自适应RBF神经网络在CDMA移动通信上行链路功率控制中的应用研究  

The application investigation of adaptive RBF neural network for uplink power control in CDMA mobile communications

作　　者：倪梁方[1] 郑宝玉[1] 

机构地区：[1]南京邮电学院信息工程系,江苏南京210003

出　　处：《通信学报》2003年第12期42-51,共10页Journal on Communications

基　　金：国家"863"计划基金(2002AA121068);广东省自然科学基金(20001641)

摘　　要：提出了一种自适应RBF神经网络功率控制方案。详细研究了该网络在DS-CDMA通信中,进行上行链路闭环功率控制(基于信扰比(SIR))的应用理论,给出了该网络参数的计算方法。最后用计算机仿真法模拟出该控制器的运行性能。结果表明基于SIR的自适应RBF神经网络功率控制器能自适应地调整移动台的发射功率,使基站接收信号的信扰比始终非常接近于一个常数,且有比定步长功率控制更小的SIR跟踪误差,从而可以降低接收信号的中断概率、提高信道容量。An adaptive radial basis function neural network (ARBFNN) power control scheme is proposed. Study is conducted on the application of ARBFNN to the uplink closed-loop power control based on signal-to-interference ratio (SIR) in DS-CDMA communications, and the approach to calculate the parameters of the network is put forward. Finally, the practical performance of this controller is simulated via computer. The simulation results have shown that the SIR-based ARBFNN power controller can adaptively raise or lower each mobile transmitting power level to keep the SIR of the signal received by the base station almost constant, and achieve a smaller SIR root-mean-squared (rms) tracking error than the fixed step power controller. This better controller performance would lead to smaller outage probability and substantial capacity improvements.

关 键 词：直接序列扩频 SIR 功率控制 自适应RBF神经网络 上行链路 异步码分多址 CDMA 

分 类 号：TN929.533[电子电信—通信与信息系统]