Near-optimal and suboptimal receivers for multiuser UWB impulse radio systems in multipath

In this paper, near-optimal and suboptimal receivers are proposed for multiuser ultra-wideband (UWB) impulse radio (IR) systems in dense multipath channels, where non-Gaussian multiuser interference (MUI) seen at each finger of a Rake receiver is taken into account. The receivers exploit twodimensional diversity that is offered by UWB signaling, namely multipath diversity and repetition diversity (i.e., repetition coding). Non-Gaussian modeling of the MUI results in employing a nonlinear process at each finger, jointly considering the desired user's path gain being estimated and the order associated with MUI statistics. For this, the MUI statistics are characterized in terms of their second and fourth order moments, especially considering the channel sparseness and cluster overlapping observed in realistic UWB channels, which leads to modeling the MUI by a generalized Gaussian distribution. It is shown that the MUI statistics exhibit non-Gaussian nature even for relatively large number of users, and the diversity gain against MUI can be significant when relatively small number of users coexist.