基于人工神经网络的移动可见光通信接收机研究  

Study on moving visible light communication receiver based on artificial neural networks

作　　者：杨恺[1] YANG Kai(School of Electronic and Electrical Engineering,Dongguan Polytechnic,Dongguan 523808,China)

机构地区：[1]东莞职业技术学院电子与电气工程学院,广东东莞523808

出　　处：《光学技术》2020年第5期591-596,共6页Optical Technique

基　　金：2020年东莞市社会科技发展一般项目“基于可见光通信的智能交通管控系统的研究与设计”(2020507156678)。

摘　　要：当前室内可见光通信系统大多考虑接收机静止的情况,无法适用于日益增多的移动设备。为了解决可见光通信系统中移动接收机的解调问题,提出了基于人工神经网络的可见光通信移动接收机方案。以广泛应用的二进制振幅键控调制技术为基础,推导出可见光通信移动场景的检测方法和解调阈值;通过动态时间规整技术提取光强度序列的距离特征,利用遗传算法对特征集进行优化,选择高显著性的少量特征子集;将特征子集送入人工神经网络进行训练,对二进制振幅键控解调阈值进行预测。实验结果显示,方案有效降低了移动场景下可见光通信系统的误码率。Most indoors visible light communication systems only consider the condition of static receivers,which are not suitable for the growing mobile equipments.In order to resolve the demodulate problem of mobile receivers in visible light communication system.A mobile receiver schema is proposed for visible light communication based on artificial neural networks.Firstly,based on the widely applied binary amplitude keying modulation technology.The detection method and demodulate thresholds are derived for visible light communication mobile situation;then,the schema abstracts the distance features of light intensity sequence through dynamic time warping technology,and it takes advantage of genetic algorithm to optimize the features,high significant feature subsets are selected;finally,the feature subset is adopted to train artificial neural networks,the networks are used to predict the thresholds of binary amplitude keying demodulation.Experimental results show that the proposed schema reduces bit error rate of visible light communication system in the mobile situation.

关 键 词：可见光通信 人工神经网络 调制解调器 遗传算法 特征选择 

分 类 号：TP394.1[自动化与计算机技术—计算机应用技术] TH691.9[自动化与计算机技术—计算机科学与技术]