Complex Deep Neural Network Based Intelligent Signal Detection Methods for OFDM-IM Systems

Abstract

Advanced signal detectors pose a lot of technical challenges for designing signal detection methods in orthogonal frequency division multiplexing (OFDM) with index modulation (IM). Traditional signal detection methods such as maximum likelihood have an excessive complexity, and existing deep learning (DL) based detection methods can reduce the complexity significantly. To further improve the detection performance, in this paper, we propose a complex deep neural network (C-DNN) and a complex convolution neural network (C-CNN) based intelligent signal detection method for OFDM-IM. Specifically, the proposed intelligent signal detection method is designed by C-DNN and C-CNN. The proposed signal detection methods for OFDM-IM use pilots to achieve semi-blind channel estimation, and to reconstruct the transmitted symbols based on channel state information (CSI). Simulation results are given to confirm the performance of the proposed signal detection method in terms of bit error rate and convergence speed.