Chinese Journal on Internet of Things ›› 2022, Vol. 6 ›› Issue (1): 65-72.doi: 10.11959/j.issn.2096-3750.2022.00256

• Theory and Technology • Previous Articles     Next Articles

Res-DNN based signal detection algorithm for end-to-end MIMO systems

Guoquan LI1,2, Yonghai XU1,2, Jinzhao LIN1,2, Zhengwen HUANG3   

  1. 1 School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
    2 Chongqing Key Lab of Photoelectronic Information Sensing and Transmitting Technology, Chongqing 400065, China
    3 Department of Electronic and Computer Engineering, Brunel University London, London UB8 3PH, UK
  • Revised:2022-01-03 Online:2022-03-30 Published:2022-03-01
  • Supported by:
    The National Key Research and Development Program of China(2019YFC1511300);The Natural Science Foundation of Chongqing(cstc2019jcyj-msxmX0666);The Natural Science Foundation of Chongqing(cstc2019jcyj-xfkxX0002)

Abstract:

Deep learning can improve the effect of signal detection by extracting the inherent characteristics of wireless communication data.To solve the tradeoff between the performance and complexity of MIMO system signal detection, an end-to-end MIMO system signal detection scheme based on deep learning was proposed.The encoder and the decoder based on residual deep neural network replace the transmitter and the receiver of the wireless communication system respectively, and they were trained in an end-to-end manner as a whole.Firstly, the features of the input data were extracted by encoder, then the communication model was established and was sent to the zero forcing detector for preliminary detection.Finally, the detection signal was reconstructed through the decoder.Simulation results show that the proposed detection scheme is superior to the same type of algorithm, and the detection performance is significantly better than that of the MMSE detection algorithm at the expense of a certain time complexity.

Key words: deep learning, MIMO system, signal detection, Res-DNN, end-to-end

CLC Number: 

No Suggested Reading articles found!