Modified Zhang neural algorithm for time-varying overdetermined linear models and applications

Abstract

The challenges of the inefficiency of the existing static least-squares method and the insufficient precision of the existing neural algorithms need to be overcome. With the aid of the ten-instant Zhang time discretization formula, a novel ten-instant discrete-time neural network algorithm is proposed and generalized to the problem-solving of overdetermined system of time-varying linear equations. Theoretical analyses illustrate that the maximal steady-state residual error of the proposed algorithm has an order-6 pattern. Comparative experiment results further substantiate the superiority of the proposed algorithm. Moreover, applications of mobile localization based on the angle of arrival technique and the time difference of arrival technique are provided to display the applicability of the proposed algorithm.