Spatial modulation signal detection algorithm of particle swarm optimization radial basis network
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
维普
万方数据
针对径向基函数神经网络(Radial Basis Function Neural Network,RBFNN)参数确定不当时易陷入局部最优的问题,提出一种粒子群(Particle Swarm Optimization,PSO)优化RBFNN的空间调制正交频分复用(Optical Spatial Modulation Orthogonal Frequency Division Multiplexing,O-SM-OFDM)信号检测算法.在离线训练阶段,接收端收集部分信号作为训练集并建立RBFNN模型,利用PSO算法搜索RBFNN的最优宽度值,再将得到的模型用于系统进行在线检测.实验结果表明,PSO-RBFNN算法的误码率性能基本近似于最大似然(Maximum Likelihood,ML)检测算法,优于其他对比检测算法,且计算复杂度在光源数目为32、64及128时相较于ML检测算法分别降低了约39.59％、70.24％及 85.24％.
For radial basis function neural network(RBFNN),it is easy to fall into local optimal problem when the parameter of RBFNN is not properly determined.A signal detection algorithm for optical spatial modulation orthogonal frequency division multiplexing(O-SM-OFDM)system is proposed based on the particle swarm optimization(PSO)of RBFNN.In the off-line training stage,the receiver collects part of the signal as a training set and builds an RBFNN model.The PSO algo-rithm is adopted to search the optimal width of the RBFNN,and the obtained model is used for the system online detection.Experiment results show that the bit error rate of the PSO-RBFNN algo-rithm is basically similar to the maximum likelihood(ML)detection algorithm,and is better than other contrast detection algorithms.When the number of light sources is 32,64 and 128,the com-putational complexity is reduced by 39.59％,70.24％and 85.24％respectively compared with the ML algorithm.
visible light communicationspatial modulationorthogonal frequency division multiple-xingradial basis function neural networkparticle swarm optimization
维普
万方数据
