论文部分内容阅读
本文叙述了通过构造一种复合神经网络实现直接序列码分多址通信系统中实时最佳多用户检测器的方法。这种最佳多用户检测器的实时性来自复合神经网络运算结构的并行性。复合神经网络由一个前向单层网络和一个竞争网络构成，其中前向单层网络进行似然函数计算，而竞争网络实现选择最大项操作。
This article describes a real-time optimal multi-user detector for direct sequence code division multiple access communication system by constructing a composite neural network. The real-time nature of this best multi-user detector comes from the parallelism of the composite neural network computing architecture. The composite neural network consists of a forward single-layer network and a competitive network, in which the forward single-layer network performs the likelihood function calculation and the competitive network implements the selection of the maximum-term operation.