Dynamic Measurement Noise Covariance Matrix R for Joint Probabilistic Data Association Filter

Abstract

In Multitarget Tracking (MTT), several targets of interest are being tracked simultaneously with the help of any optimal estimator. MTT tracking finds its applications in diverse fields like Pattern Recognition, Computer Vision, Radar Tracking, Robotics and many other research fields. In the literature, several algorithms have been implemented for MTT including Probabilistic Data Association Filter (PDAF), Joint Probabilistic Data Association Filter (JPDAF), Nearest Neighbor Standard Filter (NNSF), etc. JPDAF is the multitarget version of PDAF in which joint association probabilities are computed and tracks are then updated based upon these probabilities. Measurement noise covariance matrix R in JPDAF needs to be transformed from polar to Cartesian coordinate system. The optimal value of R should be calculated for the good performance of filter. In this thesis, measurement noise covariance matrix for JPDAF algorithm has been derived using standard radar parameters. 2D tracking is performed using scan radar and JPDAF algorithm. 3D tracking is also performed in a closed loop fashion using monopulse radar and JPDAF algorithm. For both 2D and 3D tracking, simulations are performed in MATLAB. Desired results are achieved and the error is reduced to such an extent that it lies inside the range bin for both cases.