Improvement of nulling performance in the presence of random variations in a linear sensor array

A method for analyzing and improving the nulling performance in a linear sensor array that is subjected to random variations in array weights and sensor positions is presented. In weight vector space, the nonorthogonal projection of a steering vector due to the random variations has been successfully adjusted to the nearly orthogonal projection by iterative adjustment of the initial optimum weight vector. As a result, an excellent nulling performance has been achieved. Computer simulation results are also presented.<>