An Innovative Receiver Architecture for ... - Semantic Scholar
An Innovative Receiver Architecture for Autonomous Detection of Ultra-Wideband Signals Majid Baghaei Nejad, Li-Rong Zheng Laboratory of Electronics and Computer Systems, Department of Electronics, Computer and Software Systems Royal Institute of Technology (KTH), Electrum 229,SE-164 40 Kista-Stockholm, Sweden E-mail: {majidbn|lrzheng}@imit.kth.se Abstract— Ultra wideband radio is an emerging wireless standard that uses sub-nanosecond pulses to transmit data, resulting in several GHz bandwidths. The problem of generating a synchronized template respect to the received signal grows in complexity as the signal bandwidth increases. In this paper, an innovative, low cost, non-coherent receiver architecture is proposed for autonomous detection of ultra wideband signals. The new receiver will self-generate a synchronous template and hence, no transmitter-reference synchronizer is required. We validate its performance via simulations compared with coherent receivers and conventional non-coherent receivers, the new architecture is found much more robust to timing noise and hence greatly facilitates the synchronise problem in UWB receiver. Key word- Impulse Radio, Ultra Wideband, Non-coherent
I. INTRODUCTION Recently, UWB systems are paid attention by academic research group and industries in short-distance, low power, high data rate and low cost applications. This attention increases since the American Federal Communication Committee (FCC) released the UWB spectral in FEB 2002. Because of the extremely low power density of UWB, FCC allows it to operate over the top of the bands. It is helpful to solve the frequency allocation problem that often limits high data rate wireless communication systems [1]. The wide bandwidth of UWB offers several advantages over the narrow band systems, such as low-power, impulse-like, very high data rate, diversity against multi-path, base-band transition and large number of users [2], [3]. With these advantages, UWB system can be considered as a powerful candidate for Wireless Personal Area Network (WPAN). It is also considered for low power and low cost RFID (radioidentification) and sensor systems. There are two general strategies for Ultra Wideband receivers. Optimal Rake coherent receivers use complete channel estimation to resolve paths. It is a high performance technique but also relatively expensive. On the other hand, sub-optimal noncoherent receivers basically involve energy detectors and hence relax from channel estimation [2], are attractive for low cost and low data rate applications. The coherent receivers have a good performance however, duo to the extremely number of paths and signal distortion by antennas and propagation environment, synchronization and general channel estimation in UWB
0-7803-9390-2/06/$20.00 ©2006 IEEE
schemes are critical in design [4], especially in dense multi-path and mobile applications [5]. This difficulty arises from the ultra-short pulses with low amplitude used in UWB radio. Therefore, the coherent receivers are very complex and expensive. In order to achieve low cost and low complexity transceivers, non-coherent detection has been proposed. Non-coherent receivers offer a simple and cheap architecture for the receiver in cost of losses in BER performance and lower data rate that have limited their applications [6-7]. Fundamentally non-coherent receivers need a higher SNR (signal-to-noise ratio) and therefore they are employed in short distance and low data rate applications such as RFID and sensor systems. Because of simplicity and low cost, many researches have worked on this scheme and different architectures have been investigated such as Transmitted-Reference UWB or TRUWB [8] and Differential TR-UWB systems [9]. Generally, the non-coherent receivers employ a replica of the received signal delayed by Td as the template and make a correlation between this template and the received signals. Td is a fixed delay depends on the time interval between two pulses in one doublet. If there is a mismatch between the delay value and the time interval, performance of the receiver will be degraded. This timing mismatch can be caused from process variation or from inaccuracy and jitter of the oscillator in transmitter. Another approach that relaxes from the channel estimation and synchronization is the asynchronous transceiver. Data is carrying by OOK modulation. The receiver uses an energy detector to detect the data. [10] This architecture has also poor BER and low data rate and needs high SNR input. By using non-coherent or asynchronous approach, frame synchronization is simplified to symbol synchronization. In this paper, we propose a novel non-coherent receiver that can autonomously detect received signals. The new architecture is a modification of TR-UWB system in which another multiplier is used to multiply the signal instead of using the delay component. We validate the performance of this new receiver by simulations in Matlab and compare it to traditional coherent receiver and TRUWB systems. We found that the new architecture has significant improvements in tolerance of timing noise. We will start from the channel and signal models of the UWB systems used in this paper in Section 2, then describes the autonomous receiver in section 3. Analysis and simulation results are given in section 4. Section 5 is the conclusion.
2589
ISCAS 2006
II. SIGNAL AND CHANNEL MODEL Information in UWB system is typically transmitted using a collection of pulses with short width and low duty cycle. Each user is assigned a different Pseudo-noise (PN) sequence that is used to encode the pulse in BPSK or PPM. In order to clearly compare with other architectures, in this paper we consider the BPSK modulation and single user and non multi-path environment. For BPSK the code modulates the polarity of the pulses. The received signal can be written as follow: ∞ N f −1
r (t ) = ∑ i =0
∑
bi p (t − (i.N f + j )T f ) + n(t )
Figure 2. Autonomous receiver
The output of the integrator can be expressed as:
an =
(1)
j =0
where P(t) is the monocycle template pulse. The second deviation of Gaussian pulse with 250ps width used as the template in this paper is shown in Fig. 1. T f is the duration of each frame, bi = ± 1 is the data, N f is the number of frame that is used to transmit one bit that can be between 1 to several hundred and n(t) is AWGN. III. SYSTEM DESCRIPTION A. Proposed Autonomous Receiver Scheme The structure of our autonomous receiver is illustrated in Fig. 2. As can be seen, this is a modification of the TRUWB system. The template waveform employed in the demodulation process consists of a squared replica of the received signal. Similar to coherent receiver, a band-pass filter h(t) of bandwidth B is employed to reduce the effect of noise in the demodulation process. It should be mentioned that this filer is very critical compared with that in coherent receivers. This is because, coherent operation is able to moderate the noise and maximising the SNR if a template pulse matched to the received signal is available. Therefore the filter can be designed in such a way that dose not effect the received signal. However in the non-coherent receiver, template waveform is very noisy. The bandwidth of the filter must be chosen as a trade-off between the signal energy and noise reduction. In order to synchronize the two signals at the input of the mixer, a fixed delay line is chosen to compensate the mixer delay.
( n +1)T f N f 3 1 nT f N f
∫r
(t )dt
(2)
(3) where r1 (t ) = r (t ) * h(t ) This architecture has a number of attractive properties: • Same as in other non-coherent receivers, the synchronization is only in symbol level, no frame level synchronization is needed. Therefore the complexity of the receiver is reduced, especially suitable for low data rate applications, where often T s >> T f . • Multi-path gathering is achieved. Since all components in the received signal go trough the same system and channel, the produced template pulse provides a perfect (but unfortunately noisy) template to match the received signal without any channel estimation or the need for a rake receiver with many branches. • Since the template pulse is as the same as received signal, changing in channel does not affect the receiver performance. This can be a significant advantage for systems operating in a highly mobile environment. • Because the template pulse is produced from the received signal, this architecture does not suffer from jitter and phase noise in the transmitter oscillator. It is very robust to timing noise. • In this architecture the transmitter does not send any extra pulse for synchronization purpose. It therefore improves the performance in case of multi-user interference and inter-frame interference in TR-UWB system [11]. • The transmitter power consumption will be reduced because it does not send any extra pulses. This can also be a significant advantage for self-powered systems such as RFID and passive sensor systems. B. Other Receiver Architectures for Comparison In order to compare the performance of our proposed receiver and traditional receivers, we choose two other architectures: coherent receiver and TR-UWB receiver. For a fair comparison, same signal and channels models are used.
Figure 1. Monocycle pulse
B.1. Coherent Receiver A general coherent receiver used in this paper is shown in Fig. 3. A template signal, generated according to the information acquired by bit and code synchronization and possibly channel estimation, is multiplied by the received signal. In the end the result passed trough an integrator and decision block.
2590
Figure 4. Transmitted-Reference receiver Figure 3. Coherent receiver
The output of the integrator can be written as: ( n +1) N f .T f
∫ r (t ).x(t − n.N
an =
f
.T f )dt
(4)
n. N f .T f
where
x(t ) =
N f −1
∑ p(t − m.T
f
)
(5)
m=0
p(t) is the template pulse and r(t) is the received signal. B.2. Transmitted-Reference UWB System A transmitted reference receiver is simulated for comparison. Its structure is illustrated in Fig. 4. The signal is multiplied by a delayed version of itself and the result integrated over the symbol period Ts , which can consist of many frames, each of duration T f , in low data rate applications. A threshold block makes a decision on the output of integrator to decode the data for the current symbol. As we described before, the design of the frontend filter is critical in TR-UWB system in order to reduce the effect of noise. Each bit is transmitted by a pair of pulses as:
S1 (t ) = p (t ) + p (t − Td ) S 0 (t ) = p (t ) − p (t − Td )
(6)
where S1 ( t ), S 0 ( t ) are the signals transmitted for an information bit 1 and bit 0 respectively and p(t) is the pulse shape. Td is a fixed value and chosen as
TP < Td < T f where TP is the pulse duration. The output of the integrator can be expressed as:
The FCC regulations are not considered here. In coherent and our autonomous receivers, PBSK modulation and in TR-UWB two-points-signal are utilized. Performances of these three receivers are evaluated in term of bit error rate (BER) versus signal to noise ratio (SNR). A front-end band-pass filter is considered with no distortion on the pulse shape. Both of The receiver and noise bandwidth are assumed to 5 GHz and the noise figure of the receiver is 8dB. Frame period is set to 10ns and the number of pulse per bit is changed to estimate the performance in different data rates. In order to have an accurate estimation of BER, 10000 bits are applied to each receiver. Fig. 5 shows the BER performance versus the SNR for three structures. It is clearly seen that the coherent receiver has highest performance, but as mentioned before, it is expensive and complex. The autonomous receiver and TR-UWB show similar performance. It must be mentioned that in this case, no timing and phase noise have been assumed and more improvements are expected when this effect is considered, as will be shown in next figures. Second, we introduce a time mismatch. In coherent receiver, this is a time skew in synchronization and in TRUWB and our new architectures this is the deference of Td from expected value. In autonomous receiver the delay value equals to the delay of the mixer. The accuracy of this delay is only related to the receiver and it is much easier to control during manufacture. Fig. 6 compares the performance of the receivers in the presence of time mismatch. As can be seen, any mismatch degrades the performance.
( n +1) N f .T f
an =
∫ r (t ).r (t − T
1 n . N f .T f
1
d
).dt
(7)
where r1 (t ) is the filtered signal. IV. RESULTS AND DISCUSSIONS The receiver structures described in previous section are simulated in MATLAB and the results are compared. An AWGN channel is considered as the channel model. As mentioned before, multi-path effects are not considered, though our proposed solution has better multi-path gathering effects. It is assumed that there are no interframe interference and multi-user interference. As described in section 2, the second derivative of the Gaussian pulse with 250ps duration is used to transmit the data.
2591
Figure 5. Performance of three structures for different data rates
Figure 6. Time mismatch analysis (100 Mbps)
Figure 8. Performance in the presence of jitter in transmitter and receiver (100 Mbps)
Although it shows that the TR-UWB seems most tolerant to skew, as mentioned before, the time mismatch in autonomous receiver can be easily eliminated during manufacture (since it is not related to the transmitter). Coherent receiver and TR-UWB system suffer from jitter or phase noise of the oscillator in the transmitter and receiver. But this novel architecture is not affected by the jitter. Fig. 7 and 8 show the performance in the presence of timing jitter. In coherent receiver, jitter in both of transmitter and receiver have been considered. Unlike coherent or TR-UWB receivers, the BER-SNR curve of our new autonomous receiver is independent of jitter. V. CONCLUSION In this paper a novel architecture for non-coherent receiver is proposed for autonomous detection of UWB signals. Comparison between this new architecture and conventional coherent and TR-UWB system is presented. Simulations show that the BER performance of this novel architecture is close to but better than TR-UWB system when timing noise is not considered and the new architecture is more robust against jitter and phase noise. Problem of skew in synchronization could be easily eliminated.
The results show that this architecture can be considered as a promising candidate for low complexity, low cost UWB communications. Future works will be oriented to the analysis the performance in multi-path and multi-user environment and considering the interferences and implement aspects of this autonomous receiver. REFERENCES [1] W. Hirt, “Ultra-wideband radio technology: overview and future research”, Computer Communications, Vol. 26, pp. 46–52, 2003 [2] Moe Z. Win, Robert A. Scholtz, “Impulse Radio: How It Works”, IEEE COMMUNICATIONS LETTERS, VOL. 2, NO. 2, pp. 36-38, 1998 [3] Aiello, G.R.; Rogerson, G.D.”Ultra-wideband wireless systems” Microwave Magazine, IEEE Vol 4, pp.36 – 47, 2003 [4] Ramachandran, I.; Roy, S.; ” Acquisition of direct-sequence ultrawideband signals”; Wireless Communications and Networking Conference, 2005 IEEE Volume 2, 13-17 March 2005 Page(s):752 757 Vol. 2 [5] Zhi Tian; Lottici, V.; “Efficient timing acquisition in dense multipath for UWB communications”; Vehicular Technology Conference, 2003. VTC 2003-Fall. 2003 IEEE 58th Volume 2, 6-9 Oct. 2003 Page(s):1318 - 1322 Vol.2 [6] G. Durisi; S. Benedetto; “Performance of Coherent and Non-coherent Receivers for UWB Communications”; IEEE communication Society 2004, Pages 3429-3433 [7] Idriss, A.; Moorfeld, R.; Zeisberg, S.; Finger, A. ”Performance of coherent and non-coherent receivers of UWB communication”, Wireless and Optical Communications Networks, 2005. Second IFIP International Conference on March , 2005 Page(s):117 - 122 [8] Honglei Zhang; Goeckel, D.L.; ”Generalized transmitted-reference UWB systems”; Ultra Wideband Systems and Technologies, IEEE Conference. 2003 Page(s):147 – 151 [9] Ho, M.; Somayazulu, V.S.; Foerster, J.; Roy, S.; “ A differential detector for an ultra-wideband communications system”; Vehicular Technology Conference, 2002. VTC Spring 2002. IEEE 55th Volume 4, 6-9 May 2002 Page(s):1896 - 1900 vol.4 [10] Paquelet, S.; Aubert, L.-M.; Uguen, B.; “ An impulse radio asynchronous transceiver for high data rates”; Ultra Wideband Systems, 2004. Joint with Conference on Ultrawideband Systems and Technologies. Joint UWBST & IWUWBS. 2004 International Workshop on 18-21 May 2004 Page(s):1 – 5 [11] Witrisal, K.; Pausini, M.; Trindade, A.; ” Multiuser interference and inter-frame interference in UWB transmitted reference systems”; Ultra Wideband Systems, 2004. Joint UWBST & IWUWBS. 2004 International Workshop on 18-21 May 2004 Page(s):96 - 100
Figure 7. Jitter analysis (100 Mbps)
2592
I. INTRODUCTION Recently, UWB systems are paid attention by academic research group and industries in short-distance, low power, high data rate and low cost applications. This attention increases since the American Federal Communication Committee (FCC) released the UWB spectral in FEB 2002. Because of the extremely low power density of UWB, FCC allows it to operate over the top of the bands. It is helpful to solve the frequency allocation problem that often limits high data rate wireless communication systems [1]. The wide bandwidth of UWB offers several advantages over the narrow band systems, such as low-power, impulse-like, very high data rate, diversity against multi-path, base-band transition and large number of users [2], [3]. With these advantages, UWB system can be considered as a powerful candidate for Wireless Personal Area Network (WPAN). It is also considered for low power and low cost RFID (radioidentification) and sensor systems. There are two general strategies for Ultra Wideband receivers. Optimal Rake coherent receivers use complete channel estimation to resolve paths. It is a high performance technique but also relatively expensive. On the other hand, sub-optimal noncoherent receivers basically involve energy detectors and hence relax from channel estimation [2], are attractive for low cost and low data rate applications. The coherent receivers have a good performance however, duo to the extremely number of paths and signal distortion by antennas and propagation environment, synchronization and general channel estimation in UWB
0-7803-9390-2/06/$20.00 ©2006 IEEE
schemes are critical in design [4], especially in dense multi-path and mobile applications [5]. This difficulty arises from the ultra-short pulses with low amplitude used in UWB radio. Therefore, the coherent receivers are very complex and expensive. In order to achieve low cost and low complexity transceivers, non-coherent detection has been proposed. Non-coherent receivers offer a simple and cheap architecture for the receiver in cost of losses in BER performance and lower data rate that have limited their applications [6-7]. Fundamentally non-coherent receivers need a higher SNR (signal-to-noise ratio) and therefore they are employed in short distance and low data rate applications such as RFID and sensor systems. Because of simplicity and low cost, many researches have worked on this scheme and different architectures have been investigated such as Transmitted-Reference UWB or TRUWB [8] and Differential TR-UWB systems [9]. Generally, the non-coherent receivers employ a replica of the received signal delayed by Td as the template and make a correlation between this template and the received signals. Td is a fixed delay depends on the time interval between two pulses in one doublet. If there is a mismatch between the delay value and the time interval, performance of the receiver will be degraded. This timing mismatch can be caused from process variation or from inaccuracy and jitter of the oscillator in transmitter. Another approach that relaxes from the channel estimation and synchronization is the asynchronous transceiver. Data is carrying by OOK modulation. The receiver uses an energy detector to detect the data. [10] This architecture has also poor BER and low data rate and needs high SNR input. By using non-coherent or asynchronous approach, frame synchronization is simplified to symbol synchronization. In this paper, we propose a novel non-coherent receiver that can autonomously detect received signals. The new architecture is a modification of TR-UWB system in which another multiplier is used to multiply the signal instead of using the delay component. We validate the performance of this new receiver by simulations in Matlab and compare it to traditional coherent receiver and TRUWB systems. We found that the new architecture has significant improvements in tolerance of timing noise. We will start from the channel and signal models of the UWB systems used in this paper in Section 2, then describes the autonomous receiver in section 3. Analysis and simulation results are given in section 4. Section 5 is the conclusion.
2589
ISCAS 2006
II. SIGNAL AND CHANNEL MODEL Information in UWB system is typically transmitted using a collection of pulses with short width and low duty cycle. Each user is assigned a different Pseudo-noise (PN) sequence that is used to encode the pulse in BPSK or PPM. In order to clearly compare with other architectures, in this paper we consider the BPSK modulation and single user and non multi-path environment. For BPSK the code modulates the polarity of the pulses. The received signal can be written as follow: ∞ N f −1
r (t ) = ∑ i =0
∑
bi p (t − (i.N f + j )T f ) + n(t )
Figure 2. Autonomous receiver
The output of the integrator can be expressed as:
an =
(1)
j =0
where P(t) is the monocycle template pulse. The second deviation of Gaussian pulse with 250ps width used as the template in this paper is shown in Fig. 1. T f is the duration of each frame, bi = ± 1 is the data, N f is the number of frame that is used to transmit one bit that can be between 1 to several hundred and n(t) is AWGN. III. SYSTEM DESCRIPTION A. Proposed Autonomous Receiver Scheme The structure of our autonomous receiver is illustrated in Fig. 2. As can be seen, this is a modification of the TRUWB system. The template waveform employed in the demodulation process consists of a squared replica of the received signal. Similar to coherent receiver, a band-pass filter h(t) of bandwidth B is employed to reduce the effect of noise in the demodulation process. It should be mentioned that this filer is very critical compared with that in coherent receivers. This is because, coherent operation is able to moderate the noise and maximising the SNR if a template pulse matched to the received signal is available. Therefore the filter can be designed in such a way that dose not effect the received signal. However in the non-coherent receiver, template waveform is very noisy. The bandwidth of the filter must be chosen as a trade-off between the signal energy and noise reduction. In order to synchronize the two signals at the input of the mixer, a fixed delay line is chosen to compensate the mixer delay.
( n +1)T f N f 3 1 nT f N f
∫r
(t )dt
(2)
(3) where r1 (t ) = r (t ) * h(t ) This architecture has a number of attractive properties: • Same as in other non-coherent receivers, the synchronization is only in symbol level, no frame level synchronization is needed. Therefore the complexity of the receiver is reduced, especially suitable for low data rate applications, where often T s >> T f . • Multi-path gathering is achieved. Since all components in the received signal go trough the same system and channel, the produced template pulse provides a perfect (but unfortunately noisy) template to match the received signal without any channel estimation or the need for a rake receiver with many branches. • Since the template pulse is as the same as received signal, changing in channel does not affect the receiver performance. This can be a significant advantage for systems operating in a highly mobile environment. • Because the template pulse is produced from the received signal, this architecture does not suffer from jitter and phase noise in the transmitter oscillator. It is very robust to timing noise. • In this architecture the transmitter does not send any extra pulse for synchronization purpose. It therefore improves the performance in case of multi-user interference and inter-frame interference in TR-UWB system [11]. • The transmitter power consumption will be reduced because it does not send any extra pulses. This can also be a significant advantage for self-powered systems such as RFID and passive sensor systems. B. Other Receiver Architectures for Comparison In order to compare the performance of our proposed receiver and traditional receivers, we choose two other architectures: coherent receiver and TR-UWB receiver. For a fair comparison, same signal and channels models are used.
Figure 1. Monocycle pulse
B.1. Coherent Receiver A general coherent receiver used in this paper is shown in Fig. 3. A template signal, generated according to the information acquired by bit and code synchronization and possibly channel estimation, is multiplied by the received signal. In the end the result passed trough an integrator and decision block.
2590
Figure 4. Transmitted-Reference receiver Figure 3. Coherent receiver
The output of the integrator can be written as: ( n +1) N f .T f
∫ r (t ).x(t − n.N
an =
f
.T f )dt
(4)
n. N f .T f
where
x(t ) =
N f −1
∑ p(t − m.T
f
)
(5)
m=0
p(t) is the template pulse and r(t) is the received signal. B.2. Transmitted-Reference UWB System A transmitted reference receiver is simulated for comparison. Its structure is illustrated in Fig. 4. The signal is multiplied by a delayed version of itself and the result integrated over the symbol period Ts , which can consist of many frames, each of duration T f , in low data rate applications. A threshold block makes a decision on the output of integrator to decode the data for the current symbol. As we described before, the design of the frontend filter is critical in TR-UWB system in order to reduce the effect of noise. Each bit is transmitted by a pair of pulses as:
S1 (t ) = p (t ) + p (t − Td ) S 0 (t ) = p (t ) − p (t − Td )
(6)
where S1 ( t ), S 0 ( t ) are the signals transmitted for an information bit 1 and bit 0 respectively and p(t) is the pulse shape. Td is a fixed value and chosen as
TP < Td < T f where TP is the pulse duration. The output of the integrator can be expressed as:
The FCC regulations are not considered here. In coherent and our autonomous receivers, PBSK modulation and in TR-UWB two-points-signal are utilized. Performances of these three receivers are evaluated in term of bit error rate (BER) versus signal to noise ratio (SNR). A front-end band-pass filter is considered with no distortion on the pulse shape. Both of The receiver and noise bandwidth are assumed to 5 GHz and the noise figure of the receiver is 8dB. Frame period is set to 10ns and the number of pulse per bit is changed to estimate the performance in different data rates. In order to have an accurate estimation of BER, 10000 bits are applied to each receiver. Fig. 5 shows the BER performance versus the SNR for three structures. It is clearly seen that the coherent receiver has highest performance, but as mentioned before, it is expensive and complex. The autonomous receiver and TR-UWB show similar performance. It must be mentioned that in this case, no timing and phase noise have been assumed and more improvements are expected when this effect is considered, as will be shown in next figures. Second, we introduce a time mismatch. In coherent receiver, this is a time skew in synchronization and in TRUWB and our new architectures this is the deference of Td from expected value. In autonomous receiver the delay value equals to the delay of the mixer. The accuracy of this delay is only related to the receiver and it is much easier to control during manufacture. Fig. 6 compares the performance of the receivers in the presence of time mismatch. As can be seen, any mismatch degrades the performance.
( n +1) N f .T f
an =
∫ r (t ).r (t − T
1 n . N f .T f
1
d
).dt
(7)
where r1 (t ) is the filtered signal. IV. RESULTS AND DISCUSSIONS The receiver structures described in previous section are simulated in MATLAB and the results are compared. An AWGN channel is considered as the channel model. As mentioned before, multi-path effects are not considered, though our proposed solution has better multi-path gathering effects. It is assumed that there are no interframe interference and multi-user interference. As described in section 2, the second derivative of the Gaussian pulse with 250ps duration is used to transmit the data.
2591
Figure 5. Performance of three structures for different data rates
Figure 6. Time mismatch analysis (100 Mbps)
Figure 8. Performance in the presence of jitter in transmitter and receiver (100 Mbps)
Although it shows that the TR-UWB seems most tolerant to skew, as mentioned before, the time mismatch in autonomous receiver can be easily eliminated during manufacture (since it is not related to the transmitter). Coherent receiver and TR-UWB system suffer from jitter or phase noise of the oscillator in the transmitter and receiver. But this novel architecture is not affected by the jitter. Fig. 7 and 8 show the performance in the presence of timing jitter. In coherent receiver, jitter in both of transmitter and receiver have been considered. Unlike coherent or TR-UWB receivers, the BER-SNR curve of our new autonomous receiver is independent of jitter. V. CONCLUSION In this paper a novel architecture for non-coherent receiver is proposed for autonomous detection of UWB signals. Comparison between this new architecture and conventional coherent and TR-UWB system is presented. Simulations show that the BER performance of this novel architecture is close to but better than TR-UWB system when timing noise is not considered and the new architecture is more robust against jitter and phase noise. Problem of skew in synchronization could be easily eliminated.
The results show that this architecture can be considered as a promising candidate for low complexity, low cost UWB communications. Future works will be oriented to the analysis the performance in multi-path and multi-user environment and considering the interferences and implement aspects of this autonomous receiver. REFERENCES [1] W. Hirt, “Ultra-wideband radio technology: overview and future research”, Computer Communications, Vol. 26, pp. 46–52, 2003 [2] Moe Z. Win, Robert A. Scholtz, “Impulse Radio: How It Works”, IEEE COMMUNICATIONS LETTERS, VOL. 2, NO. 2, pp. 36-38, 1998 [3] Aiello, G.R.; Rogerson, G.D.”Ultra-wideband wireless systems” Microwave Magazine, IEEE Vol 4, pp.36 – 47, 2003 [4] Ramachandran, I.; Roy, S.; ” Acquisition of direct-sequence ultrawideband signals”; Wireless Communications and Networking Conference, 2005 IEEE Volume 2, 13-17 March 2005 Page(s):752 757 Vol. 2 [5] Zhi Tian; Lottici, V.; “Efficient timing acquisition in dense multipath for UWB communications”; Vehicular Technology Conference, 2003. VTC 2003-Fall. 2003 IEEE 58th Volume 2, 6-9 Oct. 2003 Page(s):1318 - 1322 Vol.2 [6] G. Durisi; S. Benedetto; “Performance of Coherent and Non-coherent Receivers for UWB Communications”; IEEE communication Society 2004, Pages 3429-3433 [7] Idriss, A.; Moorfeld, R.; Zeisberg, S.; Finger, A. ”Performance of coherent and non-coherent receivers of UWB communication”, Wireless and Optical Communications Networks, 2005. Second IFIP International Conference on March , 2005 Page(s):117 - 122 [8] Honglei Zhang; Goeckel, D.L.; ”Generalized transmitted-reference UWB systems”; Ultra Wideband Systems and Technologies, IEEE Conference. 2003 Page(s):147 – 151 [9] Ho, M.; Somayazulu, V.S.; Foerster, J.; Roy, S.; “ A differential detector for an ultra-wideband communications system”; Vehicular Technology Conference, 2002. VTC Spring 2002. IEEE 55th Volume 4, 6-9 May 2002 Page(s):1896 - 1900 vol.4 [10] Paquelet, S.; Aubert, L.-M.; Uguen, B.; “ An impulse radio asynchronous transceiver for high data rates”; Ultra Wideband Systems, 2004. Joint with Conference on Ultrawideband Systems and Technologies. Joint UWBST & IWUWBS. 2004 International Workshop on 18-21 May 2004 Page(s):1 – 5 [11] Witrisal, K.; Pausini, M.; Trindade, A.; ” Multiuser interference and inter-frame interference in UWB transmitted reference systems”; Ultra Wideband Systems, 2004. Joint UWBST & IWUWBS. 2004 International Workshop on 18-21 May 2004 Page(s):96 - 100
Figure 7. Jitter analysis (100 Mbps)
2592
