Peak-to-average power ratio reduction in F-OFDM system using hybrid deep learning and optimised grey coded partial transmit sequence

This work presented an effective peak-to-average power ratio (PAPR) reduction with a hybrid grey code phase factor based partial transmit sequence (PTS) and tone reservation based deep convolutional neural network (CNN) technique (Hybrid Grey-PF-TRDCNN). In the first stage, the tone reservation network TRCNN reserves some of the tones to create the peak-cancelling signal. In the second stage, the resultant PAPR reduced signal of stage 1 is further reduced with the grey code phase factor using PTS. Here, the optimal phase sequence is selected by an efficient beetle swarm optimisation (BSO) technique to reduce the PAPR of the signal. The presented hybrid approach provides PAPR reduction in hybrid frequency-quadrature amplitude modulation (HFQAM) based F-OFDM signal. The MATLAB 2021a working platform with Xilinx 14.5 is used to implement the proposed technique. The experimental outcomes of the suggested strategy are contrasted with those of other existing approaches.