| 摘要:MC-CDMＡ系统是干扰受限系统，存在多址干扰。本文提出一种基于神经网络和粒子群算法的MC-CDMA多用户检测方法。在粒子位置更新中，随机选择部分粒子作为Hopfield神经网络的神经元组成的个体，进行神经网络的更新；其他粒子仍采用原粒子群算法的位置更新策略，从而能加快粒子群算法的收敛速度以及降低计算复杂度。仿真表明在算法参数相同时，该多用户检测方法在误码率、收敛速度、系统容量、抗远近能力等方面都优于基于粒子群算法的多用户检测和基于神经网络的多用户检测，更加逼近于最佳多用户检测的性能。 |
| 关键词: MC-CDMA 多址干扰 多用户检测 粒子群算法 神经网络 |
|
| A Multiuser Detection for MC-CDMA System Based on Particle Swarm Optimization Algorithm with Hopfield Neural Network |
|
|
| Abstract:There is multiple access interference (MAI) in MC-CDMA which is a interference limited system. This paper presents a multiuser detection (MUD) for MC-CDMA system based on particle swarm optimization (PSO) algorithm with Hopfield neural network (HNN). In the updating of particles’ position, choosing some random particles as individuals composed of neurons in HNN to update the network, and employing PSO updating strategy to the others, which can provide faster rate of convergence and reduce the computational complexity of PSO algorithm. Simulation results show that the performance of the proposed MUD such as bit error rate, rate of convergence, system capacity, and near-far resistance is better than that of MUD based on PSO and MUD based on HNN, and nearly reaching the performance of optimal multiuser detection (OMD), when same parameters are used in these algorithms. |
| Key words: MC-CDMA multiple access interference multiuser detection particle swarm optimization Hopfield neural network |