通信系统中的信道均衡问题研究

Research on channel equalization in communication systems

针对通信系统中的信道均衡问题进行了研究,设计了一种基于MLP多层感知器的信道均衡系统。首先,对传统线性均衡器存在的问题进行了描述;其次,针对传统线性均衡器存在的问题构建了基于MLP多层感知器的信道均衡系统模型,即以MLP均衡器为核心,增加训练模块与判决模块;然后引入一种快速学习算法对模型进行训练,即通过学习率和动量项优化后的BP算法,减少了模型训练的时间;最后,对快速学习算法进行测试,通过星座图。均方误差折线图以及误码率折线图三种方式对传统线性均衡器与基于MLP多层感知器的信道均衡系统进行测试与对比。结果表明:在复杂环境下,传统线性均衡器对星座图的校正会出现模糊和偏移的现象,而基于MLP多层感知器的信道均衡器对星座图的校正更为精准有效;均方误差折线图显示,在整体上,MLP结构的均衡器比传统LSM均衡器降低了5 dB;误码率折线图显示,在信噪比达到16 dB之后,MLP结构的均衡器误码率更低。综上可知,本研究提出的MLP结构的均衡器性能更佳且稳定性更好。

In this paper,the channel equalization problem in communication system is studied,and a channel equalization system based on MLP multilayer perceptron is designed.Firstly,the problems of traditional linear equalizers are described.Secondly,according to the problems existing in traditional linear equalizers,a channel equalization system model based on MLP multilayer perceptron is constructed,that is,with MLP equalizer as the core,the training module and decision module are added.Then,a fast learning algorithm is introduced to train the model,that is,the BP algorithm optimized by the learning rate and momentum term reduces the time of model training.Finally,the fast learning algorithm is tested,through the constellation diagram.The mean squared error line chart and bit error rate line chart test and compare the traditional linear equalizer and the channel equalization system based on MLP multilayer perceptron.The results show that in a complex environment,the correction of the constellation graph by the traditional linear equalizer will be blurred and offset,while the channel equalizer based on MLP multilayer perceptron will correct the constellation map more accurately and effectively.The mean squared error line chart shows that overall,the MLP structure equalizer is 5dB lower than the traditional LSM equalizer;The bit error rate line plot shows that the equalizer with MLP structure has a lower bit error rate after the signal-to-noise ratio reaches 16dB.In summary,it can be seen that the equalizer of MLP structure proposed in this study has better performance and better stability.