. WEGNER‡, M. REED† † Center for Ergonomics, University of Michigan, Ann Arbor, MI ‡ General Motors, Warren, MI Digital human models that depict human forms and compute biomechanical loads are important design tools. One of the limitations of these models as design tools is that they required users to either observe workers or users to guess the necessary posture inputs. Previous studies have shown that selection of posture is related to the relative load of upper limb joints and the coupling between the hand and work objects. The probability of choosing overhand posture to grasp cylinders reduces from 85% to 0% while the one of underhand posture increases from 15% to 100% as cylinder weight increases from 3.3 N to 53.3 N. The objective of this study was to provide a biomechanical explanation for hand posture selection for grasping work object. One male subject participated in a pilot study in which he lifted and placed a 70 mm diameter cylinder with a weight of 43.1 N using four hand postures (overhand, underhand, pinch, and hook grip). Hand force distribution was measured by Tekscan pressure mapping system. Sample hand force

distribution from the subject shows that hand force distributions vary by hand posture. Hand forces are highly concentrated at thumb tip (distal phalanx) and finger tips for overhand and pinch posture. The thumb tip and finger tip forces are reduced for underhand and hook grip posture. The tip forces appear to be the largest for pinch, then for overhand grasp, then for underhand grasp, and the lowest for the hook grip. Selection of posture appears to be related to the preference of reducing thumb and finger tip forces and relative joint loads.

Paper 83 A FATIGABLE MUSCULOSKELETAL MODEL FOR PREDICTION OF NECK MANEUVERING LOADINGS ON AVIATORS X. ZHOU*†, P. WHITLEY†,, A. PRZEKWAS† † CFD Research Corporation, 215 Wynn Drive, Huntsville, AL, 35758 In this paper we study cervical musculoskeletal loads and fatigue on military aviators during aerial combat maneuvering (ACM). A whole body articulated multi-body model with detailed neck musculature was utilized to predict the neck loads and muscle fatigue of a fighter pilot during high-G maneuvering. Two flight postures, look-ahead and check-6, were investigated on their effects on neck loadings for a duration of 300 seconds. To account for fatigue and decrease of muscle force capacity, a new dynamic muscle fatigue model based on the fatigue-rest-recovery mechanism was incorporated to predict fatigue of muscles responing to dynamic loading conditions. For both postures, the joint dynamics, muscle forces and fatigue levels, representing the neck biodynamic responses to the applied aircraft acceleration, were obtained and compared. For model calibration and validation purposes, digital, multiaxis in-flight data and measurments are needed in the future.

Paper 84 Development of a biomechanical diagnostic camera for consulting of occupational low back pain using a tablet PC Y. TANAKA†, T. KAWANO*‡, Y. FUKUI‡ † Graduate School of Engineering, Setsunan University, Japan ‡ Faculty of Science and Engineering, Setsunan University, Japan The purpose of this study is to develop a biomechanical diagnostic camera allowing to predict whether working postures induce low back pain (LBP) and to consult how to reduce low back load using a tablet PC; Xperia Tablet S. We focus on the working postures in the material handling or care working. Using the tablet PC a software application has been developed to take a picture of the worker’s lateral posture and draw a 2D human body link model by touching articulation nodes of the body image on the touch screen. The software finally estimates compression forces at the lumbar intervertebral disk L4/L5 solving equations derived from a static kinematical model. Some workers feel LBP if the compression force exceeds 1800∼6000N (Jäger’s criteria). Almost workers feel it if the force exceeds 6400N(NIOSH criteria). The biomechanical diagnostic camera developed in this study predicts the LBP using the criteria and indicates advice to improve the working posture. The camera is available for use on-site and improvement of the working postures and working environment.

Paper 85 Encumbered Anthropometry Protocol Development M.L.H. JONES†, P.S.E. FARRELL†, A. KEEFE† † Defence Research and Development Canada, Toronto, Ontario Anthropometry data are used to facilitate the design and sizing of personal protective clothing and equipment systems (PPE). Standard anthropometric measures are based on semi-nude data while clothing and equipment represent an additional “bulk” that must be accounted for when designing workspaces and equipment. The objectives of this study are to: (1) evaluate two different landmark paradigms, and (2) evaluate the sensitivity of a bulk metric to differentiate encumbered load conditions. Encumbered anthropometric measurement data were obtained on a convenient sample of 26 male CF operators in order to identify the dimensional changes in body size that occur as encumbered conditions were donned. Two configurations of encumbrance (PPE) were evaluated. Bulk profile was characterized by two unique

landmark paradigms: (1) Normalized to Anthropometry, and (2) Maximal Bulk. The Normalized to Anthropometry paradigm was based on reference height associated with an anatomical landmark, thereby transferring anthropometrically defined landmarks to encumbered conditions. The Maximal Bulk paradigm identified maximal bulk /girth within a specified region of the PPE at which the cross-section and corresponding reference height was defined. Unencumbered (semi-nude) baseline equivalent measures were also acquired for each cross-sectional height. Linear (1D) breadth, depth and circumferential measures were taken at each cross-section/reference height. These measurements defined an ellipse and cross-sectional 2D area. Difference in both the 1D and 2D measures of each cross-section between the unencumbered and encumbered conditions represents bulk at that reference height, thus generating a bulk (size and shape) profile. This study provided a detailed methodology of encumbered measurement. Results indicate that a combination of landmark paradigms, and calculation of bulk metrics, specifically the transformation from 1D to 2D measures, provide an effective method to parameterize encumbrance. Accurate representation of encumbrance can be used to represent body size and shape (spatial claim) and integrate into digital human modeling (DHM) software to visualize and evaluate the effects of clothing and equipment systems.

Paper 86 Variation in Real World Crash Simulation Injury Metrics Given Crash Parameter Perturbations K.A. DANELSON†‡, A.J. GOLMAN†‡, C.M.WEAVER†‡,, J.D. STITZEL *†‡ † Virginia Tech/Wake Forest School of Biomedical Engineering and Sciences ‡ Wake Forest University School of Medicine Real world crash simulations of Motor Vehicle Crashes using human body finite element models (HBMs) have the potential to elucidate injury mechanism and predict injury risks. One limitation of these simulations is the inherent uncertainty of real world crash measurements compared to laboratory measurements. The purpose of this study was to develop a method to assess a potential range of response parameters of a HBM given slight perturbations of crash parameters. The Total HUman Model for Safety (THUMS) version 4 was used in the simulation of a Crash Injury Research and Engineering Network realworld crash. Vehicle models of the Ford Taurus and Toyota RAV4 from the National Crash Analysis Center Finite Element Model database were used as the struck and striking vehicles, respectively. The case was simulated and iteratively improved until the simulation crush profile was optimized based on the recorded case data. Next, crash parameters were varied over a range of values about the final case values for a onefactor-at-a-time variation study: lateral delta-V (0-3 m/s), longitudinal delta-V (13-17 m/s), crush location (239-289), and rotational velocity (3-5 rad/s). Injury metrics (Thoracic Trauma Index, half deflection, and pelvis force) were calculated for each variation simulation and analyzed to determine the effect of variation on injury metric outcome through confidence intervals and a sensitivity analysis. The variation in the thoracic injury metrics was low indicative of small changes in the crash parameters resulting in small changes in chest injury risk. In contrast, the pelvis risk of injury changed more with changes in the crash parameters. These differences in response were attributed to the initial location of the injury metrics on the corresponding injury risk curves and the model sensitivity to these changes. The baseline model already had a high risk of thoracic injury and rib fractures in the real world case occupant. Therefore, slight changes in the crash parameters did little to increase or reduce the already high risk. The pelvis load resulted in a low risk of injury and no pelvis injury in the case occupant. Slight changes in the model inputs resulted in large variations in pelvis force with the largest coefficient of variation, over 27%, with X-location change. Given these findings, this variation study gave some insight into the relationship between crash parameters and injury metrics. Future research will enhance the variation study design to vary more parameters, including interior vehicle and occupant variables.

Paper 88 2012 Canadian Forces Anthropometric Survey A. KEEFE*†, H. ANGEL‡, B. MANGAN‡ †Defence Research and Development Canada, Toronto, Ontario ‡ Humansystems Inc®, Guelph, Ontario This paper presents the demographic, anthropometric, and methodological results of the 2012 Canadian Forces Anthropometric Survey (CFAS). A total of 2206 participants were measured at 14 Canadian Forces Bases and Wings across the country. Of these 2206 personnel, 847 were from the Canadian Army (CA), 877

were from the Royal Canadian Air Force (RCAF), and 482 were from the Royal Canadian Navy (RCN). Three-hundred fifteen of the measured personnel were female, while 1891 were male. Strategic sampling goals were established prior to data collection; stratification was by age, gender, linguistic profile, and Military Occupational Structure (MOSID). The intent was to maximize the number of personnel measured who were employed in “operational” trades as these individuals would be more likely than others to wear and use dedicated equipment or require specialized operator/passenger accommodation considerations according to their job role. Each participant was measured using traditional tools and methodologies as well as using four different digital data collection tools: Vitus XXL Whole Body Scanner, Vitus aHead Scanner, and Elinvision Footin3D Scanner. Six International Organization for Standardization (ISO)defined postures were mandated for the whole body scanner, 2 of which were in the seated position; 2 hand scans, 1 head scan, and 1 foot scan were taken. In total, ten 3-D laser scanned images were collected for each participant in this survey. This represents more than 22,000 digital human models of current CF personnel. The 2012 CFAS is the largest anthropometric survey conducted by the Canadian Forces (n=2206) and the first to fully incorporated Canada’s three military branches (Army, Air Force, and Navy). These data will benefit and inform the development of systems requirements and specification to support effective defence acquisition and ensure optimal fit, accommodation, safety and performance for clothing, equipment and workspace design.

Paper 89 A Fatigue-Included Prediction Model for Drivers’ Seat Comfort during Prolonged Driving 1XIN,

TAO; 1BO, CHENG; 1WENJUN WANG; 2CHAOYANG, CHEN 1State Key Laboratory of Automotive Safety and Energy, Tsinghua University, P. R. China 2Department of Biomedical Engineering, Wayne State University, United States The aim of the study is to develop a prediction model for the evaluation of drivers’ seat comfort located at different parts of the bodies (arm, neck, upper back, lower back, buttock, thigh and right shank). The evolution of local muscle fatigue makes important effects on drivers’ dynamic seat comfort during longterm driving. As the traditional investigation methods, specific questionnaires have long been used for the assessment of seat comfort; business software like JACK and RAMSIS was brought in vehicle design to conduct the ergonomic simulation. However, subjective questionnaires are not reliable enough and none of commercial software has the capacity of analyzing dynamic seat comfort with the consideration of local muscle fatigue. In this study a prediction model of drivers’ dynamic seat comfort including fatigue will be proposed with whole body comfort and local comfort predicted respectively. Drivers’ seat comfort during prolonged driving was investigated in a simulated driving experiment with eight male subjects. The experiments were composed of two parts. First, subjects took the maximum voluntary contraction (MVC) tests before and after long-term driving. The maximal value of tension was collected by force sensor K3D120 and NI data acquisition module. Second, each subject was required to take a 120-minute continuous driving test with desirable seat positions on a 6-DOF driving simulation platform. A Borg-scaled discomfort questionnaire was conducted to record the local subjective ratings every 5-10 minutes randomly. In both parts of the experiment, 14-channel SEMG signals were recorded continuously by Biopack MP150 system. Several parameters of SEMG signals such as root mean square (RMS), median frequency (MF), mean power frequency (MPF) were calculated by Acqknowledge 4.2 software to analyze the evolution of drivers’ local muscle fatigue. A prediction model based on support vector regression and genetic algorithm for dynamic whole body comfort and local comfort was established by Matlab software. There were significant differences in normalized RMS values between the MVC tests before and after prolonged driving, which confirmed the onset of local muscle fatigue. MF and MPF of the SEMG signals from left and right erector spinae muscle were found to grow slowly with fluctuations, which was opposite to the skeletal-muscle-fatigue manifestations reported in sports biomechanics research. RMS normalized by MVC data turned to increase slowly, which kept in line with previous study. A multi-output support vector regression with parametric optimization by genetic algorithm was used to build the prediction model of drivers’ seat comfort with minimum mean squared error (MSE) for whole body comfort and local comfort. This study contributes to a practical way for the fatigue-included evaluation of drivers’ seat comfort during prolonged driving. With large database preparation in the future, it is promising to be used for the systematic ergonomic assessment considering muscle fatigue.

Paper 90 Accuracy of reconstructed body shape from body dimensions through principal component analysis of shape variation M. KOUCHI*†, S. YAMAZAKI†, M. MOCHIMARU† †Digital Human Research Center (National Institute of Advanced Industrial Science and Technology) Tokyo, Japan Body shape generation based on principal component analysis (PCA) of homologous models is recently used for apparel applications. Combined with a large-scale body scan data collected in retail shops, this technique can be used for various kinds of services in apparel fields. The accuracy of measurements of generated digital human models can be critical when individual customers are involved. We evaluated the errors in bust, waist, and hip circumferences of the body shape generated through PCA based on body scan data obtained in a retail shop, and examined the effects of the quality of scan data and software performance. Whole body scans, automatically calculated landmark positions, and 33 scan-extracted measurements from 337 adult females were used. Three homologous models were created for each scan: a model obtained by template fitting (original model), a model calculated from PC scores (PC models), and a model calculated from estimated PC scores (estimated model). The three circumferences were calculated from the three models, and these measurements as well as scan-extracted measurements were compared. Measurement errors were evaluated according to the method of ISO 20685. The results are: 1) Measurements from the original model and the scan-extracted measurements are not comparable for the waist circumference, probably because some landmark positions used in template fitting are biased; 2) measurements from the original model and PC model are comparable; 3) measurements from the PC model and estimated model are comparable; 4) errors are reasonably small on average, but there are cases with very large errors in all comparisons; 5) individuals with body shapes further away from the average body shape tend to have large measurement errors. The quality of scans data affects the quality of the original model and performance of functions to estimate PC scores. A large-scale scan data collected in retail shops may have big potential, but should be used with careful consideration of its shortcomings.

Paper 91 Neural Networks for Performance-Measure Selection with Predictive Human Models MOHAMMAD BATAINEH, TIMOTHY MARLER The University of Iowa A significant potential use of digital human models (DHMs) is not only the simulation of motion and behavior but the study of what drives such behavior. Although there are substantial benefits to being able to replicate human motion or posture virtually, an arguably more challenging goal is simulating and understanding what drives human motion and posture. Optimization-based predictive human modeling provides unique opportunities in this regard, where imbedded objective functions within the context of a computational optimization problem inherently represent what drives human performance. These objective functions are represented by mathematical models of human performance measures or combinations thereof. These performance measures may include concepts like discomfort, energy, joint displacement, etc. However, the process of determining which performance measure (s) should be used with specific tasks has not yet been automated or generalized. With this work, we propose the use of neural networks for automatically determining the most appropriate combination of performance measures for a given postureprediction task. We essentially propose a method for giving predictive human models a very basic brain. Artificial neural networks (ANN) are mathematical models for predicting system performance (i.e., system output) and are based on the structure and function of human biological neural networks. ANNs are trained based on pre-existing data, but are then use essentially to extrapolate results for scenarios not specifically represented in the training data. Although ANNs have successfully been used to predict motion and posture with DHMs, this research investigates the use of a general regression neural network for predicting the most appropriate weighting factors in a unique multi-objective optimization formulation for human posture prediction. The ANN is trained by associating task characteristics/parameters with performance-measure weighting factors used to predict realistic postures. The training data are initially gathered based on virtual experiments and validation. For a given task, weighting factors are selected such that the resulting predicted posture is realistic based on subjective evaluation. However, we also propose a separate optimization problem for extracting the weighting factors that can be used to reproduce the results obtained from motion capture. That is, given a posture from motion capture, we automatically determine which weighting factors could have been used to obtain the same posture virtually. Thus, as the DHM is

used over time, any results obtained either from motion capture validation or from actual software use are stored and continually used to train the avatar in much the same way a child learns as he/she grows. Then, the proposed ANN is used to predict the most appropriate set of weighting factors for any task or scenario. Initial results using optimization to extract weighting factors from motion-capture data have been successful. In addition, postures predicted using weighting factors determined by the ANN, which was trained by a finite set of predicted postures and known associated weighted factors, were successfully validated subjectively. Work is ongoing to further validate the results and use additional motion-capture data to train the neural network. In addition, work continues with interpreting the predicted weighting factors as well as the parameters inherent in the ANN algorithm in order to glean insight regarding what drives human performance.

Paper 92 Digital Human Modeling for Optimal Body Armor Design NIC CAPDEVILLA, TIMOTHY MARLER The University of Iowa Personal protective equipment (PPE), which includes traditional body armor, is a critical component for Warfighter survivability and performance. Designing a suitable PPE system is a complex and time consuming task with multiple design variables and constraints. To date most effort with PPE design, especially regarding body armor, has centered on materials development in response to necessary blast and ballistics requirements. Thus, most research and development has focused on experimental and computational evaluation and design of materials. Recently, however, the design focus has started to shift towards the complete body-armor system, not just the material(s) of a single plate or component. Improvement in survivability, stemming from incremental changes in material properties, can be minimal, and increased mobility can often be more advantageous than slightly stronger and lighter materials. Consequently, human-factors is starting to play an increasingly important role. Accordingly, although digital human modeling (DHM) has not yet been used to evaluate PPE, it offers significant opportunities for improved body-armor system design with potential reductions in developmental cost and timing. Often, the design of PPE can be task specific. There are situations in which the armor must be as lightweight as possible, thus sacrificing coverage, and other scenarios require increased coverage, which may lead to reduced range of motion. Thus, new capabilities have been developed for virtually importing body armor and parenting it to the Santos DHM. Santos can then execute a variety of tasks like reaching, aiming, walking, etc. Concurrently, the virtual armor is automatically tested with respect to weight, coverage, bulk, geometric encumbrance, range of motion, joint torque, balance, and performance. This allows one to compare various PPE systems from a human-factors perspective. We propose leveraging these new evaluation capabilities and implementing an optimization-based filer that automatically evaluates and selects the most advantageous PPE system. First, a library of employable, applicable armor systems is compiled. The user then indicates which metrics are used as objectives (i.e. minimize weight) and which metrics are used as constraints (i.e. ensure that weight is no more than twenty pounds). Then, the user sets up a task or set of tasks during which the armor is evaluated. Finally, the optimization filer automatically selects the system that optimizes the specified objectives while satisfying the specified constraints. Initial tests have shown realistic results with minimal computational demands. Ongoing work involves approaching the problem of armor design from a more dynamic standpoint by actually creating new armor systems on the fly. Using the above-mentioned metrics, armor components will automatically be created or deleted, and moved over Santos’s surface. Eventually, higher fidelity models pertaining to damage of internal organs can be used to provide additional objective functions. As new and successful armor systems are found, they can then be added to the library filtering system, which can provide optimal designs more quickly. The proposed capabilities can be used, not just with body armor, but with other types of equipment as well, and thus pave the way for automatic human-centric optimal design.

Paper 93 Information flow in small unit urban operations – An approach from industrial engineering OLIVER W. KLAPROTH*, THOMAS ALEXANDER “Fraunhofer-FKIE”, Fraunhoferstr. 20, 53343 Wachtberg, Germany Task force operations in urban environments have become increasingly important. In parallel, these operations are characterized by a vast amount of uncertainty, e.g. an uncertain security situation, an unclear depth and width of operation, loss of information because of visual obstructions. In typical urban operations, little intelligence is available and recently acquired information often requires instant action (McGuiness, 2000). Therefore, the information flow and team decision making in such operations is of special importance. Cognitive task analysis and task modeling have been used to analyze and describe the information flow in small teams operating in urban terrain. Document analysis and field observations provided insight into the complex domain of urban operations, while a knowledge elicitation technique, the critical decision method (CDM; Hoffman et al., 1998), was used to reveal the actual informational requirements of the single team members. The CDM is a semi-structured critical incident interview, in which a Subject Matter Expert retells an incident from experience. Acquired data were put into a bigger picture in a hierarchical task analysis (HTA; Annett, 2003), where the general task of clearing a building was iteratively split into subtasks. Data from instructions and cognitive task analysis as well as field observations were aggregated in a HTAdiagram, which marked the final step of system analysis and a first attempt at a system model. The HTA was taken together with temporal and agent-specific information and built into a K3-model (Foltz et al., 2000) that displayed the way information is acquired, processed and communicated. As a result, the characteristics of information flow in urban operation teams were summarized and compared to general criteria for adequate information flow in teams (Schraagen & Rasker, 2003). The methods were applied to a use case of dismounted soldiers in a “typical” urban operation and close-quarters-battle setting. The results show that in close-quarters-battle the organizational structure of forces in such small unit operations can best be described as a vertical team. Due to environmental characteristics (narrow doors, etc.), buildings and rooms are usually approached and entered by single team members consecutively (phalanx formation). Thereby, recently acquired information is passed through serially from the first through each single team member to the last one in line. Team decision making therefore follows a “forces in front are always right” principle. For stable but complex situations, the team structure changes to a flatter hierarchy and decisions are made by the designated team leader.

Paper 94 Estimating Anthropometry with Microsoft Kinect M. ROBINSON & M. B. PARKINSON OPEN Design Lab, The Pennsylvania State University Anthropometric measurement data can be used to design a variety of devices and processes with which humans will interact; however, collecting these data is a very time-consuming process. In order to facilitate the rapid and accurate collection of anthropometric data, a novel system is under development which makes use of multiple Microsoft Kinect sensors to estimate body sizes. The Kinect is an appealing sensor for this application because it can be purchased for around 250 dollars, can be quickly set up in many different environments, and can easily be interfaced with a Windows computer. This paper will present the experimental setup and procedure used as well as some preliminary results.

Paper 95 Modeling Post-Accident Vehicle Egress E. BAKER†, S. BUDZIK†, M. REED, PH.D.‡ † Tank Automotive Research Development and Engineering Center (TARDEC) ‡ University of Michigan Transportation Research Institute (UMTRI) This paper describes the application of the Jack human figure model to the analysis of vehicle egress by military personnel following an accident. In post-accident situations, the vehicle attitude may preclude

egress by the normal methods that have been addressed in most previous research. For this analysis, the Human Motion Simulation Framework reference implementation was used to aid the simulation. With the current level of functionality, only basic clearance issues could be investigated; improvements to the Humosim Framework and Jack software are needed to enable rapid and accurate generation of motion simulations for these non-standard situations.

Paper 97 Markerless Motion Tracking to Study Physical Stresses during Climbing S. HAN*†, S., S. LEE‡,, T.J. ARMSTRONG‡ † University of Illinois at Urbana-Champaign, Urbana, IL, USA ‡ University of Michigan, Ann Arbor, MI, USA Musculoskeletal disorders and falls are major causes of nonfatal occupational injuries―accounting for 32.8% and 25.3% of all workplace injuries and illnesses requiring days away from work, respectively. In an effort to prevent such injuries in workplaces, we developed a non-invasive markerless motion tracking system to measure and analyze postures and movements as workers climb on and off of equipment. The proposed study involves extracting the 3D positions of body joints from video recordings of workers, and modeling 3D skeletons that represent the workers’ movement over time. The extracted 3D skeletons can be used as a basis for biomechanical analysis that computes physical stresses on particular body parts. To this end, we conducted a laboratory study in which 3 human subjects climbed up and down a ladder 10 times, which allowed us to evaluate the performance of the proposed approach. Specifically, the locations of hands and feet in the x-, y-, and z-directions were estimated 30 times per second over all of the frames of the 10 trials, and these estimated locations were compared with the known positions of the ladder (e.g., rungs and rails where subjects made contact with their hands and feet) to assess the precision of the proposed system. As a result, the average Root Mean Square (RMS) error of the difference between estimated and known positions was 10.5 cm. The result may be hard to confirm yet that the proposed markerless motion tracking approach can produce outcomes as accurate as those from a marker-based system (e.g., VICON that tracks reflective markers attached to a body with 6 or 8 cameras); nevertheless, the system still enables us to identify different climbing patterns (e.g., hand-over-hand versus hand-to-hand), and to determine climbing speed. Moreover, there is still the potential to improve the accuracy by dealing with the estimated 3D locations through a post-process such as data smoothing (e.g., moving average filter) for the conversion of positions to angular data, which is the input format for biomechanical analysis. The converted angular information is readily available to manipulate 3D biomechanical models and assess musculoskeletal loads. In this respect, the proposed system has great potential for conducting ergonomic studies in field settings.

Paper 98 Measuring the maximum range of motion to create a valid database used for ergonomic analysis in the automotive industry. R. BICHLER†; M. AMERELLER†; A. SCHWIRTZ‡; R. KAISER† † BMW Group ‡ Technische Universität München For the year 2050 the U.N. estimates that the proportion of the world's population age 65 and over will more than double, from 7.6% today to 16.2%. Growing older implicates age dependent limitations regarding coordinative or biomechanical factors. One factor is flexibility, which is supposed to decrease with age because articular cartilage and ligaments will become less elastic and proprioceptive afferents are fewer or less competent. The loss of flexibility would possibly affect the manner of reaching tasks within a car, e.g. fasten the seatbelt. But valid data regarding the maximum range of motion (RoM) are difficult to find, a representative database which contains reliable data for all important joints of the human body is currently not available. A cooperation between the Technische Universität München and the BMW Group aims to make a first step solving this shortcoming. For collecting reliable data regarding the flexibility of all relevant body joints and their degrees of freedom(DoF) three measurement devices were developed by Max Amereller (unpublished PhD). It is possible to adapt these devices with respect to the individual anthropometrics of the subjects, within a range from the 1st to the 99th percentile of the corresponding body part length. The maximum range of motion for the right and the left side is measured in a 2-dimensional plane. All relevant DoF can be taken into account, which means 85 directions in summary.

A first study accomplished by Max Amereller was aimed to validate the developed measurement devices regarding learning effects, the effect of the measuring sequence, the reliability of the devices as well as inter- and intra-observer reliability and the laterality of subjects. Furthermore the subjects RoM reached during free motions were recorded with an optical Motion Capturing System for a later comparison. The validation study includes 50 subjects, 24 female and 26 male with the age of 29,84 ± 2,92 years and a body length of 169,7 ± 23,68 cm. Six joints were measured on the right as well as on the left side, each side at two different days. This means 58 directions in summary were measured with the devices developed by Max Amereller, plus 10 directions with standard equipment CROM und BROM. Every motion direction was performed 5 times by the subjects. For the hip external rotation (s=0.048), the shoulder extension (s=0.040) and the foot adduction (s=0.002) significant differences (α=0.05) were found regarding learning effects. For all other DoF it can be stated that there is no learning effect when performing the measurements more than 2 times by the subjects. No significant effect regarding the measurement sequence was found except for the Hip inner rotation (s=0.021). The reliability of the devices was proofed by measurements after repositioning the subjects at 2 different times in the devices, no DoF was found significantly different. Inter- and intraobserver reliability were confirmed with a correlation coefficient of 0.99 (Pearson). Significant differences were found between the subjects RoM of the left and the right side (for all pretests α=0.05). The consequence is the need for measuring both sides of the subjects. Beyond that it shows that assuming a symmetric RoM for subjects is mostly wrong. The comparison between the data captured with the Vicon system is actually ongoing; the results are expected end of 2012. Actually serial measurements are undertaken at the Technische Universität München with the devices and the method developed. The number of subjects is n=420, half female and male. 6 age groups are included, starting with the age of 25 up to the age of 85 years. The experimental measurements require 90 minutes per subject, including a questionnaire and a Bodyscan in three different postures. The measurements will be finished in February 2013. An important achievement of the serial study is a large and consistent database containing the maximum range of motion of the 420 subjects. The data will allow calculating percentiles for the RoM of relevant DoF of the human body, as it is known for anthropometrics. Amongst other things the experimental data will also be analyzed regarding the correlation between age and flexibility of each joint measured. Furthermore the correlation between the fitness, defined as hours of sports per week, and the maximum range of motion will be tested because it is assumed that the fitness has more impact at the flexibility than the age. Besides the knowledge obtained, these results will give a hint to the automotive industry to choose adequate subjects for evaluations when no objective RoM-measurements were performed by these subjects. At least it is aimed to implement the data in Digital Human Models used for ergonomic evaluations. After that, analyzing the reachability will be based on up-to-date data and the flexibility limits of a clearly defined customer group can be taken into account in an objective way.

Paper 99 Application of a character animation SDK to design ergonomic human-robotcollaboration F. BUSCH*†, S. WISCHNIEWSKI ‡, J. DEUSE † † Institute of Production Systems, TU Dortmund University, Dortmund, Germany ‡ Federal Institute for Occupational Safety and Health (BAuA), Dortmund, Germany Common digital human modeling (DHM) systems are sophisticated models, which need high efforts for simulation programming, in some cases disproportional to expected benefits. In consequence, these complex models require high expertise and training. For this reason, the usage of DHMs is often limited to larger companies with specialized planning divisions or to service providers. There are furthermore limited interfaces for implementing or using these DHMs in other software environments (e. g. offline programming systems for robots). This paper discusses the integration of a character animation system (CAS), originating from the entertainment industry, into an offline programming software enabling the virtual ergonomic assessment of a multi-robot system with direct human-robot-collaboration. The character animation system is used to implement a simple digital human model with basic functions. The aim is to simulate the entire human-robot-collaboration to plan robot trajectories in a way that physical stress on the human body is minimized. Final result is a classification of injury risks for the musculoskeletal system, which indicates the need and urgency for corrective actions. The main issues addressed in this paper are advantages and challenges of the described approach, for example in the field of anthropometric variances and accuracy of digital ergonomic assessment, and a discussion of further development needs.

Paper 100 A Mathematical Model of Motion Sickness in 6DOF Motion and Its Application to Vehicle Passengers T. WADA *†, NORIMASA KAMIJI‡,, SHUN’ICHI DOI‡ † Ritsumeikan University, 1-1-1 Noji-higashi, Kusatsu, Shiga, Japan ‡ Kagawa University, 2217-20 Hayashi-cho, Takamatsu, Kagawa, Japan A mathematical model of motion sickness incidence (MSI) is derived by integrating neurophysiological knowledge of the vestibular system to predict the severity of motion sickness of humans. Bos et al. proposed the successful mathematical model of motion sickness based on the neurophysiological mechanism based on the subject vertical conflict (SVC) theory. We expand this model to 6-DOF motion, including head rotation, by introducing the otolith-canal interaction. Then the model is applied to an analysis of passengers’ comfort. It is known that the driver is less susceptible to motion sickness than are the passengers. In addition, it is known that the driver tilts his/her head toward the curve direction when curve driving, whereas the passengers' head movement is likely to occur in the opposite direction. Thus, the effect of the head tilt strategy on motion sickness was investigated by the proposed mathematical model. The head movements of drivers and passengers were measured in slalom driving. Then, the MSI of the drivers and that of the passengers predicted by the proposed model were compared. The results revealed that the head movement toward the centripetal direction has a significant effect in reducing the MSI in the sense of SVC theory.

Paper 101 Simulation Validation Methods J. PELLETTIERE*† † Federal Aviation Administration, Washington DC Modeling and simulation is increasingly being used to represent occupant behavior, both of human subjects and of Anthropomorphic Test Devices. Future trends are towards the application of certification by analysis whereby a product, such as an aircraft seat, can be partially certified through analytical means that evaluates the occupant responses to dynamic events. If certification by analysis is to be utilized, then an objective method of determining model validity is needed so that the certification authority can make an objective determination of the model acceptability. Just using expert opinion could introduce bias and lead to inconsistent results depending on the particular engineer or office consulted. This objective method must be robust and able to discriminate different features of the resulting test data such as magnitude, phase, and pulse shape. Another consideration is in developing the baseline comparison data. In any testing, there is some inherent uncertainty and variation in the measured responses. If just an average is calculated, the result may not be representative of the true response as some features may get smoothed out. Likewise, calculating standard deviations around this average has similar issues. Each individual response should be treated as a separate case. A method derived from Sprague and Geers has been developed and applied to occupant models of seat simulations. Using the test data, acceptable error levels can be calculated that represent the uncertainty and provide consistent and objective results.

Paper 102 Building 3D Deformable Body Model and Torso Shape Estimation System S. SHUNTA*†, M. KOUCHI‡, M. MOCHIMARU‡, Y. AOKI† † Keio University, Graduate School of Integrated Design Engineering, Yokohama, Kanagawa, Japan ‡ Digital Human Research Center, National Institute of Advanced Industrial Science and Technology, Koto-ku, Tokyo, Japan Human body shape represented as a 3D mesh model can be achieved by various methods, for example, laser scan, projector camera system and volume intersection, etc. However, a more easier method is required in apparel markets. Because a virtual fitting technique or virtual cloth evaluation system is increasingly demanded on the back of online shopping and those techniques commonly requires the 3D body shape of users. Therefore, more simple and accessible method to obtain the 3D shape is necessary. Then, we focus the estimation of upper body shape in 3D using ordinary cues, silhouette images.

In this paper, to obtain 3D body trunk shape approximately but with sufficient degree of accuracy to represent individual differences, we propose a body model. The model has deformable body type and pose and is synthesized from a human shape database provided by DHRC. We use the model to estimate the 3D torso shape by optimizing the model shape and pose parameters. It is achieved by comparing input silhouettes and model silhouettes. The resulting estimation errors are evaluated by measuring the average errors in the nearest vertex distance between model and answer shape.

Paper 103 ema (Editor for Manual Work Activities) – A Holistic Method for Efficient Production Planning based on Digital Human Simulation L. FRITZSCHE*†, B. ILLMANN‡, W. LEIDHOLDT‡, S. BAUER‡, A. MORENO† † imk automotive, Inc., 5 Research Drive, 29607 Greenville SC, U.S.A. ‡ imk automotive GmbH, Annaberger Str. 73, 09111 Chemnitz, Germany In 2011, Fritzsche et al. introduced the EMA (Editor for Manual Work Activities) software tool to address the need for realistic and holistic assessments of physical workload in manual assembly tasks. Compared to other human modeling software, EMA is specifically designed for the production planner because it is easy to use by drag and drop and it uses the typical planning language based on MTM codes, for instance “pick part, go to car, place part, use screwdriver”. Thus, EMA greatly benefits the ideas of cost-efficient human simulation and realistic human movement generation. The clue to efficiency lies within the self-initiated motion generation and object reference of EMA. A suchlike generated simulation can easily be used for further analysis of results with high practical relevance and other applications. The integration of methods for ergonomics risk assessment, such as EAWS (Ergonomic Assessment Worksheet), allows a holistic evaluation of physical workload across the entire work process. Furthermore the use of standardized planning languages, such as MTM, and the linkage of human movements with all objects within the 3D environment allow a fast variation of processes, products and resources for testing alternative planning solutions in a cost-efficient way. Different design options can be compared in reference to needed production time (based on MTM) and physical workload (based on EAWS). Moreover, EMA now provides a set of tools that facilitate the design of efficient work processes by avoiding “waste” (with reference to Toyota Production System), such as ergonomic strains, long walking ways, and double-handling of parts and tools. In summary, these EMA tools enable the production planner to compare variants of product and process design by means of objective quantitative analyses of efficiency and ergonomics based on human simulations.

Paper 104 Dynamic Digital Human Model for ergonomic assessment based on human-like behaviour and requiring a reduced set of data for a simulation GIOVANNI DE MAGISTRIS*†, ALAIN MICAELLI†, JONATHAN SAVIN‡, CLARISSE GAUDEZ‡, JACQUES MARSOT‡ † CEA, LIST, Interactive Simulation Laboratory, 18 route du panorama, BP6, Fontenay aux Roses, F92265 France ‡ Institut national de recherche et de sécurité (INRS), rue du Morvan, CS 60027, Vandœuvre-lès-Nancy, F- 54519 France Biomechanical risk factors assessment of a work activity is usually based on the study of a human operator’s postures and forces while performing the work task. Hence, assessing the ergonomics of a future workstation at the design stage requires that an operator performs the work on a prototype or a similar equipment. An alternative solution has emerged through the use of digital human models (DHM) for ergonomics analysis. Yet, using industrial DHM software packages available for ergonomic assessment is usually a complex and time-consuming task. A challenging aim therefore consists in developing an easy-to-use DHM capable of computing dynamic, realistic movements and internal characteristics (position, velocities, accelerations and torques) in quasireal time, based on a simple description of the future work task, in order to achieve reliable ergonomics assessments of various work task scenarii at an early stage of the design process. We have developed such a dynamic DHM automatically controlled in force and acceleration, inspired by human motor control and based on robotics and physics simulation. In our simulation framework, the

DHM motion is dictated by real-world Newtonian physical and mechanical simulation, along with automatic control of applied forces and torques. Our controller handles multiple simultaneous tasks (balance, contacts, manipulation) in real time along with human-like feedforward force and impedance control. An experimental insert-fitting activity has been simulated and assessed based on the OCRA ergonomic index. A comparison with experimental human data showed consistent results: joint torques, DHM movements and their related OCRA assessment were realistic and coherent with human-like behaviour and performance. The main interest of our DHM is that it requires minimal information for a simulation: a starting point, an intermediate point for obstacle avoidance and an end point, along with the applied force for insert clipping. Moreover, changing the subject’s anthropometry and the scenario does not require new trajectory specification nor additional tuning.