Nonlinear Distortion Cancellation using Predistorter in MIMO-GFDM Systems

Abstract

Generalized frequency division multiplexing (GFDM) with offset quadrature amplitude modulation (OQAM) is an alternative non-orthogonal modulation scheme for future generation wireless broadband systems. The nonlinearity of high power amplifiers (HPAs) has a very significant effect on the performance of GFDM systems. In this paper, we investigate the effects of nonlinear distortion on the multiple-input multiple-output (MIMO)-GFDM system when the signal is passed the HPA, which is modeled with amplitude and phase distortion. The effects of nonlinear distortion due to the HPA include amplitude distortion, phase distortion, and the spread of signal constellations. These effects also produce harmonic signals and intermodulation outside the frequency band which results in spectral spread. This will then reduce the performance of the MIMO-GFDM system. The contributions of this paper concern three key areas. Firstly, we investigate the effects of nonlinear distortion on the MIMO-GFDM system. We also derive the new closed-form expression bit error rate (BER) in MIMO-GFDM systems that use a memoryless HPA, which is modeled using the Saleh model when passed through the additive white Gaussian noise (AWGN) channel. This model was chosen because it is simple and has AM/AM and AM/PM curves. Secondly, we propose the application of techniques for the linearization of each HPA predistorter on the transmitter side of the MIMO-GFDM system separately. This predistorter is able to compensate for nonlinear distortion caused by the HPA without memory operating in the saturation region. The main contribution of this paper is to investigate the predistorter, which can linearize nonlinear distortion in MIMO-GFDM transmitters. The performance of the proposed scheme is evaluated in terms of spectrum analysis, PAPR analysis, a constellation diagram, and bit error rate (BER) analysis. The simulation results show that the proposed predistorter design succeeds in compensating for nonlinear distortions caused by the HPA for large input back-off (IBO) values.