The Applicative Investigation of Adaptive BP Networks for Multi-user Detection in Asynchronous DS-CDMA Mobile Communications

摘要

Three-layer Adaptive Back-Propagation Neural Networks (TABPNN) are employed for the demodulation of spread spectrum signals in a multiple-access environment. A configuration employing three-layer adaptive Back-propagation neural networks is put forward for the demodulation of spread-spectrum signals in asynchronous Gaussian channels . The theoretical arguments and practical performance based on the neural networks are analyzed . The results show that whether the resistance to the multiple access interference or the robust to near-far effects, the proposed detector significantly outperforms not only the conventional detector but also the BP neural networks detector and is comparable to the optimum detector.