Application of Kalman Filtering

Abstract

This chapter is devoted to the estimation of nondeterministic quantities with special focus being on estimation of the state of a dynamic system, e.g., the location and orientation of a mobile robot, and also estimation of the coordinates of a detected object of interest. The Kalman Filter is presented and utilized to a large extent. Simulations are used to illustrate the capability of this methodology. The development of the Kalman Filter will begin with a well-known estimation problem, that of estimating a fixed quantity using batch processing. The chapter illustrates the way the filter allocates weighting on the measurement versus weighting on the propagation of the model. The confidence in each depends on the respective measurement noise and process disturbance covariances. By using the Kalman Filter, it is possible to obtain estimates having lower error variance than the measurement itself.