Interference Mitigation by Statistical Interference Modeling in an Impulse Radio UWB Receiver

Some impulse radio UWB (IR-UWB) networks may allow concurrent transmissions without power control (for example MAC protocols that do not use power control, or co-exisiting, non-coordinated piconets). In such cases, it has been proposed to mitigate multi-user interference (MUI) at the physical layer, but existing proposals for interference mitigation do not account for the multipath nature of UWB channels. We address this problem and propose a receiver that employs a combination of statistical interference modeling and thresholding to mitigate MUI. We find that in a multipath environment the proposed receiver significantly outperforms existing receiver designs that either completely neglect the effect of MUI or only use a simple threshold to reject samples from interfering users. Further, in contrast to successive interference cancellation schemes, our receiver does not require active decoding of each interferer. Thus there is no need to synchronize the receiver with all the interfering users, which would be impractical in an IR-UWB system that is likely to be run in ad hoc mode. To model MUI we consider a hidden Markov model (HMM) and a Gaussian mixture model (GMM). We find that the HMM models interference better than the GMM. However, the resulting performance difference is not huge and comes at the cost of increased receiver complexity.