摘要
提出一种自适应变尺度暂态混沌神经网络,并将其应用于CDMA的多用户检测技术.该算法在优化搜索过程中根据暂态混沌神经网络能量函数的变化调整网络参数,自适应地控制能量函数对神经网络动力学特性产生良好的影响.仿真结果表明,基于改进的自适应混沌神经网络算法的多用户检测器能够有效地逼近CDMA的最优多用户检测器的性能.
An improved adaptive chaotic neural network algorithm was proposed and applied to the CDMA multiuser detection technique. During the optimal searching process, this algorithm adjusts network parameters according to the change of the energy function of transient chaotic neural network. So it can control the energy function to produce favorable influence on the dynamics of neural network adaptively. The simulation results show that the multiuser detector based on improved adaptive chaotic neural network can approach the performance of the optimum CDMA multiuser detector effectively.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2004年第5期697-700,共4页
Journal of Shanghai Jiaotong University
基金
国家自然科学基金资助项目(60272082
60372076)
关键词
混沌
神经网络
码分多址
多用户检测
组合优化
Adaptive algorithms
Chaos theory
Code division multiple access
Combinatorial mathematics
Optimization
