Research On Multiuser Detection Algorithm In Wireless Ultra-wideband Communications

Both large capacity and precise ranging are persuing goals in IR-UWB communication systems. Extremely high bandwidth, and very low-bound transmit power is the fundamental characteristics for reaching this expective performance. IR-UWB carries information on base-band impulse while traditional wireless system on continuous carry modulation. The traditional signal detection and processing method are confronted with a significant handle in practice due to several drawbacks of the impulse transmission, such as rapid path fading, density multipath, long delay spread, and multiuser interference. In view of problems mentioned above, IR-UWB gains importance in its own right. This thesis concentrates on the signal detection and the multiuser signal processing. It tries the best to take advantages of the received impulse signal to obtain simple and effective policies of multiuser detection. The main contribution is concluded as follows:1. The demodulation performances are analyzed for IR-UWB multi-access system. The thesis provides a simple approach to compute bit-error-rate (BER) in multiuser environment. Furthermore, considering the multipath fading channel in UWB, the expression on multi-access BER performance of coherent detection and Rake division detection are concluded and verified via simulation. Through analysis and experiments in multiuser UWB channel, we find the internal relation between the detection performance and the channel factors, including the fading of the multipath components and the length of the delay spread;2. We propose a least mean fourth (LMF) based multiuser detection criterion, which has the outstanding decorrelation capability at high SNR environment. By time-domain window filtering, the high E_b/N₀ can be obtained for the concentrative energy in impulse signal. The thesis analyzed the convergence of different high-order statistics (HOS) based blind multiuser detection criterions. The Hessian matrix is derived in noise environment, and furthermore, is used to analyze and prove the decorrelation and convergence of LMF criterion. In multuser CDMA system, the least square method is adopted to construct a RLS-LMF algorithm with faster convergence in multipath channel;3. In view of the non-overlap characteristics of impulse spreading signa, we propose a framework, called "multipath-free", for simplifying multiuser detection. Under this framework, the code matrix in the common multipath constrain is not necessary for multiuser detection. That makes many efficient detection approach in AWGN channel, for example, Kalman algorithm, could be directly applied in the impulse radio system. No doubt, it simplifies the algorithms and boosts the proformance. Moreover, least-square derivation is implemented in adaptive LMF detection and the MF-RLS-LMF algorithm is obtained. It has lower complexity than multipath RLS-LMF proposed above;4. A modified TR-UWB signal and detection are proposed for suppressing the multiuser interference. The approach not only uses the information in the reference signal template, but also the reference symbol information. It modulates the reference impulse by user pseudo-random code, and uses analog LMS feedback to obtain the optimal detection template. This approach, in fact, takes advantage of the statistics independence of multipath signal to construct an effective signature to lessen the correlative interference.5. For low data rate and large user capacity IR-UWB communicaiotn, we propose a multiuser detection algorithm with high performance and low computation complexity. The algorithm adopts an interval-guarded impulse modulation with direct sequence spread. Due to the non-overlap characteristics of impulse spreading signal, the multipath-free framework is applied in the proposed dual-interpolating multiuser detector. The effective equalization is applied in the algorithm at the same time. The analysis and simulation verify the efficiency of the proposed algorithm even in the NLOS channel.