Multi-target tracking algorithm in conditions of multipath radar reflection

This work presents a modified Poisson multi-Bernoulli mixture (PMBM) algorithm, adapted to perform trajectory processing tasks under conditions of multipath radar reflection. Based on the evaluative data of the previous kinematic state, size, and orientation of the target obtained using the extended PMBM algorithm, a pre-processing stage is carried out. During this stage, the origins of primary radar detections are initially classified. Then, after classification, these radar detections are further processed to estimate the state of targets at the next point in time. To assess the effectiveness of the proposed algorithm, a simulation modeling method based on MATLAB is used. With its help, scenarios for tracking distributed targets are modeled under conditions of the emergence of false signals from multipath radar reflections. The simulation results endorse the algorithm's capability to concurrently track multiple targets upon receiving primary radar detections from an assortment of radars amidst multipath radar reflections. The application of the algorithm finds use in the field of autonomous driving, driver assistance systems, applications ensuring driver safety, and other areas.