Practical application of compressive sensing to ultra-wideband channels

Compressive sensing (CS) is proposed by many researchers as a solution to the formidable sampling requirements of promising ultra-wideband (UWB) technology. Previous research was evaluated by simulating the IEEE UWB channel model. This study examines the behaviour of compressive sensing for practical UWB signals and proposes approaches to enhance signal reconstruction performance. Four practical dictionaries are proposed to increase the sparsity of the realistic UWB signals. Those dictionaries account for the practical effects of channel-like pulse dispersion and the unavoidable effects of antenna directivity. Both measured data and a more practical directional model are used for evaluation. It is shown that CS based on the new dictionaries is able to reconstruct the practical UWB signals more efficiently with reasonable complexity. Performance improvement of the proposed practical dictionaries in channel estimation for two different receivers is quantified.