Analysis of modified SMI method for adaptive array weight control

The authors characterize the performance of the diagonally loaded sample matrix inverse (SMI) algorithm versus the number K of snapshots used in the covariance matrix estimate by providing O(1/K) statistics (bias and variance) of the array weights, output powers, and output power ratios such as signal to interference noise ratio (SINR) and INR. The approach accommodates wideband signals. Monte Carlo simulations verify the theoretical analysis.<
>