Target Tracking Based on Kalman Filtering Techniques

Abstract

Target tracking techniques aim at estimating desired target states online overtime. This paper introduces and derives algorithms for both single and multiple targets tracking, most of which are based on Kalman filtering technique. We demonstrate that multiple target tracking is more complicated that single target tracking in that data association techniques are required to match estimate target tracks with true targets. We present simulation examples of applications of both single and multiple target tracking in MAT LAB to demonstrate and analysis the performance of our introduced algorithms. In both tests, moving targets are successfully tracked via RGB camera with object recognition algorithms to extract the target features.