Efficient algorithm with WINNER II lognormal distributions for MIMO systems

Genetic algorithm (GA) has achieved high interest in the field of wireless communication due to its outstanding search strength and highly design component steps. For multiple-input/multiple-output (MIMO) wireless communications systems employing spatial multiplexing transmission, we evaluate the effects of initialization for the performance and complexity of genetic algorithm (GA)-based detection, against the maximum-likelihood (ML) approach. We consider transmit-correlated low rank Rician fading with realistic Laplacian power azimuth spectrum. The values of the azimuth spread (AS) and Rician K-factor are set to the means of the lognormal distributions obtained by widely recognized WINNER II channel models. We first confirm that GA coincides with ML throughout the SNR points without incest prevention. Then, we compared the GA performance in WINNER II average over random AS and K values with average AS and K values and finally, compared GA-based detection complexity with conventional ML detection. We find that, achievable performance average over random AS and K values are lower than the average AS and K values, GA complexity is much lower than the ML, and thus, is an advantage in field programmable gate array (FPGA) design.