Development of intelligent receiver for MIMO-OFDM system

Abstract

An intelligent receiver based on Radial Basis Function (RBF) neural network for MIMO system is developed. Due to its fast maximum-likelihood(ML) decoding, STBC is commonly used in MIMO system. However, most of the existing STBC methods rely on the availability of accurate channel state information (CSI) at the receiver. Furthermore, the complexity of the ML algorithm grows exponentially with the number of transmit antennas and constellation size. Especially when the number of transmit antenna is more than two, in order to enhance the symbol transmission rate, its complexity increases greatly. Therefore, intelligent receiver based on RBF neural networks is designed for 3 transmit antennas and 4 receive antennas MIMO system, in which a PCA approach is applied to process the train samples and online sequential extreme learning machine (OS-ELM) is adopted to adjust the parameters of the of RBF neural network. Compared with ML decoder, the proposed receiver has a high precision and good performance to track the variations of the fading channels. The result of simulation illustrates the effectiveness and feasibility of the receiver introduced.