Comparative Analysis of CMA and MMSE in MIMO-OFDM System

Abstract

Channel estimation is one of the techniques used to achieve high data rates and low bit error rates in wireless communications. In wireless communication system, where Multiple Input Multiple Output - Orthogonal Frequency Division Multiplexing (MIMO-OFDM) exists, the effect of channel causes the received signal to be distorted which necessitates the receiver to have an insight of the channel known as the channel estimation. However, most of the existing techniques such as Least Square Error (LSE), Minimum Mean Square Error (MMSE) and Best Linear Unbiased Estimation Algorithm (BLUE) employ pilot symbols. High errors are observed in addition to computational complexity and in the platform other than MIMO-OFDM. In this paper, performances of Constant Modulus Algorithm (CMA) and MMSE are evaluated, and compared with each other in the 3x3 MIMO-OFDM systems. The system model for 3x3 MIMO-OFDM system incorporating each of CMA and MMSE consists of a transmitter, frequency selective channel and the receiver. 1000 bits are generated randomly and served as input signal. Three antennas configurations at the input of the frequency selective channel radiate the signal. The three antennas at the output of the channel receive the radiated power, processed by appropriate signal processing techniques. Each of MMSE and CMA techniques is performed at SNR of 5, 10 and 15dB. The system model is simulated using MATLAB 7.2 application package and evaluated using Mean Square Error (MSE) and convergence value. The results obtained show that CMA gives lower error than the MMSE and converges faster. Therefore, the study has shown the significant reduction in computational complexity and can be used by wireless design.

Keywords: Constant Modulus Algorithm, Orthogonality, Channel Estimation, Multiple Antenna, Cyclic Prefix.