The Research Of Adaptive Multi-user Detection Algorithms

CDMA (Code Division Multiple Access) is a priming technology for future mobile communication system because of its several features with which other multiple access technology can not compare: higher efficiency of frequency allocation, robust anti-multi-path effect, macro diversity, easy access, and so on. But CDMA communication system is a interference-limited system. That is to say that the interference directly influences the capacity of this system and the QOS of the user in the system. In order to rid the multi-access interference, multi-user detection plays a important role. It executes the jointly detection fully using the spreading sequence,time delay,phase information and amplitude. It considers the difference between the multi-user interference and the white noise. The uplink frequency resource is used efficiently. The performance has enhanced as a whole.This paper introduces correspondence multi-user examination principle and existing multi-user detection algorithm in CDMA wide frequency. It embarks from the multi-user signal statistical independence with the distributed statistical characteristic and proposes two kind of blind multi-user examinations strategies: Based on improved information maximization multi-user detection algorithm and based on fast fixed-point independent component analysis multi-user detection method. These two methods do not need the training sequence and fully uses the multi-user signal characteristic code. They overcome the blind separation algorithm on the indefinite question. Simultaneously they have also used the multi-user signal statistical independence and the inherent characteristic of blind separation algorithm. Under unknown situation to the signal source, they separate and detect the object user. Under big signal and noise environment, two algorithms could detect target user with less computation than information maximization algorithm and the traditional detection algorithms. They also have good performance in BER (bit error rate), convergence and capability. Some simulations confirm the conclusions. They save the frequency spectrum resources and simultaneously obtained the better separation and detection effect.This paper also embarks from the statistical information theory angle, in independent parameter: KL (Kullback-Leibler), two kind of blind separations algorithm are unified under a theory frame, and simultaneously used in two algorithms in the multi-user examination to do the comparison. The simulation result has proven that the theory inferential is reasoning and accuracy.