Research Article Open Access

A Noval Solution of Implementation Issues of Kalman Filter for Tracking the Targets

Abstract

The Kalman filter gives a linear, unbiased and minimum error variance recursive algorithm to optimally estimate the unknown state of a system from noisy data taken at discrete real-time. This paper present the practical aspects of implementing Kalman filter estimator applied to time varying stochastic non-linear models. The conventional implementation of the Kalman filter is particularly sensitive to round off errors, errors in the linearization process and ill conditioning in connection with matrix inversion. To reduce the complexity, the methods based on UD factorizations, aiming to simplify the update of the sate co-variance matrix P=UDUT. During the update of P, there is no need to find the FK , the Jacobian of the state transition matrix. Simulation of linear non linear target trajectory is performed using MATLAB R2009b. The result shows that the relative improvement in the convergence achieved using the UD factorization method.

Beena M Varghese, Sija Gopinathan, Daisykutty Abraham

To read the full article Download Full Article | Visit Full Article