Neural Network compensator based MMSE receiver for HPA nonlinearity in MIMO OFDM systems

Abstract

In this paper, we present a method based on Neural Network (NN) technique and accompanied with MMSE (Minimum Mean Square Error), which corrects at the receiver level, the Non-Linear (NL) distortions due to the HPA (High Power Amplifier). The neural network consists on a feed-forward Multi-Layer Perceptron (MLP) associated with Levenberg-Marquardt learning algorithm. The results show that the neural network compensator brings perceptible in a complete VBLAST MIMO OFDM (Vertical Bell Laboratories Layered Space-Time Multiple-Input Multiple-Output Orthogonal Frequency Division Multiplexing) system running under a Rayleigh fading channel.