An M-Estimation EKF-CMAC-Based Algorithm for Robust Position Estimation and Guidance

Tracking a target in a mixed line-of-sight (LOS)/non-LOS (NLOS) environment is a challenging nonlinear uncertain problem. Its difficulty is compounded by the condition that received measurements are probabilistically corrupted by known LOS noises and unknown NLOS biases and noises. Time of arrival (TOA) is the target detection method considered herein. To enhance the accuracy of locating and guiding a target moving along a pre-specified trajectory, we propose an M-estimation based extended Kalman filter with cerebellar model articulation controller (M-EKF+CMAC) algorithm. The M-EKF processes received measurements and utilizes M-estimation for yielding estimated results. The outcomes of the M-EKF are fed into the CMAC for adaptively reducing tracking errors. Simulation results illustrate that our proposed algorithm is efficient and superior to the EKF-CMAC method in the literature.