POR-based channel estimation for UWB communications

In an ultrawideband (UWB) system, typically, a Rake receiver is employed to detect information symbols. To fully capture signal energy spread over multiple paths, the desired user's channel parameters are necessary. In this paper, we first convert a time-hopping (TH) UWB signal model using pulse position modulation (PPM) to a linear form similar to that of a code division multiple access (CDMA) system. Then, we apply a power-of-R (POR) technique to blindly estimate multipath parameters of the desired user in a multiple-access (MA) UWB system, where "R" represents the data covariance matrix. The POR technique is related to the subspace technique in that it directly estimates the unknown noise-subspace component in the subspace method but avoids rank estimation where errors may be incurred by perturbations in practical conditions. Performance of the proposed POR channel estimator is analyzed and shown to converge to that of the ideal subspace approach. Detection performance of the Rake receiver constructed from the estimated channel is also studied in terms of output signal-to-interference-plus-noise ratio (SINR) and bit error rate (BER). The proposed method demonstrates superiority to existing maximum-likelihood (ML)-based approaches in terms of either channel estimation or symbol detection. It also outperforms the subspace method for heavily loaded systems.