Power Awareness In Context Aware Middleware for Ubiquitous Computing Systems Shu Lei, N.Q. Hung, L.X. Hung, S.Y. Lee Department of Computer Engineering Kyung Hee University, Korea {sl8132, nqhung, lxhung, sylee}@oslab.khu.ac.kr Abstract: Context aware middleware as the common approach that is used by the smart spaces or ubiquitous environment can provide the intelligent services to the end users. Power as the first class resource in the sensor network should be paid much attention. How to achieve using power efficiently is always the key issue for designing the ubiquitous computing system. In this paper we present the power aware functionality of our current research CAMUS project, a context aware middleware for ubiquitous Computing Systems. The principal contribution of our project is that we classify contexts into three categories: User Context, Physical Environment Context and Application Requirements. Based on this concept, CAMUS can provide power aware functions for end users. By using these functions, we can notify the end users about the changing of resource status to enable them to adapt their behaviors and complete tasks or change the battery before the power drains.

1. Introduction The goal of ubiquitous computing is to bring computation into the real physical world and to allow people to interact with them in a more natural way: by talking, by moving, pointing and gesturing. Once users can communicate directly with their computer-equipped objects and the environment, users are able to achieve their goals more easily and free their minds to think even further ahead of their current tasks and problems. The emerging of sensor networks that are built up by using hundreds or thousands small sensor nodes is the key to achieve the designing of ubiquitous computing environment. Every sensor node has the sensor to sense the useful information for its nearby physical environment and send this sensed data to the context aware middleware by going through the sensor networks. Context aware middleware is needed to acts as the mediator between the environment and users, to maintain sensing information and context data, as well as to provide reasoning/delivery service for context-aware applications. Context aware middleware is always working closely with the sensor networks, and based on the sensor networks to get the context data. Since power is the key issue to prolong the working life of the whole sensor networks, to make our context aware middleware be power aware and work for the whole sensor network is very necessary. Our CAMUS [1,12], a Context Aware Middleware for Ubiquitous Computing Systems, which has the final aim to deal with all kinds of contexts to provide intelligent services to the end users, also considers power as the first class resource that needed to be used efficiently to prolong the working life of sensor networks and provide better quality of services. In a short word, our CAMUS can play with two roles: one is the context aware middleware, which collects the contexts from the sensor networks to provide intelligent services for the end users, the other one is the power aware middleware, which well knows the power status of the whole sensor networks, and can provide several potential ways to prolong the working life of the whole sensor networks. In this paper we present the power aware functionality of our current research

CAMUS. The principal contribution of CAMUS is that we classify current contexts into three categories, User Context, Physical Environment Context and Application Requirements. Based on this concept, we use these three kinds contexts to provide power aware functions: 1) Tradeoff between the application requirements and different levels’ power consumption. 2) Using Physical Environment Context, especially, the power status information for directing power aware routing and notification. 3) Using user location and intension context for smart power management. 4) Finally, We also distinguish the two kinds of tasks that are under the timing constraint and power constraint respectively for using the dieing sensor networks’ power more efficiently. In next section, we present several related research works. In section 3 we introduce the definition of context and classify three kinds of contexts for power aware. In section 4 we present our system architecture and several functions for using power more efficiently. In sections 5, we describe two scenarios as the large area sensor network and small area sensor network respectively to show how our middleware works. In section 6, we propose assigning the different priorities to the tasks within different situations for using power more efficiently. We talk about the future work in section 7 and conclude in section 8.

2. Related research A tremendous amount of research has already been done to achieve context middleware and power aware middleware. In [2], the author presents the FlashBack, a peer-to-peer replication middleware designed specifically for power-constrained devices. In [3] the authors present a distributed middleware framework (parm), which is inherently power-aware and reconfigures itself to adapt to diminishing power levels of low-power devices. In paper [4] the author presents a middleware framework coordinating the Processor/Power Resource Management (PPRM) in mobile computing environment. From our previous survey [5], we can know several famous context aware middleware: Gaia [6], RCSM [7], Smart Kindergarten [8], and Context toolkit [9]. As we can know from these related research work, current context middleware just mainly use the contexts that collected by the sensor networks to provide intelligent services for the end users, for example, the people’s location, the situation of around physical environment in the Smart Classroom or the Smart Kindergarten. Most of them do not classify different contexts for using power to be aware and more efficiently. They also do not pay much attention on the system itself that is used to collect these contexts, for example the sensor networks. On the other hand, most of the current power aware middleware are not sensor networks based. They just get the power status information from the single device and make the single device or upper layer applications use power to be aware. Our CAMUS works not only as a new context aware middleware, but also provide the power aware functionality to use the power of the whole sensor networks more efficiently by combining both kinds of middleware’s concepts. Power awareness in CAMUS is one of our unique features that are different to those existing context aware middleware.

3. Classifying contexts into three categories Context always plays a very important role in developing intelligent applications in ubiquitous computing environments. Around how to define the context exactly, different researchers have different definition for the context. For example, Schilit describes context as the following [10]: In a mobile distributed computing system, contexts are the location of the user, the identity of people and physical objects that are nearby the user, and the states of devices that the user interact with. Dey defines context as [11] “Any information that can be used to characterize the situation of an

entity. An entity is a person, place, or object that is considered relevant to the interaction between a user and an application, including the user and applications themselves”. In our previous paper [5], we also give the definition as: “Context” is any information about the circumstances, objects, or conditions by which a user is surrounded that is considered relevant to the interaction between the user and the ubiquitous computing environment. We can understand that “Context” is a general word and has a loose definition. In our CAMUS, we classify the general contexts into three categories for using power more efficiently. “Users Context”, this kind information is mainly focus on the information about the users, such as the users’ location, profile, action, and intention, etc. “Physical Environment Context”, it consists of the hardware information, device capacity, power status information of each sensor nodes or the data sensed by the temperature sensor and the data sensed by the weight sensor, etc. “Application Requirements Context”, it consists of the sensing time, query frequency, quality level of contexts, etc.

4. System architecture 4.1 CAMUS architecture [1,12] The heterogeneity and dynamism inherent in smart/active spaces leads to high levels of heterogeneity at the link and transport layers. Services can be developed using existing technologies and deployed at the infrastructure level to mask such heterogeneity and provide support for handheld devices and applications executing in resource constraint environments. These services provide communication, amongst the framework components at different levels, and assistance to applications.

Fig. 1. CAMUS Architecture

In the CAMUS architecture, infrastructure services exist at three levels: Core Feature Services, Context-Feature Mediator Services and Core Context Services, as figure 1 shows.

Core Feature Services deal with extraction of features from sensor signals at the sensor layer, event subscriptions, triggering event notifications, handling queries and generating responses related to the basic feature information at the Feature Extraction Layer. Context-Feature Mediator Services include Context to Feature Mapping service (CFM), Context Aggregator service and the ontology middleware sub-system. Collectively, these services provide reasoning facility over the features through context modeling. CFM acts as a mediator between the feature and context layers by facilitating collection of features relevant to particular context e.g. to determine simple sensed context of “presence of a person X in room Y”, we can get the Feature Tuple related to RFID reader in Room Y with associated value of person equal to X or from the Feature Tuple {camera_room_X, “person”, “X”, timestamp}. The context received from CFM can be termed as ‘simple sensed context’. Information flow between application and context layer is based on contextual information that is mapped to corresponding features at the feature extraction layer. CFM also retrieves specifications, such as feature format, wrapper access patterns and sensor descriptions from the repository at the CA layer, for the consumption of FE and SN layers. These specifications are stored in OWL format for utilizing semantic representation advantages and to make change management easy. Context Aggregator service, after getting useful context data through CFM, saves them in OWL representations into the ontology middleware repository. The reasoning module provides reasoning functionality over the stored context knowledge in the repository for deducing higher-level context. Repository and pluggable reasoning module forms the heart of the ontology middleware. It can also act as an independent module helping services external to the infrastructure to query and gather reasoned context information of their interest. This repository is the first point of contact for context delivery services. In case the required context data is not available with the repository yet, the Context Aggregator service is requested to facilitate such a requirement. Core Context Services lie at the top of the architecture and consist of asynchronous Context Event service and synchronous Context Query service. These services handle the overall query/response and event registration/ notification tasks for the applications. 4.2 Context flow Users’ information, physical environments’ information and applications’ requirements are used as the contexts in our CAMUS. All of these contexts are

Fig. 2. Context Flow

sent to the middleware layer through the sensor networks and the application

layer as showed in figure 2. In order to help reader to understand this more clearly, we also provide CAMUS system architecture from the power awareness perspective, as showed in figure 3. In the context Awareness layer, Repository has the three different fields. One is used to store the Users Contexts, and one is used to store the Physical Environment Contexts, the last one is used to store the Application requirements.

Fig. 3. System architecture from power awareness perspective

4.3 Power aware functionalities By using three kinds different contexts, our CAMUS can mainly provide the following functions for using power more efficiently: 1) Power and Quality Aware Feature Extracting 2) Directing Power Aware Routing and Power aware Notification 3) Using Users’ location and intention for power management 4) Assign the priority based on power and timing constraint In next section, we describe how our CAMUS achieves these functions from two different cases: large area sensor networks and small area sensor network.

5. Large area case and small area case When the battery dies, the sensor node dies. In large area sensor networks case the batteries are generally un-changeable or unfeasible to be changed, for example the battlefield. Only in the personal network systems due to the area is small and the number of sensor nodes is limited, the batteries of sensor nodes can be recharged or changed. We have two approaches to get the power status information of sensor nodes. First, by using the power aware operating systems, sensor nodes can report power status information to the base station by themselves. But in this approach, for sending the power status information to the base station sensor nodes also will consume some power. Second, the sensor nodes have different working models, for example sensing model and relaying model, the base station can track each sensor node’s working model and working time to calculate the consumed power of that sensor node, then the base station can well know the status of the remaining power. By using the second approach, we can move the burden from the sensor networks to the base station. Generally we assume the base station is power line supported, so we don’t need to worry about the power consumption in the base station.

5.1 Large area case First, let us have a look at the large area sensor network, for example the battlefield. Several thousand sensors are deployed in remote terrain. The sensors coordinate to establish a communication network, divide the task of mapping and monitoring the terrain amongst themselves in an energy-efficient manner, adapt their overall sensing accuracy to the remaining total resource, and re-organize upon sensor failure. When additional sensors are added, the sensors re-organize themselves to take advantage of the added system resources. In this figure 4, we have five sub sensor network areas. Each area consists of many sensor nodes and one cluster. Every area sends the power status information to the Base station, or the Base station can calculate the working model and get the power status information by tracking each area. Because this is a large area sensor network, these entire sensor networks are deployed step by step.

Fig. 4. Large area case

Area A is the first area that was deployed, so compared with the other areas, area A’s power already was consumed for a long time and the status is low. We can have the clear result about the average power status of each area, as showed in figure 5.

Fig. 5. The power status of each area

Figure 5 shows us that area A has the lowest power status and the area E has the highest power status. Threshold is used to classify the sensor groups to the dieing stage. If power status level under this threshold, then CAMUS will automatically

consider this area is in the dieing stage and report to the end users or applications. Since the batteries of the large area sensor network are generally un-changeable or unfeasible to be changed, sometimes, we hope this very limited remaining power can be reserved for emergent using. Base on this result, our CAMUS can change the way of using this sensor networks, and adopt some modifications to balance the power of different areas for prolonging the working life of the whole sensor networks. Several potential ways are provided to use power more efficiently: 1) Using greedy algorithm [2] A greedy algorithm repeatedly executes a procedure that tries to maximize the return based on examining local conditions, with the hope that the outcome will lead to a desired outcome for the global problem. In some cases such a strategy is guaranteed to offer optimal solutions, and in some other cases it may provide a compromise that produces acceptable approximations. Sometimes, some sensor networks are deployed for remote monitoring as the figure 4. The scientists use these sensor networks to obtain experimental data. Because the power levels of different sub sensor networks are different as the figure 5, in order to prolong the working time of the whole sensor network, they need to balance the power of different areas. For this reason, they can use the greedy algorithm, which means the base station always requires the data from the area that has the richest power. By using the greedy algorithm we can keep the whole sensor networks as long as possible. 2) Directing power aware routing in large area [13] In paper [13], Mohamed Younis et al. had described an energy-aware routing protocol in cluster-based sensor networks. Each node in the sensor networks can have several working models, such as sensing model and relaying model. And each working model consumes different amount power of the node by considering the working time. The base station can based the working time and working model of each sensor node to calculate the power status of each node. Based on the power status of each sensor node, the base station can create the routing table to direct the message routing. In addition, as the time goes by, the power status of each sensor node is changing. Sequentially, the base station will modify the routing table based on the changing of power status. In our CAMUS middleware, we also adopt this approach to provide routing table by using the power status information and direct the routing in the sensor networks. 3) Power and Quality Aware Feature Extracting [14] Feature that extracted from the sensor network layer generally should satisfy the end users or applications’ quality requirements. The quality of the feature mainly relates to three factors: (1) the sensor’s sensing time, the more time that is used for sensing the more power is consumed; (2) the frequency of the sensor’s sensing, the more frequent to sense the data the higher correctness is provided, but the more power is consumed; (3) the number of the sensor nodes that are used to provide this feature, the more sensor nodes are used to sense the data the higher reliability is guaranteed, also the more power is consumed. Obviously, the higher quality level that required by the end users and applications will consume more power in the feature-extracting layer. Sometimes, we can change these three factors to reduce the quality of the feature to the minimum level, but still can provide enough information and can be accepted by the end users and applications. This is the tradeoff between the power consumption and the quality requirements from the applications and end users.

5.2 Small area case

In small area case, for example Context-aware Home [15] or Intelligent Hospital [16], the batteries of sensor nodes generally can be recharged or changed. Here we present the wired [17] and wireless [8] network architecture for a Context-aware Home. Both are the current sensor network architectures for personal network environment. As the following figures 6 and 7 show, in each room we deploy several sensor nodes and one cluster. Sensor nodes can send their messages directly to the cluster, and then the

Fig. 6. Small area case (1) Wired approach

cluster can forward these messages to the computer center or send to other clusters. By using this approach we can make the messages routing steps between different rooms reduced from the multi-hop to three steps: 1) Sensor node A Cluster A; 2) Cluster A Cluster B; 3) Cluster B Sensor node B [18]. If messages are routed by multi-hops, then a lot of sensor nodes just work as the routers. We hope all the power in the sensors just is used for sensing, processing, and communicating with cluster. By reducing the transmission steps, we can save the power in the sensor nodes.

Fig. 7. Small area case (2) Wireless approach Clusters can be designed to provide stronger processing and communication ability, and power is not provided by the battery but provided by the power line, so we don’t

need to worry about the power consumption in the clusters. Compared with the wireless approach, the wired approach is cheaper but has less mobility and more difficult to deploy. Sometimes only combining two architectures can get the best utility. In the small area sensor network case, CAMUS can provide the following functions: 1) Power aware notification As we assumed that this is a smart home, and in this home there are workroom, playroom, parents’ and child’s bedrooms, drawing room and toilet. We assign a room identifier to every room as the following figure 8.

Fig. 8. The table for room name and room number Because the number of these sensor nodes in each room is limited, so we can assign each sensor node a unique number as the identity. We design the following table to record these sensor nodes, as showed in figure 9.

Fig. 9. The table for room and sensor Since our CAMUS can well know the power status information of each sensor, we can create the power status table for each room as showed in figure 10. Here the power status of sensor node No. 3 of Room A is 17% and the power status of sensor node No. 4 of Room B is 24%, both sensor nodes are going to run out of power. Basing on this power status table, CAMUS can easily find out the dieing sensors and notify the applications or end users that these two sensor nodes’ power status is low. The applications can base on this information to adapt their behavior, for example stop sending more commands to the dieing sensors and just let them finish the existing tasks.

Fig. 10. The power status of each sensor node

By using the result from the figure 10, and searching in the figure 9 and figure 8, the end users or sensor network administrator also can exactly know which sensor node is going to run out of power and can know the exactly room and place of this sensor node, so that they can change the batteries or prepare new batteries for it before this sensor node stop working. 2) Use User location and intention for power management [19] Beside the Physical Environment Context and Application Requirements, users’ information is also very useful for power management. By using RFID, smart building, such as smart home and office can easily get the residents’ or officers’ location information for providing intelligent power management. For example, this family has the habit that they always stay in the Drawing room to watch TV programs from 8:00 PM to 10:00 PM in the evening. Then during this time only the sensor nodes that in this room need to work in the active status. After 10:00PM, parents and child go back to their own bed rooms, the changing in the location is automatically send to our CAMUS middleware, CAMUS then send command to all of the sensor nodes and appliances that are deployed in the Drawing rooms to change their working model into sleep or turn off. Another example, if there is one digital camera is used to capture the image with several frequencies in a room. When the RFID detects that some user is staying in the room, CAMUS can send command to this digital camera to capture the images with a higher frequency. If the user leaves this room, CAMUS can send command to this camera to change it working model into sleep or turn off.

6. Assign the priority based on power and timing constraint [20] For a certain sensor network, as the time goes by, the average power status is changing to be lower and lower. Based on the different power stages of the whole sensor network, we should provide the different approaches to correspond to its power status. Many researches, such as power aware routing and power aware broadcasting, had been invested to use the sensor networks when the power status of the whole sensor network is ample. But none of them consider the dieing stage of the sensor networks. In the dieing stage of the sensor network, due to the sensor nodes are dieing one by one, the messages that are routed in the sensor networks cannot avoid the losing. In order to use this limited remaining power as efficient as possible, in the dieing stage of the sensor network, CAMUS can inform this sensor network to change its routing approach. Here we distinguish the characteristic of current tasks in the sensor networks. Some of these tasks have the timing constraint and must be finished within a deadline, but some of the others do not have the timing constraint. For every task, it will consume some power when it is executing in the sensor networks. And different task will consume different amount of power. As the power is very limited in sensor network, we proposed to assign the priority to these tasks of the sensor network based on the power and timing constraints. Our key idea is to guarantee the timing requested tasks, and at the dieing stage of the sensor network we hope it can process as many as possible tasks that have no timing requirement. In our approach, every task can be assigned two kinds priorities based on power constraint or timing constraint. Timing Constraint If the tasks are timing constraint, we can assign the priority based on the timing. For example: Three tasks based on timing Task A: the most emergent Task B: not emergent Task C: the second emergent So the priority: Task A – 1; Task B – 3; Task C – 2.

Power Constraint If the tasks are power constraint, we can assign the priority based on the power. For example: Three tasks based on power Task A: Power consumption 6 W Task B: Power consumption 2 W Task C: Power consumption 4 W So the priority: Task A – 3; Task B – 1; Task C – 2. As showed in figure 11, each task is assigned the priority based on the timing and power constraint, and the final result will be decided by the corresponded situation.

Fig. 11. Assigning the priority to the tasks For one sub sensor network or special sensor node, we can get the Power Status Information of the sensor network as PSI. Our principle is as following: First: PSI = Power Status Information; Second: If PSI > Threshold then Working model = “Timing Constraint”; Else Working model = “Power Constraint”. Since CAMUS can well know the power status information of the whole sensor network as well as the single sensor node, it can change the working model once the power status under the threshold. In order to get the maximum performance of using power, the threshold can be decided after the testing work based on the specific working purpose and working model of that sensor network or sensor node. Sometimes, the end users for their specific reasons also can design the threshold. As a microscopical comparison, we assume the power of a certain sensor node in the dieing stage is only 5 W, and three kinds messages as A, B, and C are waiting for this sensor node’s routing. If we still use the priority from the Timing Constraint, then this sensor node should send the Message A first, but unfortunately the remaining power cannot support this sensor to finish this sending, because the remaining power 5 W is less then the 6 W, which is the nature requirement to route this message. But if we use the priority from the Power Constraint, we can finish sending of one Message C (5 W > 4 W) or two Message B (5 W > 2 W + 2 W). In our simulation, as the macroscopical comparison, we assume that many messages are flying in a sensor network, and they have to pass some sub sensor networks that the average power status is lower than the threshold. In the simulation, we randomly create 300 messages with the nature power consumption 2, 4, 6 W respectively. 90 dieing sensor nodes are deployed in this sensor network to route these messages, each dieing sensor node has the remaining power 10 W. When we route these messages by using the priority only from the Timing Constraint, 76% messages can be routed. When we use the priority from the Power Constraint, 84% messages can be routed. We get the improvement as 8%, which is showed in the figure 12. The key advantage

of this approach is that we can use the limited power more efficiently, but we must allow some messages can be lost in the sensor network.

Fig. 12. Comparison of performance

7. Future work A ‘smart office/Lab’ test bed has been setup and a prototype CAMUS system is under development for step-by-step evaluation of the proposed architecture. This will enable us to verify the architecture, witness strong and weak spots, improve upon the architecture and provide effective solutions for real world applications. As most of current context aware middleware have not tackled the issues of privacy and security. Some context information may be private and very important, for example the personal information in the context-computing environment [21]. Since we want to provide better security functionality for our context middleware, we must think about the excess power consumption for the providing security management. In the future we are going to find out the tradeoff between the power consumption and security.

8. Conclusion With the developing of wireless sensor networks, a middleware layer was introduced that whose functionalities are collecting raw sensor information, translating it to an application-understandable format, and disseminating it to interested applications. Our CAMUS provides an easy way for developers to specify how an environment should automatically respond to different contexts. Since power is considered as the first class resource in the sensor network and context aware middleware is always working closely with sensor networks, to provide the power awareness functionality for our CAMUS is very necessary to prolong the whole life of the sensor network. By classify the three kinds contexts, CAMUS can provide several power aware functions. This helps the end users and applications to be power aware for the whole sensor network or certain special device. End users or applications can adapt their behaviors and complete tasks or change the battery before the power drains.

References 1. Hung Q.N., Saad Kiani et al.: CAMUS – A Context-Aware Middleware Framework for Ubiquitous Computing Systems. Submitted to Ubicomp http://ucg.khu.ac.kr (2004) 2. Loo, B.T., LaMarca, A. and Borriello, G.: FlashBack: Power-Aware Peer-To-Peer Replication Middleware. The First International Conference on Mobile Systems,

Applications, and Services, Oct 16 (2002) Mohapatra, S., Venkatasubramanian, N.: PARM: Power Aware Reconfigurable Middleware. IEEE International Conference on Distributed Computer Systems (ICDCS-23), (2003) 4. Yuan, W.H., Nahrstedt, K.: A Middleware Framework Corrdinating Processor/Power Resource Management for Multimedia Applications. Proceedings of IEEE Globecom 2001, San Antonio, Texas, November (2001) 5. Hung, N.G., Shu Lei, Lee, S. Y.: A Survey on Middleware for Context-Awareness in Ubiquitous Environment. Korean Information Processing Society Review ISSN 1226-9182, July (2003) 6. Anand Ranganathan, et al: A Middleware for Context-Aware agents in Ubiquitous Computing Environments. http://choices.cs.uiuc.edu/gaia/papers/ GAIA Project (2003) 7. Yau, S. S., Gupta, S., Karim, F., Ahamed, S., Wang, Y., and Wang, B.: A Smart Classroom for Enhancing Collaborative Learning Using Pervasive Computing Technology, to appear in Proc. 6th WFEO World Congress on Engineering Education & 2nd ASEE International Colloquium on Engineering Education (ASEE2003), Nashville, Tennessee, USA, (2003) 8. Srivastava, M., Muntz, R., and Potkonjak, M.: Smart Kindergarten: Sensor-Based Wireless Networks for Smart Developmental Problem-Solving Environments, Proc. 7th Int’l Conf. Mobile Computing and Networking (MobiCom 2001), ACM Press, New York, (2001) 9. Dey, A.K., et al: A Conceptual Framework and a Toolkit for Supporting the Rapid Prototyping of Context-Aware Applications, anchor article of a special issue on Context-Aware Computing, Human-Computer Interaction (HCI) Journal, Vol. 16, (2001) 10. Bill S., Norman A., and Roy W.: Context-aware computing applications. In IEEE Workshop on Mobile Computing Systems and Applications, Santa Cruz, CA, US, (1994) 11. Anind K., Dey and Gregory D. A.: Towards a Better Understanding of context and context-awareness. Technical Report GIT-GVU-99-22, Georgia Institute of Technology, College of Computing, June (1999) 12. Hung, N.Q., Hung, L.X., Lee, S. Y.: A Middleware Framework for Context Acquisition in Ubiquitous Computing Systems. Second International Conference on Computer Applications (ICCA 2004), Myanmar 8th January (2004) 13. Younis, M., Youssef, M., and Arisha, K.: Energy-Aware Routing in Cluster-Based Sensor Networks. 10th IEEE International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunications Systems (MASCOTS'02) October 11 - 16, (2002) 14. Sinha, A., and Chandrakasan, A. P.: Operating System and Algorithmic Techniques for Energy Scalable Wireless Sensor Networks. Proceedings of the 2nd International Conference on Mobile Data Management, January (2001) 15. Meyer, S., Rakotonirainy, A.: A Survey of Research on Context-Aware Homes. Workshop on Wearable, Invisible, Context-Aware, Ambient, Pervasive and Ubiquitous Computing, Adelaide, Australia, February (2003) 16. Mitchell, S., Spiteri, M.D., Bates, J., and Coulouris, G.: Context-Aware Multimedia Computing in the Intelligent Hospital. Proceedings of the 9th ACM SIGOPS European Workshop Kolding, Denmark, September (2000) 17. Cho, S. Y.: Framework for the Composition and Interoperation of the Home Appliances Based on Heterogeneous Middleware in Residential Networks. IEEE Transactions on Consumer Electronics, Vol. 48, No.3, Software Center, Samsung Electronics, Korea, August (2002) 18. Ghiasi, S., Srivastava, A., Yang, X.J., and Sarrafzadeh, M.: Optimal Energy Aware Clustering in Sensor Networks. Sensor, University of California at Los Angeles, (2002) 19. Dalton, A.B., Ellis, C.S.: Sensing User Intention and Context for Energy Management. 9th workshop on hot topics on operating systems, USENIX, Duke University, February (2003) 20. Liu, J.F., Chou, P.H., Bagherzadeh, N., and Kurdahi, F.: Power-Aware Scheduling under Timing Constraints for Mission-Critical Embedded Systems in Proceedings of the 38th Design Automation Conference, pages 840-845, Las Vegas, NV, USA, June (2001) 21. Chen, G.L., and David K.: A survey of context-aware mobile computing research. Technical Report TR2000-381, Dartmouth College, Computer Science, Hanover, (2000) 3.

T

T

P

P

Power Awareness In Context Aware Middleware for ...

context aware middleware for ubiquitous Computing Systems. The principal ... thousands small sensor nodes is the key to achieve the designing of ubiquitous ..... the wired [17] and wireless [8] network architecture for a Context-aware Home. ... building, such as smart home and office can easily get the residents' or officers'.

196KB Sizes 2 Downloads 207 Views

Recommend Documents

A Middleware for Context-Aware Agents in Ubiquitous
computing is essentially a reactive approach to information access, and it ..... Context-sensitive object request broker (R-ORB) hides the intricacies of ad hoc ...

context-aware communication - CiteSeerX
Context-aware computing applications examine and react to a ... aware communication applications from the research .... net office network, desktop computers, and office ...... 102–4. [9] K. Nagel et al., "The Family Intercom: Developing a Con-.

Context-Aware Computing Applications
have been prototyped on the PARCTAB, a wireless, palm-sized computer. .... with Scoreboard [15], an application that takes advantage of large (3x4 foot),.

context-aware communication - CiteSeerX
phone 1. Receptionist assistant 2. □ Table 1. Context-aware software dimensions .... Text messages sent to a ... Anyone wishing to send a message invoked.

Context-Awareness for Indoor Spaces
Standard laptops/phones. Other APs. Processing server ... Not necessary RSSI. ▻ All are CoS ... E.g. Samsung Galaxy S5. Contributions. Ubiquitous ...

CARE: Context Awareness in Residences for Elderly
In order for elders to extend their lives at home, some form of automatic health monitoring is required. ... in three forms of assistance: “(1) by proving assurance that the elder is safe and is performing necessary daily ... and well documented AP

Context-aware applications - Semantic Scholar
Oct 22, 2004 - All providers of context information registered with Context Weaver ... correspond to Conference Room A, and that Conference Room A is ..... active source, the adapter is also responsible for calling a particular method ...

Context-aware applications
Oct 22, 2004 - Copies may be requested from IBM T. J. Watson Research Center , P. .... of a method call in a Java program is familiar with the semantics of ...

Context-Aware Computing Applications - CiteSeerX
with computers while in changing social situations. ... Three important aspects of context are: where you are, who you are with, and what ... Context includes lighting, noise level, network connectivity, communication costs, communication.

context-aware communication
particular media space research and awareness systems. ... situational and social awareness for others, in part to help ..... Both Context-Call [20] and Calls.calm.

Implementation of Power Aware Features in AODV for ...
Ad hoc wireless networks are power constrained since nodes ... proposed three energy efficient routing protocols at [4] and from .... service in the entire network.

Implementation of Power Aware Features in AODV for ...
and 0.95% for heavy data like 8 flows. As we can observe the improvement of energy consumption is greater for high data simulations and this is important.

Context-Aware Computing Applications
dialogs use the familiar alphabetical ordering by name or should they be ordered by location. ... Proximate selection may also be used to choose virtual objects.

Context-Aware Personal Route Recognition
contextual information (e.g. time of the day) into route recognition from ... personalized real-time information about location-specific events, implementing.

Context-aware HCI service selection
aDepartment of Computer Science and Engineering, Shanghai Jiao Tong University, ... associates interactions with services, and provided service selection ...

Context-Aware Service Composition in a Home Control ...
protocols, security policies, and such are dynamic and differ from one ... speech system if no display is available. If he comes at ... system is then given to using distinct service providers ..... Conference (CCNC'07), Las Vegas, January 2007.

Context-Aware Paradigm for a Pervasive Computing Environment ...
The proposed research, Context-Aware Paradigm for Pervasive Computing ...... The standardization provides an abstraction layer that separates the underlying ...

A System Architecture for Context-Aware Mobile ...
store user, device, and, for each geographic region, context information. In order to ...... tracking a co-worker you wish to talk to and tracking the office coffee cart in order to .... One such approach is to employ an “open” system composed of

A Non-Intrusive Context-Aware System for Ambient ...
... de Bretagne-Sud. Centre de Recherche, BP 92116, F-56231 Lorient Cedex, France. E-mail: [email protected]. 2 Electronic Lab of Kerpape Functional Reeducation and Rehabilitation Center ..... DogOnt (Domotic OSGI Gateway) [19, 20].

Efficient, Context-Aware Privacy Leakage Confinement for Android ...
Android; Privacy leakage; Context-aware policy; Bytecode rewrit- ing. Permission to make ..... cleaned to be “false” if imei is set as a constant (Ln.10). Without.

Context-Awareness in Geographic Information Services
Sep 23, 2014 - Information Science, for example: Geographic Information Retrieval, Location-Based. Services, Mobile Cartography, and Recommender Systems. Different approaches have been proposed in the literature to model and use context information i

Beacon-based context-aware architecture for crowd ... - Gerardo Canfora
According to Apple guidelines for iBeacons17,18, the beacon payload contains editable static data such as a 16 byte. UUID field that represents the particular ...

Augmenting Conversations through Context-Aware ...
The amount of video content available on the Internet has increased dramatically .... indexed into a special data structure to optimize the speed and performance in .... [18] SoX - Sound eXchange. http://sox.sourceforge.net. [19] Taylor, A. and ...

Augmenting Conversations through Context-Aware Multimedia ...
Oct 20, 2011 - call-center conversations and multimedia contents [14,20,9]. Most of them relied on the fact that the retrieval accuracy is not significantly ...