Fuzzy Kalman Filter based trajectory estmation

This paper presents an algorithm of fuzzy based Kalman filter for trajectory estimation of dynamical objects. The Fuzzy subsystem is designed to tune dynamically the process noise covariance matrix of the discrete time Kalman Filter. The main adaptation strategy is based on the heuristic knowledge/practical expertise of the human observer/control engineer. The Fuzzy Kalman Filter attempts to offset some of the assumptions made in the original discrete Kalman Filter formulation. In order to illustrate the proposed algorithm, the state estimation of a Weather Balloon is considered, in which the noises affecting the system are highly non-stationary. The performances of the Fuzzy Kalman Filter is compared with existing Discrete Time Kalman filter.