ALGORITHM FOR MULTIPLE EXTENDED TARGET TRACKING

This paper introduces a novel algorithm for multi–extended object tracking, which is based on probabilistic–statistical modelling, machine learning, and optimization theory. The proposed approach encompasses methods for modelling the state of extended objects, data association, and predicting and updating the state of extended objects at each time step. The performance of the algorithm was assessed and compared with other existing algorithms using simulation modelling in MATLAB. This simulation emulates scenarios in which extended objects are tracked while moving under various noise levels. Simulation results demonstrate that the algorithm can effectively detect and track the trajectories of multiple extended objects under different noise levels, utilizing both linear and non–linear measurements from an array of sensors. Furthermore, the algorithm estimates not only the number and kinematic states of the targets but also provides an approximate evaluation of their size. The performance of the proposed algorithm, in terms of accuracy in target number and state estimation, is superior to that of comparable algorithms, particularly in high–noise environments. Potential applications of this algorithm include tracking moving targets in the context of autonomous driving, driver assistance, safety assurance, and various other domains.