Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753 www.elsevier.de/aeue

Adaptive digital phase modulation schemes using transition-initiated phase acceleration Rajarshi Mahapatra∗ , Anindya Sundar Dhar, Debasish Datta Department of Electronics and Electrical Communication Engineering, Indian Institute of Technology, Kharagpur 721 302, India Received 16 November 2006; accepted 24 September 2007

Abstract Link adaptation (LA) schemes, where signal transmission features, such as, modulation, coding rate, etc. are dynamically selected based on the feedback information from receiver regarding the channel condition, have recently emerged as powerful tools for increasing data rate and spectral efficiency of wireless networks. In this paper, we examine an adaptive modulation scheme for LA, based on digital phase-modulation techniques, employing a novel phase-shifting scheme. In the proposed technique, the abrupt phase shift of carrier waveform, required after each data transition, is realized by incrementing the frequency of the voltage-controlled oscillator (VCO) for a short duration, immediately following the data transition. The momentary increase in VCO frequency is realized by feeding the VCO with narrow control pulses, derived from the baseband data stream using a simple supporting circuit. The proposed technique can be used for very large scale integration (VLSI) of adaptive modulation schemes, and one such candidate scheme has been designed for VLSI implementation with triple options of binary phase-shift keying (BPSK), quadrature PSK (QPSK), and quadrature amplitude modulation (QAM). 䉷 2007 Elsevier GmbH. All rights reserved. Keywords: BPSK; QPSK; QAM; VCO; VLSI; Transition-initiated phase acceleration (TIPA); BER; Link adaptation

1. Introduction In recent years, wireless communications industry has gone through unprecedented developments in both fixed and mobile applications. Continued increase in demand for all types of wireless services (voice, data, and multimedia) have fuelled the need for higher capacity. Although improved compression technologies have reduced the bandwidth needed for voice calls, data traffic will demand much more bandwidth as new services come on-line. The wireless data channels are subject to significant interference and fading, resulting in widely varying received signal quality. Signal variation occurs due to three major causes: variation of received signal strength with distance from the transmitter, shadow fading caused by large obstructions and Rayleigh ∗ Corresponding author.

E-mail address: [email protected] (R. Mahapatra). 1434-8411/$ - see front matter 䉷 2007 Elsevier GmbH. All rights reserved. doi:10.1016/j.aeue.2007.09.011

fading caused by local scatterers around the receiver [1,2]. Furthermore, in packet-based data systems, the bursty nature of data traffic also causes rapid changes in the interference characteristics. In order to handle such rapid variation in signal conditions, techniques that adapt the bit rate, transmission power and other transmission features/parameters to channel conditions are being proposed for next-generation wireless system. In this context, emerging technologies that can improve wireless spectrum efficiency are becoming a necessity, especially in broadband applications. One useful approach in this direction is to employ link adaptation (LA) [3] using smart antennas with adaptive beam formation [4], in particular the multiple-input multiple-output (MIMO) technology [5], coded multi-carrier modulation, and transmission schemes using dynamic modulation and coding techniques [6]. The basic idea behind LA techniques is to adapt the transmission parameters to cope up with the prevailing channel

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753

conditions. It exploits the variations of the wireless channel (over time, frequency, and/or space) by dynamically adjusting certain key transmission parameters observed between the base station and the subscriber. The major parameters to be adapted include modulation and coding schemes, but other parameters can also be adjusted for the benefit of the system, such as, power level (as in power control), transmission rate, and more. Implementation of LA requires feedback information from the receiver to the transmitter about the link conditions [3,7]. More specifically, the growing popularity of space-time coding (STC) [8] and availability of a wide range of modulation schemes can support the need for LA solutions that integrate temporal, spatial, and spectral control functionalities together. The systems are now-a-days becoming more and more complex and their implementation can only be carried out efficiently with the help of very large-scale integration (VLSI). The key issue therein is the design of robust low complexity and cost effective solutions for compact VLSI implementation. An important indication of the popularity of such techniques is the current proposals for third generation wireless packet data services, such as code division multiple access (CDMA) schemes like CDMA2000, wideband CDMA (WCDMA) and general packet radio system (GPRS, GPRS-136), wherein LA is recommended as a means to provide higher data rates. In practical LA implementations, the values for the transmission parameters are quantized and grouped together in what we refer to as a set of modes. Each mode is usually limited to a pair of modulation schemes or more. One of the goals of an LA algorithm is to ensure that the most efficient modulation is always used when channel conditions change, based on a mode selection criterion [9]. Therefore, in LAbased wireless transmission systems, one of the major functionalities is to implement adaptive modulation schemes offering reconfigurable digital phase-modulation options, which in turn needs simpler ways to realize various modulation functionalities on a single VLSI fabric. In digital phase modulation, phase-shifting mechanism of carrier at data transitions is in general an essential operation in modulators, such as in systems using binary phase-shift keying (BPSK), quadrature phase-shift keying (QPSK), minimum-shift keying (MSK), quadrature amplitude modulation (QAM) and so on. The phase-shifting mechanism of BPSK is usually conceived as a multiplier functionality with unmodulated carrier and data stream as its two input waveforms. In other phase modulations, the data stream is divided into in-phase and quadrature components and then used to modulate the carrier by two multipliers. However, multiplier functionality being complex in general for VLSI implementation, several other methods have been explored for implementation of digital phase modulation systems. Two such methods, using 180◦ phase shifter [10] and a phase-splitter circuit [11] were proposed in microwave band, and are useful for VLSI implementation. In another study [12], BPSK modulator has been proposed with a voltage-controlled

741

current source, a resonator and a limiting amplifier. An implementation of digital modulators was proposed in [13] without true hardware multiplier. In order to reduce the hardware, the author utilized the periodicity and other properties of sine and cosine functions to implement the digital modulators. In [14], the author proposed a transmultiplexer, which used a multiplier-free modulation scheme. It was pointed out that the frequency shifting can be carried out without requiring multiplications by numbers other than +1, −1, and 0, and that multistage direct modulation schemes for large-channel systems can be devised this way without the need for multipliers in the modulators. In practical implementation, the digital phase modulation in microwave range has been implemented either by using a digital phase shifter [15] or by numerically controlled oscillator governed by the data sequence. In the present work, we propose an adaptive modulation scheme with triple options for BPSK, QPSK and 16-QAM modulations, which obviates the need of the multiplier functionality. In particular, to simplify the modulation hardware, the phase shift of the carrier waveform (for BPSK, QPSK and 16-QAM) required at the moment of a given data transition is realized by incrementing the instantaneous frequency of a voltage-controlled oscillator (VCO) for a short duration immediately following the data transition. During this short duration, due to the increment in instantaneous frequency, the carrier at VCO output experiences a transition-initiated phase acceleration (TIPA). Performance of basic TIPA implementation for BPSK modulation has been examined in our earlier work [16], while in the present work, we propose a VLSI implementation of a BPSK/QPSK/16-QAM-based adaptive modulation scheme using TIPA-based digital phase modulation of a single VCO. First, we examine the operational features of the TIPABPSK and analyze its performance in terms of receiver bit error rate (BER) and spectrum. The proposed method provides a generic VLSI implementation scheme for phaseshifting operation and one can use the same as a building block to implement other phase-modulation schemes, such as, in MSK and higher-order QAMs. As mentioned earlier, LA requires feedback information from the receiver to the transmitter about the link conditions. However, in our work, we assume that this information is available from the relevant receiver end from the measurements on its short-term signal-to-noise ratio. The rest of this paper is organized as follows. Section 2 presents the TIPA-based phase-shifting technique with the governing equation to be used for designing a given BPSK modulator. Subsequently, the proposed technique for the BPSK modulator is extended for QPSK and 16-QAM modulators. In Section 3, we present an analytical model for evaluating BER at the receiver for the proposed TIPA-BPSK modulation scheme and the spectrum of the BPSK waveform obtained using this scheme is examined through simulation. In Section 4, we report a possible CMOS VLSI implementation of the proposed TIPA-based modulators and examine

742

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753

their design issues. Section 5 provides a realistic approach of adaptive modulation with the help of TIPA-based modulation schemes. Finally, Section 6 presents the concluding remarks on our work.

2. TIPA-based phase-shifting scheme In this section, we explain the basic principle of the TIPAbased phase-shifting scheme, which obviates the need of multiplier functionality, and its realization into BPSK, QPSK and 16-QAM modulation schemes. Under the TIPA scheme, the phase swing of the carrier waveform at each data transition is achieved by accelerating the phase of the carrier waveform during a short interval immediately following each data transition. In the present work, all the above-mentioned TIPA-based modulation schemes have been realized with the help of a single VCO and appropriate control circuits.

2.1. TIPA-BPSK In this subsection, we reproduce the theoretical model used in our earlier work [16] for evaluation of the TIPABPSK system, for the sake of completeness of the present paper. The phase-shifting scheme for the TIPA-based BPSK under consideration is illustrated in Fig. 1, wherein the carrier frequency fc of a VCO is momentarily increased (using a rectangular control pulse at VCO input) by an amount f for a short duration , following each data transition, with  being much less than the bit interval (T). During this interval (), due to the increment in instantaneous frequency, the carrier waveform at VCO output experiences a TIPA. If  and f are chosen appropriately during the TIPA interval, carrier phase can undergo an additional phase shift of 180◦ (or its odd multiple) at the end of the TIPA interval, and thereafter the carrier returns to its original frequency fc but with a phase shift of 180◦ with respect to its waveform in the preceding bit interval. Thus the carrier waveform achieves the desired phase shift following the data transition without any multiplier operation, albeit with an imperfect waveform during the short duration () of the TIPA interval. Indeed,

to make the TIPA-based scheme comparable in performance with an ideal BPSK generation scheme, one needs to minimize  (as compared to T) and we examine this issue later. The governing equation for  and f is obtained by ensuring that, at the end of a TIPA interval, the phase-accelerated carrier with an instantaneous frequency (fc + f ) accumulates an additional phase equal to an odd multiple of  with respect to the original carrier waveform of frequency fc , i.e., 2(fc + f ) = 2fc  + (2n + 1),

(1)

where n represents a positive integer including zero. This in turn implies that one should ensure the following design constraint on  and f , given by 2f  = (2n + 1).

(2)

The BPSK waveform, generated in this process after a given data transition, can be represented in two parts (first part during the TIPA interval, and the second part in the remaining period in a bit interval following the completion of the TIPA interval) as  Ac cos(2fc + 2f )t 0 t , vT X = (3) Ac cos(2fc t)  < t T , where Ac represents the carrier amplitude. The abovementioned TIPA processing (earlier used in the optical DPSK system [17]) can be implemented using a simple block schematic as shown in Fig. 2. In Fig. 2, first, one can generate the rectangular control pulses after each data transition by applying a delay-and-EXOR (DEO) operation on the baseband data stream, wherein the delay is set at  with the help of suitably loaded chain of inverters. Further, the height (h) of the control voltage pulses of DEO output is adjusted (level translated) by an appropriate value (K), so that

Input Data Stream

Pulse Amplitude Control

Voltage Controlled Oscillator

BPSK Output

Delay

Fig. 2. Phase acceleration of VCO using DEO scheme.

Input Bit Stream

T Narrow Pulse to VCO Input

ε

BPSK Output

Fig. 1. Waveforms for illustrating the TIPA-BPSK scheme.

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753

743

Table 1. Minimum frequency required for the VCO to satisfy the phase reversal condition Free running frequency ( = Carrier frequency fc ) (MHz)

Bit rate (kbps)

n (/T )

Minimum frequency required during TIPA (fc + f ) (MHz)

900

270.8

1/10 1/50 1/100 1/500 1/1000

901.353984 906.769984 913.540032 967.370016 1035.400000

Table 2. Frequency required for the VCO to satisfy the phase-shifting condition for a fixed value of n ( = 1/100) Free running frequency (MHz)

Bit rate (kbps)

Phase shift (Deg.)

Required frequency obtained analytically (MHz)

Frequency measured from simulation (MHz)

900

270.8

0 90 180 270

900 906.769 913.540 920.310

900 906.726 913.545 920.319

the consequent frequency increment for VCO during TIPA interval equals f (which increases with increasing h). It may be noted that, although one can readily choose a suitable combination of  and f to satisfy the constraint in Eq. (2), performance of the TIPA-BPSK would be closer to the conventional BPSK only if its imperfection during the TIPA interval is at minimum with >T . Also, a smaller setting for  will demand a higher value for f . Table 1 shows how to design the frequency fc +f required for the VCO to satisfy phase-shifting condition corresponding to some possible values of normalized TIPA interval (n = /T ) in the range of 1/1000 to 1/10. For this purpose, we have considered fc =900 MHz and bit-rate=270.8 kbps, which are used in GSM application. As seen in Table 1, the value of required frequency fc + f for 180◦ phase shift is 901.35 MHz for n =1/10 and for n =1/1000, it is as high as 1035.400 MHz. But for n = 1/100, the value of fc + f is not very high, i.e., close to the carrier frequency fc . Since, the required frequency fc + f is inversely proportional to n , therefore, in reality, their practical design values will depend on the respective realizable upper/lower limits (presumably, with upper limit being applicable for f and the lower limit for ). We examine in Section 3 the impact of the non-ideal aspect of the TIPA-BPSK (due to the finite TIPA interval) on the system performance in terms of receiver BER. Next, we extend the TIPA-BPSK scheme to realize other two modulation schemes, viz., QPSK and 16-QAM.

Q 1011

1001

b2

b1

1110

1111 a2 a1

1010

1000

1100

1101

I 0001

0000

0100

0110

d1

d2

0101

0111

c1 c2

0011

0010

Fig. 3. Typical constellation diagram of a 16-QAM.

of every new symbol, the phasor of the modulated signal exhibits a phase shift of 0◦ , 90◦ , 180◦ or 270◦ according to the data bits in the present and the next symbols. Therefore, the design equation for the TIPA-QPSK is given by 2f  = (2n + k/2),

(4)

where n represents a positive integer including zero and k is 0, 1, 2, 3. By using Eq. (4), we present in Table 2 an estimate of the required frequencies needed for different QPSK phase shifts during TIPA interval for a given value of n = 1/100.

2.2. Carrier frequencies for TIPA-QPSK 2.3. Carrier frequencies for the TIPA-16-QAM In QPSK, there are four possible combinations of data bits in a pair to be transmitted through in-phase and quadrature components. The QPSK-modulated signal transmits two bits at a time, such as 00, 01, 10 and 11. Thus, at the beginning

QAM is a modulation technique which combines ASK and PSK. In 16-QAM, there are a total of 16 possible combinations for data bits. The transition from any one state to

744

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753

Table 3. Required phase shift for changing the states of data bit in 16-QAM (Fig. 3)

1101 1100/1111 1110 1001 1000/1011 1010 0001 0000/0011 0010 0101 0100/0111 0110

Degree 18.43 45 71.56 108.43 135 161.56 198.43 225 251.56 288.43 315 341.56

Previous constellation points (ZYXW)

Amplitude level

1101

1100/1111

1110

18.43◦ 0◦ 26.56◦ 53.13◦ 90◦ 116.56◦ 143.13◦ 180◦ 206.56◦ 233.13◦ 270◦ 296.56◦ 323.13◦

45◦ 333.43◦ 0◦ 26.56◦ 63.43◦ 90◦ 116.56◦ 153.13◦ 180◦ 206.56◦ 243.13◦ 270◦ 296.56◦

71.56◦ 306.86◦ 333.43◦ 0◦ 36.86◦ 63.43◦ 90◦ 126.56◦ 153.13◦ 180◦ 216.56◦ 243.13◦ 270◦

another is possible depending on the data sequence during two consecutive symbol periods. From the typical rectangular constellation diagram of 16-QAM, as shown in Fig. 3, we observe that data points are located at 12 different angles, viz., 18.56◦ , 45◦ , 71.43◦ , 108.56◦ , 135◦ , 161.43◦ , 198.56◦ , 225◦ , 251.43◦ , 288.56◦ , 315◦ and 341.43◦ with three different amplitude levels. However, TIPA-based phase-shifting scheme is applied during the phase change between previous state to present state. In the design of QAM modulation, different amplitude levels are to be taken into account along with the phase changes. Table 3 shows that a total of 20 unique phase shifts (including 0, which implies no shift at all) with three different amplitude levels are required when the data sequence changes from any constellation point (ZYXW) of a particular quadrant to any other point (SRQP) of the constellation diagram in Fig. 3. Thus, only 20 phase shifts are sufficient to represent all possible changes from any constellation point to any other point in the constellation diagram. This is due to the reason that the same phase shift occurs, when the constellation point moves from a2 to b1, b2 to c1, c2 to d1 and d2 to a1 or a2 to b2, b2 to c2, c2 to d2 and d2 to a2 or a1 to a2, b1 to b2, c1 to c2 and d1 to d2 or so on. This makes the constellation diagram quadrant symmetric. Therefore, for the TIPA-16QAM, the carrier frequency of VCO requires the provision of 20 different phase shifts during TIPA interval, depending on the data sequence, along with an amplitude variation of the carrier waveform out of three possible levels. We make use of this observation in Section 4, while describing the VLSI design scheme of TIPA-16-QAM modulator.

3. Performance analysis of the TIPA-BPSK scheme In this section, we analyze the performance of the TIPAbased BPSK scheme. First, the spectrum of TIPA-BPSK

Level Level Level Level Level Level Level Level Level Level Level Level

2 1/3 2 2 1/3 2 2 1/3 2 2 1/3 2

0 Conventional BPSK TIPA-BPSK

−10 Power Spectrum (dB)

Present constellation points (SRQP)

−20 −30 −40 −50 −60 −70 −80 2985

2990

2995

3000

3005

3010

3015

Frequency (Hz)

Fig. 4. Spectrum of the TIPA-BPSK signal with n = 0.1 and the conventional BPSK.

signal is obtained through simulation and the effect of the TIPA interval () on the spectrum is observed. Next, the BER is evaluated for a given transmission system with the TIPABPSK modulation. Subsequently, these results are compared with the spectrum of conventional BPSK.

3.1. Spectrum of the TIPA-BPSK waveform In order to understand the spectral behaviour of TIPABPSK, we carried out simulation of the TIPA-BPSK waveforms using Matlab/Simulink. To evaluate the spectrum, we have considered the same ratio between the carrier frequency and the bit rate as used in GSM system. The spectra of a TIPA-BPSK are shown in Figs. 4 and 5 for different values of the TIPA interval and compared with the spectrum of conventional BPSK. Fig. 4 shows that there is an asymmetry between the two sides of the TIPA-BPSK spectrum around

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753 0 TIPA- BPSK Conventional BPSK

−5

Power Spectrum (dB)

−10 −15 −20 −25 −30 −35 −40 −45 2985

2990

2995

3000

3005

3010

3015

Frequency (Hz)

Fig. 5. Spectrum of the TIPA-BPSK signal with n = 0.001 and the conventional BPSK.

the carrier frequency with a higher spectral density on the higher frequency side of the carrier. This is expected due to the presence of incremental frequency (f ) during the TIPA interval (). Because the TIPA-BPSK would have the same average power as the conventional BPSK, the spectrum exhibits higher side lobes on higher-frequency side with a reduced side lobe on lower-frequency side, thereby preserving the total power of the waveform as in conventional BPSK. As expected, the spectrum of the TIPA-BPSK waveform approaches that of the conventional BPSK waveform if we reduce the TIPA interval, as shown in Fig. 5.

3.2. Receiver bit-error rate (BER) for TIPA-BPSK transmission system For BER evaluation of a TIPA-BPSK system, we consider an ideal correlation receiver for BPSK demodulation, with the assumptions of ideal carrier and clock recovery subsystems. It may be noted that, the TIPA-induced waveform impairment (during TIPA intervals) does not occur in a TIPABPSK waveform segment, when the data stream carries a continuous chain of binary zeros (space) or ones (mark). However, the impairment in the TIPA-BPSK waveform (and also in its demodulated baseband version) can occur for a binary zero when it is preceded by a binary one in the parent data stream, and similarly for a binary one when it is preceded by a binary zero (i.e., in any bit interval that is preceded by a data transition). Thus, the demodulated baseband signal, whether a binary one or a binary zero, may not be (or may be) affected by the TIPA impairment, if its preceding binary symbol was same (or different). With this observation, we therefore express the receiver BER with four components as follows: Pe = P (0)[P (00)P (1|0) + P (10)P  (1|0)] + P (1)[P (11)P (0|1) + P (01)P  (0|1)],

(5)

745

where P (0) and P (1) represent the transmission probabilities for binary zero and one, respectively. P (1|0) and P (0|1) represent the conditional error probabilities for the receiver when a binary symbol is transmitted without being preceded by a data transition, and P  (1|0) and P  (0|1) are the conditional error probabilities for the receiver when a binary symbol is transmitted, having been preceded by a data transition (and, hence with a TIPA impairment). Given that a binary one has been transmitted, P (11) represents the probability that a binary one is preceded by a binary one itself and P (01) represents the probability that a binary one is preceded by a binary zero. Similarly, given that a binary zero has been transmitted, P (00) represents the probability that a binary zero is preceded by a binary zero itself and P (10) represents the probability that a binary zero is preceded by a binary one. It may be noted that, each of P (01) and P (10) causes the occurrence of data transition, leading to the consequent TIPA-induced impairment. With this representation and assuming that P (0) = P (1) = 21 , P (00) = P (10) = 21 , and P (01) = P (11) = 21 , Eq. (5) can be simplified as Pe = 41 [P (1|0) + P  (1|0) + P (0|1) + P  (0|1)].

(6)

Assuming white Gaussian noise and a binary symmetric channel, one can have P (1|0) = P (0|1). Moreover, the imperfection in the carrier waveform during the TIPA interval is also expected to be identical for 10 and 01 combinations of binary symbols, leading to the identical degradation in baseband signal at the correlation receiver output. This enables us to assume that P  (1|0) = P  (0|1). With these observations, we finally simplify the above expression for BER as Pe = 21 [P (0|1) + P  (0|1)].

(7)

Considering the above formulation, we next evaluate P (0|1) and P  (0|1) at the correlation receiver output as follows. Assume that the received signal has an amplitude A and the locally recovered carrier has an amplitude B. One can, therefore, express the baseband noise and signal components at the correlator output (integrate and dump operation with a gain parameter = 1/int ) as follows:  T 1 n(t) × B cos(2fc t) dt, (8) int 0  T 1 A cos(2fc t) × B cos(2fc t) dt, (9) s0 (T ) = int 0   1 A cos(2fc +2f )t×B cos(2fc t) dt s0 (T ) = int 0  T 1 A cos(2fc t)×B cos(2fc t) dt. (10) + int 

n0 (T ) =

In Eq. (8), n0 (T ) represents the noise component of the correlator output in the receiver, while in Eqs. (9) and (10), s0 (T ) and s0 (T ) represent the baseband signals for binary one reception without and with TIPA imperfections, respectively. Using these expressions, we finally arrive at the

746

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753 0

10 10

Bit Error Rate

10 10 10 10 10

Conventional BPSK TIPA BPSK with ε = 0.1

1

n

TIPA BPSK with εn = 0.01 TIPA BPSK with ε = 0.001 n

2

3

4

5

6

10 0

5

10

15

Signal to Noise Ratio (dB)

Fig. 6. Plots of BER vs. SNR with different n .

10

4. VLSI design

1

a

Bit Error Rate

10

10

10

10

10

b

c

d

e

f

2

3

4

5

6

10

4

10

3

10

2

10

1

0

10

Normalized TIPA Interval ( εn )

Fig. 7. Plots of BER vs. normalized TIPA interval (n ): (a)—conventional BPSK with SNR = 10 dB; (b)—TIPA-BPSK with SNR = 10 dB; (c)—conventional BPSK with SNR = 8 dB; (d)—TIPA-BPSK with SNR8 = dB; (e) conventional BPSK with SNR = 4 dB; (f) TIPA-BPSK with SNR = 4 dB.

receiver BER, given by Pe = 41 [erfc(X) + erfc(X  )],

(11)

where, X= ABT  =signal-to-noise-ratio (SNR) of the receiver X =

interval = /T , T = T − n1 fc , where n1 represents the rounded-down whole number of Tf c . Next, we present the results of our analysis with plots of BER vs. SNR and normalized TIPA interval in Figs. 6 and 7, respectively. As expected, it is evident from Fig. 6, with decrease in n , the BER vs. SNR plot for the TIPA-BPSK approaches the same for the conventional BPSK. However, one needs to estimate from this plot the maximum acceptable value for n , that can offer a specified BER. From these two figures, it is observed that, for achieving a BER of 10−9 with n = 1/100, TIPA-BPSK requires power penalty well below 1 dB with respect to conventional BPSK. Moreover, from Table 1, we also find that with n = 0.01 and fc = 900 MHz one needs a minimum value for fc + f of 913 MHz, which is reasonable enough (i.e., not too high) for implementation. Next we discuss the VLSI design of different TIPA-modulation schemes in the following section.

 ABT f/2fc n +  4fc T (1 + (f/2fc ))  sin(4fc T ) × sin(4fc ) + −1 , 4fc T

(12)

where  is the double-sided power spectral density of receiver white noise, n is the normalized value of TIPA

The proposed TIPA scheme for BPSK modulation has been designed for VLSI implementation and examined through computer simulation using analog VLSI simulation software of Cadence with 0.18 m CMOS technology. Subsequently, the TIPA-BPSK design has been extended further to propose the designs for TIPA-QPSK and TIPAQAM modulation schemes. As mentioned earlier, the VCO is required to be operated at different desired frequencies governed by the TIPA-BPSK/QPSK/16-QAM requirements, the lowest of which is the carrier frequency fc prevailing most of the time. When phase shift is intended following a data transition, the VCO is driven at a higher frequency that is appropriate to catch up the required phase in the given interval of time during the TIPA by applying a narrow-pulse at its control input. The narrow-pulse is generated during the TIPA intervals following each data transition by using delay-and-EXOR (DEO) operation on the baseband data stream. Fig. 8 shows the transistor-level circuit diagram of narrow-width pulse generator block. It consists of two functional units, the first one provides a delay to the input data, where the delay is same as the required TIPA interval, which is fed along with the original data to the inputs of the second unit, which is an EXOR circuit and thus produces a narrow-pulse of width equal to the TIPA interval at every transition of input data stream. The delay block consists of a couple of inverters, loaded with capacitance. The inverter should always be used in pairs to maintain the same phase with input data. The inverters provide the delay of relatively large amount, whereas loading capacitance Cp and biasing voltage Vbias are used for fine adjustment of the delay. For longer delay, we can use more number of inverter pairs. This narrow-width pulse is used as the input to the VCO. Fig. 9 shows the transistor-level circuit diagram of the VCO that employs a current-starved architecture with three controllable delay stages consisting of transistors M3 through

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753

Input Data

747

Output Data

Cp

Vbias

Inverter Pair

EXOR

Delay Block

Fig. 8. Transistor level schematic for narrow-width pulse generator block.

M3

M4

M5

M6

M7

M8

M1

M0

Ouput M9

M10

M11

Signal

From Current Source M2

M12

M13

M14

Fig. 9. Transistor level schematic for the current starved VCO along with the current-to-voltage converter block.

M14. While the control voltage input directly drives the pMOS transistors (M3, M4 and M5), the appropriate voltage level required to drive the nMOS transistors (M12, M13 and M14) is derived with the help of M1-M2 combination. The transistors M0-M1-M2 form a current-to-voltage converter which is employed to achieve better controllability. The transistor dimensions (W/L ratios) are chosen such that the frequency of the VCO under normal condition is equal to the carrier frequency (fc ), that shoots up to (fc + f ) under pulsed condition during the TIPA interval. The voltage level-translation unit ensures that appropriate voltages are fed to the control input of the VCO for generating proper frequencies (i.e., fc and fc + f ). Since the VCO is operated only at discrete frequencies and not at continuously variable ones, linearity requirement is not stringent,

although controllability improves with linearity. Thus, in our scheme, current is chosen as the independent variable that controls the frequency of the VCO, and in that sense, the combined structure may be more appropriately called a current-controlled oscillator (CCO). Using this CCO, we have designed our proposed TIPA scheme for different modulation schemes.

4.1. TIPA-BPSK The transistor level schematic of VLSI design for the TIPA-BPSK is shown in Fig. 11. The narrow-width pulse is generated during TIPA interval following every data transition by a narrow-width pulse generator block (as shown in

748

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753

Fig. 8). It may be noted that, for designing BPSK, two distinct control-current values are required to drive the CCO at two desired frequencies. In our design, two current sources (CS0 and ACS) are employed as shown in Fig. 10, wherein the reference current source (CS0) remains operational always, to keep the CCO running at the carrier frequency fc . During a TIPA interval following a data transition, additional current is injected from the other current source (controlled by the narrow-width pulse generator block, derived with data stream and DEO block appropriately) to the CCO to increase its instantaneous frequency by f . This additional current source (ACS) becomes active only when there is a phase shift requirement of 180◦ during the TIPA interval. Table 4 gives the values of simulation width and the length of every transistor, which have been used to generate TIPA-BPSK output using Cadence software with 0.18 m CMOS technology. The current values required for satisfying phase-shifting condition for TIPA-BPSK scheme are shown in Table 5 along with the sizes of the transistors used in CS0 and ACS for different values of n . As expected, it is evident from Table 5 that, ACS should provide higher

VCO/ Input Data

QA

Narrow Width A Pulse Generator

BPSK Output

CCO SW

Additional Current Source (ACS)

Reference Current Source (CS0)

Fig. 10. Block diagram for the TIPA-BPSK scheme, used in VLSI simulation. Table 4. Design parameters for the TIPA-BPSK scheme (Fig. 11) MOS name

Simulation width (nm)

Simulation length (nm)

M0, M1 M2 M3-M8 M9-M14 M15 M16

2800 1880 1280 1040 2080 1080

740 1020 340 460 1020 1020

current for lowering n to satisfy phase-shifting condition for TIPA-BPSK. Therefore, there is always a need to have a trade-off between the current value of ACS and n . In our design, we consider that the value of n = 0.002 and the corresponding current value is 195.3 A, which produces a frequency of 967.37 MHz during TIPA interval. It may be noted that, the same ACS can be programmed to serve for the other values of phase shifts as well (say, 90◦ , 270◦ and so on) with different settings of incremental currents for m-ary phase modulators, leading to a simple adaptive hardware for multi-level digital phase modulation systems. Fig. 11 shows the transistor level schematic of TIPA-BPSK scheme, where two nMOS transistors M15 and M16 provide the biasing voltage to the current sources.

4.2. TIPA-QPSK It may be noted that for designing QPSK, provisions for having four different phase shifts must be there while only two of them (including 0, which implies no shift at all) are sufficient for designing BPSK. Therefore, four distinct control current values are required to drive the CCO at the four desired frequencies during TIPA transition (note that, TIPA interval  is kept fixed at the same value for all phase shift). In our design, four current sources are employed as shown in Fig. 12, where the reference current source (CS0) remains operational always to keep the CCO running at the carrier frequency fc . During data transition, additional current is injected from the additional current sources (ACS1, ACS2 and ACS3), controlled by the data bits appropriately, to the CCO to increase its output frequency. To control the four current sources, for proper functioning according to data sequences, at least two binary control pulses are required. These two control pulses A and B of fixed width  are derived from QA and QB , which are again generated from the input data using a conventional odd/even-bit identifier block. These two narrow pulses (A and B) are used to control the currents from ACSs. ACS1 and ACS2 become active when the phase-shift requirement becomes 90◦ and 180◦ , respectively. For a phase shift of 270◦ , in addition to ACS1 and ACS2, ACS3 comes into action to satisfy the current requirement for 270◦ which is more than the currents required for shifting by 90◦

Table 5. Design parameters of ACS to satisfy phase-shifting condition of TIPA-BPSK scheme (Fig. 11) for different values of n Design parameters

MOS parameters for CS0

Simulation width (nm) Simulation length (nm) Current (A) Frequency (MHz)

2000 580 126 900.01

ACS

n = 0.1

n = 0.01

n = 0.002

n = 0.001

800 1100 135.7 906.71

800 1020 140.7 913.54

900 860 195.3 967.37

1020 780 250.5 1001.21

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753

M15

M3

M4

M5

M6

M7

M8

749

M1

M0

BPSK M16 CS0 M9

Output

M11

M10

SW1 Input Data

Narrow Width Pulse Generator

M2

M12

M14

M13

ACS

Fig. 11. Transistor level schematic for TIPA-BPSK scheme, used in VLSI simulation. QPSK Output

VCO / Narrow Width Pulse (QA) Generator

ODD

Input Data

ODD/ EVEN Bit Selection Block

EVEN

(QB)

Narrow Width Pulse Generator

CCO

A

B

ACS0

ACS1

ACS2

CS0

Fig. 12. Illustration of the basic TIPA-QPSK scheme. Narrow Width Pulse Generator

QA

Previous Constellation Point (ZYXW)

Input Data

Serial to Parallel Converter

Narrow Width Pulse Generator

QB

Control Signal Generator Block

B

Narrow Width C Pulse Generator

QC

Q

A

Narrow Width

D

Current Source Array and Selection Logic (include CS0 and ACSs )

D

Pulse

Q

Narrow Width Pulse Generator

E

VGA

a

b

Amplitude Control Logic

Generator

Present Constellation Point (SRQP)

Output Signal

VCO / CCO

E

P Q

Fig. 13. Logic diagram for TIPA-16-QAM scheme.

and by 180◦ taken together. Table 2, as discussed earlier in Section 2 provides an idea about the required frequencies for the desired phase shifts for a particular value of .

4.3. TIPA-16-QAM The block diagram of VLSI design scheme for the TIPAbased 16-QAM is shown in Fig. 13, which consists of six different functional blocks. Beside the narrow-width pulse

generator blocks and VCO which are used in the earlier TIPA scheme for QPSK, the other four are control signal generator block, current source array, selection logic block and a variable gain amplifier (VGA). The control signal generator block controls the five narrow-width pulse generator blocks in such a way, that the appropriate pulse generators are selected from a group of five, to generate narrow pulses of TIPA interval following each data transition. In a standard constellation, as shown in Fig. 3, any particular constellation point differs from its neighbouring points typically by

750

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753

Z

X

Y

W

S

Q

R

A

P

Additional Current Sources (19 in Numbers)

B C

To VCO / CCO Selection Logic Reference Current Source (CS 0)

D E

Z

Z

Y

X

W

X

19 Lines

W

Fig. 16. Block diagram for the current-source array and selection logic unit in VLSI simulation. Subtractor P Q

a

Q D QC QB QA

QE

Fig. 14. Block diagram for control-signal generation of 16-QAM.

1111

1101

1100

1011

1000

Q

a

b

0 0 1 1

0 1 0 1

0 1 1 1

1 0 0 1

Amplude Level Level 1 Level 2 Level 2 Level 3

Fig. 17. Circuit diagram for the amplitude control logic unit in VLSI simulation.

Q 1110

b

P

1010

1001

I 0001

0000

0100

0111

0010

0011

0101

0110

Fig. 15. Typical constellation diagram of a 16-QAM after mapping.

one bit to facilitate error control mechanism. A small combinational circuit, as shown in Fig. 14, is employed to map the standard constellation in Fig. 3 into an intermediate one, as shown in Fig. 15, where the binary-weighted values associated with the constellation points increase monotonically as one moves in a counter-clockwise direction. This rearrangement in the constellation enables the usage of the conventional arithmetic (viz., subtraction) to control the current sources for implementing 16-QAM effectively. The implementation of TIPA-16-QAM has been carried out in two steps. In the first step, we have designed the control signal generation block and in the second step, a selection logic has been employed to control the current sources. In these two steps, the current sources are controlled in a structured manner. As shown in Fig. 14, five control signals QA , QB , QC , QD and QE are generated, which are sufficient to control 19 ACSs with a reference current source (CS0). In 16-QAM, the various required phase shifts (20 in number) can be selected by 19 ACS’s in the current-source array (Fig. 16), which are controlled by five narrow pulses A, B, C, D and

E derived from QA , QB , QC , QD and QE . Fig. 16 shows that 19 ACSs and CS0 are controlled by five narrow pulses A, B, C, D and E with the help of a selection logic circuit and used to feed the appropriate current level to VCO. Thus, QA , QB , QC , QD and QE become active for realizing the desired value of phase-shift, while the two amplitude control bits P and Q (the same one, which are shown in Fig. 3) are used to control the amplitude of the carrier waveform of VCO with the help of VGA control by amplitude control logic unit. The circuit diagram of amplitude control logic unit is shown in Fig. 17, wherein the table provides the logic to control the three different amplitude levels. Table 6 provides the current values corresponding to the required phase shifts with n = 0.002. As mentioned earlier, in order to examine the feasibility of TIPA scheme, we have carried out a sample simulation of basic TIPA scheme for BPSK. In our simulation, the desired phase shift (i.e., 180◦ for TIPA-BPSK) has been realized with a supply voltage of 1.8 V at a temperature of 27 ◦ C with typical process parameters using Cadence software for 0.18 m CMOS technology. However, in a practical system, variation of supply voltage, temperature and process will make the phase shift wander around the desired value. In order to have a reliable system operation, the designer has to ensure that the circuit satisfies the given specification despite the variation of the ambient temperature, supply voltage and the process variation during device fabrication. From our preliminary study, it is observed that the drift of phase shift is not significant (less than ±6◦ ) with supply voltage variation within the considered voltage range of 1.4 to 2.0 V (for a nominal supply voltage of 1.8 V), while the phase shift shows stronger sensitivity (upto ±35◦ ) to temperature changes in

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753

751

Table 6. Current values for satisfy phase-shifting condition of TIPA-16-QAM scheme with n = 0.002 Phase shift

Frequency (MHz)

Current value (A)

Phase shift

Frequency (MHz)

Current value (A)

0 26 36 53 63 90 116 126 143 153

900 906 913 919 923 933 943 947 953 957

126 (CS0) 135.7 (ACS1 = 9.7) 140.3 (ACS2 = 14.3) 146 (ACS3 = 20) 150 (ACS4 = 24) 155 (ACS5 = 29) 165 (ACS6 = 10) 169 (ACS7 = 14) 177 (ACS8 = 22) 183 (ACS9 = 27)

180 206 216 233 243 270 296 306 323 333

967 977 981 987 991 1001 1011 1015 1021 1025

195 209 215 224 232 250 268 280 295 315

the range of −20 ◦ C to 80 ◦ C. Impact of process variation has not been examined in the present work.

5. Adaptive modulation scheme The goal of an adaptive modulation scheme is to ensure that the most efficient modulation is always used when channel conditions change, based on a mode-selection criterion. In the following, we propose a comprehensive adaptive modulation scheme with triple options: BPSK, QPSK, and 16-QAM. It may be noted that, link adaptation requires feedback information from the receiver to transmitter about the link conditions and in our work we assume that this information is available from the relevant receiver end after due measurements on its short-term signal-to-noise ratio. To design the adaptive modulation scheme with the triple options of BPSK, QPSK, and 16-QAM, we have combined the VLSI design of all modulation schemes, which have already been discussed. The block diagram of the adaptive modulation scheme under consideration is shown in Fig. 18. As shown in Fig. 18, a DEMUX is used to select the mode (modulation scheme, in this case) based on adaptive decision logic depending on the feedback information and pass on the appropriate data to subsequent stages of control circuit for a given modulation scheme. For different modulation schemes, appropriate pulse-generators are selected from a group of five pulse generator blocks to generate streams of narrow pulses of required width following each data transition. These control pulses of appropriate width and timing are in turn used to choose appropriate current sources from the current-source array, using an appropriate logic. These current sources having been selected for the desired modulation scheme, are fed into the CCO, for the necessary TIPA operations following data transitions. In the following, we describe each functional block of Fig. 18 in further details. First, we describe the function of the control logic (as shown by the dotted rectangle in Fig. 18), for selecting as well as enabling/disabling the pulse generator blocks, i.e., current sources from the current-source array. The outputs

(ACS10 = 69) (ACS11 = 14) (ACS12 = 20) (ACS13 = 29) (ACS14 = 37) (ACS15 = 26) (ACS16 = 44) (ACS17 = 56) (ACS18 = 71) (ACS19 = 91)

of this control unit are QA , QB , QC , QD and QE , which are passed onto the CCO through narrow pulse generators for operation with the desired modulation scheme. For this control logic of different modulation schemes, we introduced the control signal generation block for each modulation, as described earlier. In particular, for BPSK, one ACS (in addition to CS0 which sets the CCO to operate at the carrier frequency) has to be controlled for the desired phase shift of 180◦ when there is a data transition (Fig. 10). This is realized by the pulse A of fixed width  (TIPA interval), which is derived from input data only in QA by DEO operation (Fig. 2); thus, in BPSK mode, QA becomes active and others (QB , QC , QD and QE ) remain inactive/disabled. Therefore, for BPSK, when there is a data transition, pulse A enables the appropriate current source for CCO to realize the desired phase shift of 180◦ during the TIPA interval, while current from the other sources are blocked. Similarly, as mentioned earlier, for QPSK, four different phase-shifts (including zero phase shift) are provided by three ACSs and a CS0, which are controlled by two control pulses A and B of the fixed width  are derived from QA and QB (Fig. 12). Thus, in QPSK mode transmission, only QA and QB , become active and others remain inactive/disabled. In 16-QAM, the various required phase shifts (20 in number including zero phase shift) can be selected by 19 ACSs and the CS0 in the current-source array, which are controlled by five narrow pulses A, B, C, D and E derived from QA , QB , QC , QD and QE . Thus, QA , QB , QC , QD and QE become active for realizing the desired value of phase-shift, while the two amplitude control bits P and Q are used to control the amplitude of the carrier waveform of CCO. Although, QA and QB carry control information for the BPSK-QPSKQAM and QPSK-QAM, respectively, the preceding combinational logic stages enable them to pass on one of the three possible modulation schemes at a time to the subsequent stage (i.e., narrow-width pulse generators). In the following block, selection of appropriate ACS from the current source array is made by using appropriate selection logic (as used in Fig. 16 for QAM), which in turn provides the necessary frequency increment f during TIPA interval to

752

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753

BPSK

Input Data

QPSK 1:3 DEMUX (Modulation Mode Selection Block)

QA

ODD/EVEN Bit Selection Block

Odd QB

Even

QC ZYXW

QAM Serial to Parallel Converter Adaptation Decision Logic Based on Feedback From Rx

Control Signal Generator Block

QD

QE

RSQP

Narrow Width Pulse Generator

Narrow Width Pulse Generator Narrow Width Pulse Generator Narrow Width Pulse Generator Narrow Width Pulse Generator

A

B

C

Current Source Array and Selection Logic (include CS0 and ACSs)

VCO / CCO

VGA

a

Output Signal

b

D Amplitude Control Logic E

P Q

Fig. 18. Block diagram for the adaptive modulation scheme.

satisfy the phase-shifting requirement and the CS0 remains operational always to keep the VCO running at the nominal carrier frequency fc . Immediately after a data transition, additional current is injected from ACSs, controlled by the data bits, to the VCO to increase its output frequency during TIPA interval. Finally, it may be noted that the two control bits (P and Q) in Fig. 18 would control the amplitude of CCO for 16-QAM modulation, but for BPSK and QPSK the CCO output would maintain a constant envelope waveform.

6. Conclusion An adaptive modulation scheme with triple options of BPSK/QPSK/16-QAM has been examined for application in wireless networks using a phase-shifting technique for digital phase modulation. The proposed phase-shifting scheme uses TIPA, employing a VCO and narrow pulse generator blocks. The spectrum and BER analysis of TIPABPSK scheme show that it approaches ideal BPSK performance for lower value of TIPA interval, which needs higher value of f and thus a trade-off is necessary. Simulation of a TIPA-BPSK modulation has been carried out using Cadence software for 0.18 m CMOS technology, ensuring the feasibility of the proposed modulation technique using TIPA. The adaptive modulation scheme with triple options has been designed for possible VLSI implementation with five number of pulse generator blocks. The number of pulse generator blocks depends on the modulation schemes under consideration. The proposed scheme offers a simple and scalable means for VLSI implementation, while offering a performance similar to other conventional schemes. This feature will prove useful for LA realization, as the LA-based systems would demand multiple modulation schemes to remain available concurrently on the same chip.

References [1] Jakes WC. Microwave mobile communication. 2nd ed., IEEE Press; 1994. [2] Rappaport TS. Wireless communication: principle and practice. 2nd ed., Pearson Education; 2003. [3] Catreux S. et al. Adaptive modulation and MIMO coding for broadband wireless data networks. IEEE Commun Mag 2002; 108–15. [4] IEEE Pers Commun. Special issue, Smart Antennas, 1998; 5(1). [5] Foschini GJ, Gans MJ. On limits of wireless communications in a fading environment when using multiple antennas. Wireless Pers Commun 1998;6:311–35. [6] Gesbert D. et al. Technologies and performance of non lineof-sight broadband wireless access networks. IEEE Commun Mag 2002;40(4):86–95. [7] Gesbert D. et al. From theory to practice: an overview of MIMO space time coded wireless system. IEEE J Select Areas Commun 2003;21(3):281–302. [8] Foschini GJ. Layered space-time architechture for wireless communication. Bell Labs Tech 1996;1:41–59. [9] Nanda S, Balachandran K, Kumar S. Adaptation techniques in wireless packet data services. IEEE Commun Mag 2000; 54–64. [10] Mazumder SR, Waterman RC. A novel 6 to 18 GHz 180degree bit phase shifter configuration having very small amplitude and phase errors. IEEE MIT-S Int Microwave Symposium Digest, 23–27 May, vol. 1, 1994. p. 83–6. [11] Goldfarb ME, Cole JB, Platzker A. A novel MMIC biphase modulator with variable gain using enhancement-mode FETs suitable for 3 V wireless applications. IEEE Microwave and Millimeter-Wave Monolithic Circuits Symposium, 22–25 May, 1994. p. 99–102. [12] Chen JH, Tsao HW. BPSK modulator using VCCS and resonator without carrier signal and balanced modulator. Electron Lett 1997;33(5):1286–7. [13] Gazsi L. On the reduction of hardware in digital modulators. IEEE Trans Commun 1979;COM-27(1):1575–86.

R. Mahapatra et al. / Int. J. Electron. Commun. (AEÜ) 62 (2008) 740 – 753

[14] Fettweis A. Transmultiplexers with either analog conversion circuits, wave digital filters, or SC filters A review. IEEE Trans Commun 1982;COM-30(7):221–8. [15] Kamilo Feher, Digital communications: microwave application. Prentice-Hall of India; 1987. [16] Mahapatra R, Dhar AS, Datta D. On feasibility of a multiplier less phase-shifting scheme for digital phase modulation and its VLSI implementation. Int J Electronics 2007;94(2): 171–81. [17] Datta D. Analysis of a new technique for optical DPSK transmission without external modulation. Int J Opto 1993; 8:451–7. Rajarshi Mahapatra is a Research Scholar at Department of Electronics and Electrical Communication Engineering of Indian Institute of Technology, Kharagpur. He obtained B. Tech. and M. Tech. degrees in Optics and Optoelectronics from the Department of Applied Physics, Calcutta University, Kolkata in 1998 and 2000, respectively. His current research interests include adaptive modulation, space–time coding and emerging techniques in wireless networks. Anindya Sundar Dhar received his Bachelor degree in Electronics and Telecommunication Engineering from Bengal Engineering College, Sibpur, India in 1987. In 1989, he received his M. Tech. degree in Integrated Circuits and Systems Engineering from Indian Institute of Technology, Kharagpur. He received his Ph.D. degree from the same Institute in 1994, where he is presently serving as an Assistant Professor in the Department of Electronics and Electrical Communication Engineering. His research interests include VLSI for communication and DSP architectures for real time signal processing.

753

Debasish Datta received his B. Tech. degree in 1973 from the Institute of Radiophysics and Electronics, Calcutta University, and M. Tech. and Ph.D. degrees from IIT Kharagpur in 1976 and 1986, respectively. He has been engaged in teaching and research at IIT Kharagpur in the Department of Electronics and Electrical Communication Engineering during the last 28 years and currently he serves therein as the Head of the Department. During the period 1999–2002, he also served as the Chairman of G.S. Sanyal School of Telecommunications at IIT Kharagpur. In the early phase of his career, he worked for Transmission R&D Division in Indian Telephone Industries, Bangalore, during 1976–1978, and in Production Management Division of Audio and Intercom Systems of Philips India Ltd, Calcutta, during 1980–1981. During his stay at IIT Kharagpur, he was awarded Indo-US Fellowship by the Department of Science and Technology, Government of India, and the United States Agency for International development, to carry out research at Stanford University for one year during 1992–1993 in the area of Coherent Optical Communications. Thereafter, he visited University of California at Davis during 1997–1999 and Chonbuk National University, South Korea, during 2003–2004 to carry out collaborative research in the area of optical networking. He received Sir J.C. Bose Premium Award in 1985 during his doctoral work from the Institution of Radio Engineers (IERE), UK, for a paper on Optical Receiver in the Journal of IERE. In the recent past, he served as Guest Editor for IEEE Journal of Selected Areas in Communication for the January-2002 Special Issue on WDM-based Network Architectures, and presently he serves as an Editor for the Elsevier Journal of Optical Switching and Networks. He has been and also serves presently in the technical program committees of several national as well as international conferences in the area of optical communications and networking. His current research interests include survivable wavelength-routed optical networks, optical access networks, optical burst switching and link adaptation in wireless networks employing diversity aided adaptive modulation and power control schemes.

Adaptive digital phase modulation schemes using ...

Department of Electronics and Electrical Communication Engineering, Indian Institute of Technology, Kharagpur 721 302, India. Received 16 November ... all types of wireless services (voice, data, and multime- dia) have fuelled the need ... services, such as code division multiple access (CDMA) schemes like CDMA2000 ...

379KB Sizes 0 Downloads 203 Views

Recommend Documents

Joint Adaptive Modulation and Switching Schemes for ... - IEEE Xplore
Email: [email protected]. Tran Thien Thanh ... Email: thienthanh [email protected] ... the relaying link even it can provide better spectral efficiency.

Estimation of the phase derivative using an adaptive window ...
rivative contains the information regarding the measur- and. During recent years, there has been an increased in- terest in developing methods that can directly estimate the phase derivative from a fringe pattern because the phase derivative conveys

Joint Adaptive Modulation and Distributed Switch-and ...
bit error rate (BER) in fading channels [9]. Recently, the effectiveness of adaptive modulation in cooperative wireless communication systems in which power ...

MODULATION FORENSICS FOR WIRELESS DIGITAL ...
... College Park, MD 20742 USA. Email: {wylin, kjrliu}@eng.umd.edu ... achieves almost perfect identification of the space-time coding, and high accuracy rate of ...

Implementing Implementing Digital modulation ...
Several available FPGA programming softwares can be broadly classified as Xilinx's VHDL (Verilog hardware description language), National Instruments. LabVIEW (Laboratory virtual instrument engineering workbench), and MATLAB etc. By considering those

Adaptive Modulation for Distributed Switch-and-Stay ...
Posts and Telecommunications Institute of Technology. Email: [email protected]. Abstract—In this letter, we investigate the performance of distributed ...

Direct adaptive control using an adaptive reference model
aNASA-Langley Research Center, Mail Stop 308, Hampton, Virginia, USA; ... Direct model reference adaptive control is considered when the plant-model ...

Direct Adaptive Control using Single Network Adaptive ...
in forward direction and hence can be implemented on-line. Adaptive critic based ... network instead of two required in a standard adaptive critic design.

Phase Adaptive Integration for GNSS Signals
In the GNSS signal acquisition operation, long time coherent integration is a problem if the coherence of the baseband signal is not guaranteed due to the.

Genetic Algorithm based Phase-Only Nulling in Adaptive Antennas
A simple alternative to complex hardware design of adaptive antennas along with optimizing SNR is proposed by the application of genetic algorithms. Adaptive antennas are antenna arrays with phase angle associated with each element of the array. The

h-adaptive grid generation - phase field modelling ...
This work deals with the application of adaptive grid generation technique to ..... A suitable data structure is then generated by traversing the region and finding ...

Cheap White Mini Frequency Modulation Fm Radio Digital Signal ...
Cheap White Mini Frequency Modulation Fm Radio Digi ... Earphone Radio Free Shipping & Wholesale Price.pdf. Cheap White Mini Frequency Modulation Fm Radio Digit ... h Earphone Radio Free Shipping & Wholesale Price.pdf. Open. Extract. Open with. Sign

Recognizing Stress Using Semantics and Modulation of Speech and ...
Our experiments are run on an audiovisual dataset with service-desk interactions. The final goal is ..... the annotation software Anvil [22] used by the raters.

Color Schemes
Name. Period ______. Color Schemes. Define Color Scheme: 1. The first color schemes is: Definition: Examples of colors: 2. The second color scheme is:.

Comparing single and multiple carrier schemes using ...
... [email protected], {marcioft, mdm}@mackenzie.br ... The OFDM systems divide the available band- ... Block diagram of a typical COFDM system; CP: Cyclic.

adaptive modulation in time and space for space time ...
For Cases 1, 5, and 9, we may use the best code ... Cases 5 and 9: Use the best non-adaptive encoders ..... Broadband Wireless Data Networks," IEEE Comm.

Fast-switching phase gratings using in-plane ...
Dec 20, 2011 - Related Articles ... There are alternative LC electro-optic effects that exhibit .... intensity in the zeroth order is decreased as energy is trans-.

Low Phase Noise Wideband VCO using MEMS
grown market demand and technology advancement. In such wireless High-speed communication circuits and systems,. Balasaheb S. Darade is final .... MEMS tunable lower capacitors have advantages of lower loss, larger tuning range and ...

Fringe demodulation using the two-dimensional phase ...
Received June 22, 2012; revised September 4, 2012; accepted September 4, 2012; posted September 5, 2012 (Doc. ID 171145); published October 11, 2012. The Letter proposes a method for phase estimation from a fringe pattern. The proposed method relies