基于深度学习的OFDM-IM信号检测方法  

OFDM-IM Signal Detection Method Based on Deep Learning

作　　者：张志晨 王昱凯 王荣 李军[1] 郑文静 ZHANG Zhichen;WANG Yukai;WANG Rong;LI Jun;ZHENG Wenjing(School of Information and Automation,Qilu University of Technology(Shandong Academy of Science),Jinan 250353,China)

机构地区：[1]齐鲁工业大学(山东省科学院)信息与自动化学院,山东济南250353

出　　处：《无线电工程》2023年第7期1572-1577,共6页Radio Engineering

基　　金：国家自然科学基金(12005108);山东省自然科学基金(ZR2020QF016)。

摘　　要：正交频分复用(Orthogonal Frequency Division Multiplexing,OFDM)技术在无线通信领域中拥有着重要地位,但OFDM系统中存在子载波间干扰和较高的峰均比的缺点,使得OFDM系统在信号检测方面的表现不太理想。针对OFDM系统中信号检测性能较差的问题,提出一种基于自归一化网络的索引调制(Index Modulation for Self Normalizing Network,IM-SNN)算法,并采用4QAM、8QAM、16QAM的调制方式验证系统的信号检测性能。结果表明,所提出的算法提高了接收端解调信号的性能,有效增强了信号检测的能力,并表现出优于传统技术中最大似然检测(Maximum Likelihood Detection,MLD)算法及现有技术中基于深度神经网络的索引调制(Index Modulation in Deep Neural Network,IM-DNN)算法的系统误码率及网络损失。在3种调制方式下,性能改善0.6~8 dB。Orthogonal Frequency Division Multiplexing(OFDM)technology plays an important role in the field of wireless communication.However,the OFDM system has the disadvantages of inter-carrier interference and high peak-to-average ratio.The performance of OFDM system in signal detection is not ideal.To deal with the problem of poor signal detection performance in OFDM system,an Index Modulation for Self Normalizing Network(IM-SNN)algorithm is proposed.The signal detection performance of the system is verified by 4QAM,8QAM and 16QAM modulation modes.The results show that the proposed algorithm improves the performance of the receiver demodulation signal,effectively enhances the capability of signal detection,and performs better in terms of bit error rate and network loss than the traditional technique of Maximum Likelihood Detection(MLD)algorithm and the current technique of Index Modulation in Deep Neural Network(IM-DNN)algorithm.The performance improvement is 0.6~8 dB in the three modulation modes.

关 键 词：正交频分复用 索引调制 信号检测 自归一化网络 深度学习 

分 类 号：TN92[电子电信—通信与信息系统]