Deep Learning-based Signal Detection Technique for FTN Signaling-based Emergency Alert Communication System

In this paper, for spectrally efficient transmission of emergency alert signal, a faster-than-Nyquist signal is considered. FTN signaling is a useful communication technique for high spectral efficiency. However, since the use of a faster symbol rate than Nyquist rate destroys orthogonality between symbols, ISI is generated inevitably. To reduce the ISI effect, interference cancellation and signal detection process is required for the FTN receiver. Generally, interference cancellation techniques based on the trellis algorithm are adopted for ISI reduction. The complexity of the trellis algorithms is highly increased according to the increase of the number of states. And the increase of interference symbols augments the number of states exponentially. This paper investigates the interference cancellation technique based on deep learning technology for FTN communication systems. To reduce the continuous interference between adjacent signals, LSTM algorithm-based RNN is applied. The simulation results show that the proposed deep learning-based signal detector can provide similar performance to the trellis-based BCJR algorithm.