Reduced-Rank Detection for Hybrid Direct-Sequence Time- Hopping UWB Systems in Nakagami-m Fading Channels

This paper investigates and compares the performance of hybrid direct-sequence time-hopping (DS-TH) ultrawide bandwidth (UWB) systems using full-rank and various reduced-rank detections when minimum mean-square error (MMSE) principles are applied, while communicating over Nakagami-m fading channels. In addition to the full- rank MMSE detection, three classes of reduced-rank MMSE techniques are investigated, which are derived based on the principles of principal components (PC), cross-spectral metric (CSM) and Taylor polynomial approximation (TPA), respectively. Our study and simulation results show that the reduced-rank techniques constitute a family of promising detection techniques for UWB systems, which are capable of achieving similar bit error rate (BER) performance as the full-rank detection, while with a detection complexity that is significantly lower than that of the full-rank detection.