Extended Kalman Filter Algorithm for Target Tracking

This document presents the Extended Kalman Filter algorithm and approximate grid filtering algorithm for target tracking applications. The Extended Kalman Filter algorithm, implemented in the main function file kal_demo.m, is primarily designed for tracking target position and velocity states. This implementation uses nonlinear state transition models and measurement models with linearization techniques to handle system nonlinearities. The approximate grid filtering algorithm, with its main function file bayes_demo.m, focuses on target localization tasks. For the grid partitioning approach, we employ a specialized method where the grid is centered around the target's previous position coordinates, with each grid cell dimension set to 1 unit and the total coverage area maintained at 5x5 units, creating a localized search region around the estimated position. Finally, we have developed an enhanced version of the algorithm that incorporates optimization techniques for improved computational efficiency and tracking accuracy, implemented in the new main function file trackiing_demo.m, which includes refined state estimation and measurement update procedures.