Clustered Channel Estimation for UWB Multiple Antenna Systems

Due to their extremely large transmission bandwidth, ultrawideband (UWB) communications systems have the potential for significant multipath capture. However, such capture can be dependent on the accuracy of channel estimation, which is the problem considered herein. Strategies that exploit properties of the UWB channel can offer performance improvement over schemes which are less parametric. To this end, recent UWB propagation studies and associated channel modelling suggest that clustering occurs in both space and time for indoor wireless office/laboratory environments. Motivated by this study, a two-stage channel estimation scheme is proposed. First, coarse channel estimation is conducted which estimates the location and dispersion of a cluster around this location. Then, channel estimation efforts are localized around the nominal location to reduce the dimensionality of the estimation of the parameters associated with the different multipath components. For cluster estimation, methods based on maximum-likelihood and correlation matrix fitting are considered. Different strategies are presented for multipath estimation, based on the expectation maximization algorithm and correlator output maximization. The bit error rate of a Rake type receiver is simulated with the estimated channel parameters for a multipath scenario based on clusters. Simulation results indicate that the clustered approach offers solid performance gains over conventional methods