Comparative performance analysis of a streamlined iteration cancellation technique for MIMO-OFDM systems with memoryless nonlinearity

We propose a streamlined iteration cancellation technique, reduced in complexity and implementation, to mitigate the nonlinear distortion for 2 by 2 Alamouti coded Multiple-Input Multiple-Output Orthogonal Frequency-Division Multiplexing (MIMO-OFDM) systems caused by memeoryless nonlinearity. The iteration cancellation technique based on two different estimators Minimum-Mean Square Error (MMSE) and Maximum-Likelihood (ML) estimator is presented, to form the basis for comparison to our modified approach. The modified reduced complexity iteration cancellation technique gives identical results to the ML estimator in terms of the Bit-Error Rate (BER), with significant savings in computational cost and implementation.

I. Introduction

A MIMO system combined with the OFDM modulation technique has been widely used in wireless communications as it increases the system's capacity, spectral efficiency and data rate whilst reducing the effect of fading due to increased diversity [1]. MIMO systems use multiple transmitter and receiver antennas to send parallel data streams at the same time through independent channels, enhancing spectral efficiency without increasing the bandwidth of the transmitted signal [2]. The main disadvantage of an OFDM signal is a high Peak-to-Average Power (PAP) ratio, which makes it more sensitive to nonlinear distortion [1], [3]. When the OFDM signal is amplified by the High Power Amplifier (HPA) which has nonlinear characteristic, the output spectrum exhibits bandwidth expansion and nonlinear in-band distortion [4]. To mitigate the nonlinear distortion a simple iterative cancellation method was presented in [4] and used in [5] on the Alamouti MIMO-OFDM systems with 2 transmitter and 1 receiver antennas. In both [4] and [5] the ML detector is utilized to make the final decision on the received signal. In this paper we extend the idea presented in [4] on 2 by 2 Alamouti coded MIMO-OFDM systems and analyse the performance of the ML and MMSE symbol detection techniques with the iterative cancellation algorithm. We propose a reduced complexity iterative cancellation algorithm for 2 by 2 Alamouti coded MIMO-OFDM systems where the ML detector is avoided and show that in terms of the BER values the same performance results can be achieved as with the iterative method presented in [4] and [5], but with significant reduction in computational complexity.