基于RBF神经网络多用户检测方法的设计

基于免疫算法和免疫进化网络,提出了一种训练RBF网络的混合算法.该算法采用了一种可以实现数据聚类的免疫进化网络,根据输入数据集合自适应地确定RBF网络隐层中心的数量和初始位置;采用免疫算法训练RBF网络,使优化过程趋于全局最优.通过计算机仿真证明,将该方法应用于多用户检测中获得了比传统检测器和其他方法训练的RBF网络多用户检测器更好的误码率检测性能.

Design of multiuser detection method based on RBF neural network

Based on immune algorithm and an evolutionary immune network(aiNet),a hybrid algorithm in designing the RBF network is proposed .An evolutionary immune network for data clustering is used to adaptively determine amount and initial positions of RBF centers according to input data set; then RBF network is trained with immune algorithm that makes the training process tend to global optimum. Simulation results show that the RBF network trained with the algorithm achieves better BER performance than traditional detectors and RBF detectors trained in other methods.