An Integrated Deep Learning-Based Detector for OFDM with Index Modulation Under Imperfect Channel Conditions

Index modulation (IM) is a notable option for fifth-generation wireless communication due to its efficient spectrum usage and lower complexity compared to traditional modulation schemes. However, in an environment with uncertain channel conditions, IM encounters challenges in reliably detecting indices and effectively optimizing dynamic index assignments. Deep learning (DL) may be an important candidate for overcoming this issue, as it can enhance detection accuracy and robustness by learning complex patterns amidst varying channel quality. To leverage DL, in this paper, we propose an integrated deep neural network (InDNN) based detection technique for orthogonal frequency division multiplexing with IM systems under uncertain channel conditions. Our integrated detector is designed with a one-dimensional convolutional neural network (1D-CNN) and a bi-directional long short-term memory (Bi-LSTM) model. The channel matrix and received signal undergo preprocessing using domain knowledge before entering the network. To infer the signal at the receiver terminal, the 1D-CNN extracts features from the input signal and feeds them into the time-series BiLSTM network. The simulation results confirm that our proposed integrated detector outperforms both traditional detectors and other DL-based detectors in performance comparisons.