论文部分内容阅读
传统的单用户检测算法当系统中用户数及由其他用户信号形成的多址干扰较大时，其检测性能将大为恶化。而最佳多用户检测算法虽能有效的消除多址干扰，但它极高的计算杂度大大限制了它在实践中的应用。因此，采用多层感知器神经网络来实现同步ＣＤＭＡ通信中用户信号的检测接收，并用文中提出的改进的反向传播（ＢＰ）算法来训练网络，其性能大大优于传统单用户检测算法，且逼近最佳多用户检测性能。
The traditional single-user detection algorithm when the number of users in the system and multiple users formed by the signal interference larger, the detection performance will be greatly deteriorated. The best multi-user detection algorithm can effectively eliminate multiple access interference, but its extremely high computational complexity greatly limits its practical application. Therefore, using multi-layer perceptron neural network to detect and receive user signals in synchronous CDMA communication and training the network with the improved backpropagation (BP) algorithm proposed in this paper is much superior to the traditional single-user detection algorithm Approaching the best multi-user detection performance.