Intelligent Multiuser Detection Algorithms In DS/CDMA Communication System

As a key technique in CDMA communication system, the multiuser detection can jointly detect the signals transmitted by all the users, thereby alleviate the "near-far" effect, cancel the MAI and increase the system capacity. In this paper by applying intelligent signal processing algorithms to multiuser detection, several DS/CDMA signal receiving algorithms are proposed.The main works of this paper can be summarized as follows:A multiuser detection method based on immune algorithm is proposed. This method decreases greatly computation complexity of optimal multiuser detector and achieves suboptimal performance.A multiuser detection method based on adaptive wavelet network is proposed. This method can effectively eliminate MAI and white noise.A novel space-time matched filter combined with adaptive wavelet network in asynchronous multipath CDMA channels is proposed. It is shown this filter is far superior to the conventional space-time matched filter in eliminating near-far effect and white noise with theoretical analysis and computer simulations.The latest progress in independent component analysis (1CA) is reviewed systematically, then two novel classes of multiuser detection methods based on ICA algorithms and feedforward neural networks are proposed.This paper is supported by the National Natural Science Foundation of China (Grant No. 60073053) and the Natoinal "863" Project.