PAPR reduction of OTFS using an automatic amplitude reduction neural network with vendermonde matrix-based PTS and SLM algorithms

Abstract

Abstract Orthogonal time frequency selectivity is considered one of the most promising advanced waveforms for high-mobility conditions in beyond-fifth-generation networks. A high peak-to-average power ratio in orthogonal time frequency introduces nonlinearity and degrades the power amplifier performance. In this work, we propose a novel automatic amplitude reduction neural network algorithm combined with partial transmission sequence and selective mapping methods using the Vendermonde matrix for generating phase sequences. In a conventional selective mapping methods or partial transmission sequence, selecting an optimal phase increases the complexity, which is overcome via Vendermonde matrix in a proposed framework. The minimization of the peak-to-average power ratio is obtained at the double-stage process: in the first stage, the threshold amplitude is set, and the amplitude exceeding the threshold is reduced by using an automatic amplitude reduction neural network and then the partial transmission scheme or selective mapping methods. The simulation results reveal that, compared with the conventional algorithms, the proposed algorithms yield a significant peak average-to-power ratio, bit error rate, and power spectrum density performance.

Keywords

• OTFS
• Neural network
• PAPR
• B5G
• Vendermonde matrix
• PTS