Sudhanshu Gaur

UWB Receiver Design

Ultra-wideband (UWB) systems have received considerable attention in the recent past. UWB communication systems can be deployed in dense multipath environment to provide short-to-medium range communications. As the UWB systems operate at very low power and have bandwidths in excess of few GHz, multipath components of UWB signals tend to have very low SNR per diversity branch. Given the fine time resolution of UWB signals, a rake receiver can be used to exploit the multipath diversity. However, for large number of multipath components rake receiver is not practical because of increased receiver structure complexity. Moreover, large spreading bandwidth of UWB signals will preclude the implementation of rake receivers with huge number of antennas.

In the recent past, several receiver structures have been considered for UWB communication systems. As depicted in Fig. 1(a), a GSC(N, L) receiver adaptively combines a subset of N signals with the best instantaneous SNR out of L available paths. The performance of GSC receiver improves as N is increased and is equivalent to that of maximal-ratio combining (MRC) for the case N = L. However, there is a trade-off involved as this improvement in GSC performance comes at the cost of increased selection circuit complexity.

Fig. 1(a) GSC(N, L) receiver that adaptively combines N paths with strongest instantaneous SNRs.

Fig. 1(b) MRC(N, L) rake receiver that combines N paths with strongest mean SNRs.

Theoretical throughput performance of binary PAM in conjunction with GSC(N, 20) and MRC(N, 20) receiver structures over i.i.d Rayleigh fading environment is illustrated in Fig. 2. As expected, GSC receiver outperforms MRC for fixed number of combined paths N (N ≠ L). For instance, GSC(1, 20) and MRC(5, 20) have identical throughput performances. From Fig. 2, it is also apparent that as the number of combined paths-N are increased, throughput performance of N-tap MRC becomes virtually identical with GSC(N, L). Based on these observations, we can conclude that deployment of GSC(N, L) receiver is desirable for the cases when N << L. Whereas, for N ≥ L/2 an N-tap MRC rake receiver is preferable because of its low complexity and virtually identical throughput performance when compared to GSC(N, L).

Fig. 2 Comparison between the ABER performance of 2-ary PAM that employs either GSC(N, 20) receiver or a MRC receiver that combines first N paths in a i.i.d Rayleigh channel.

Fig. 3 draws comparison between the throughput performance of GSC(N, L) and MRC(N, L) receiver structures in a Rayleigh environment with exponentially decaying multipath intensity profile (δ = 0.2). The average SNR of the n-th diversity path is  where  denotes the average SNR/bit and the parameter is chosen such that the constraint  is satisfied, solving for C yields . It is apparent from Fig. 3 that even for smaller ratio of N/L, MRC(N, L) gives comparable performance with respect to the GSC(N, L) receiver. As N increases, the gap between the MRC and GSC performance curve narrows down. In general, it may be concluded that for greater values of δ, throughput performance of MRC(N, L) is virtually identical to the GSC(N, L) receiver.

Fig. 3 Comparison between the ABER performance of 2-ary PAM that employs either GSC(N, 12) receiver or a MRC receiver that combines first N paths in a Rayleigh channel with nonidentical fading statistics and δ = 0.2.