Performance analysis of OFDM and GFDM techniques over additive white impulsive noise channels

Abstract

Generalized Frequency Dividing multiplexing (GFDM) technique due to overcoming orthogonal frequency division multiplexing (OFDM) drawbacks such as high out-of-band emissions and high sensitivity to time and frequency offsets, has received increasing interest as a waveform solution for the 5G networks and beyond. However, these two multi-carrier techniques suffer from high peak power to average power ratio (PAPR) and loss in bit error rate (BER) performance over challenging scenarios of impulsive noise channels. This paper investigates the performance of these two techniques from BER and PAPR viewpoints in the presence of two well-known impulsive noise models of Middleton Class A and Bernoulli-normal. In the case of BER, the results confirm that the more impulsive noise intensity is increased, the more loss in the BER of both techniques is caused. However, the GFDM technique is more robust against the adverse effects of such noises. In the case of PAPR, although the GFDM performance is worse than OFDM, we remarkably moderate this challenge by employing a precoder at the GFDM transmitter side and a noise mitigation method for its receiver.