Channel Capacity Estimation of 4 × 4 MIMO Antenna by Machine Learning, Considering SNR, Power Imbalance, and Correlation Coefficient

This study aimed to develop an accurate method for estimation of the channel capacity of 4 × 4 MIMO antenna systems by using a neural network. First, the channel capacity of an omnidirectional array antenna was calculated using the Monte Carlo method by changing the incident-wave signal-to-noise power ratio, power difference between the array elements, and correlation coefficient between the MIMO channels. Second, a neural network was used to evaluate the causal relation between the channel capacity and these parameters. The channel capacity predicted by the trained neural network matched well with the ones calculated using the Monte Carlo method.