Advanced data association technique using integrated track splitting filter for multi-target tracking in clutter and occlusion

Multi-target tracking (MTT) in cluttered environments with occlusion is a complex challenge due to the uncertain number of targets, their motion variability, and measurement ambiguities. Traditional MTT methods employing joint data association such as joint integrated track splitting (JITS), often encounter prohibitive computational costs due to the exponential growth of measurement-to-track associations. This paper presents a modified integrated track splitting (MITS) algorithm with an advanced data association technique. The MITS algorithm enhances tracking performance by adjusting the data association probabilities based on track likelihood ratios, detection and gating probabilities which significantly decreases the number of false measurement detections. In addition, the MITS refines the validation gate by varying gating threshold to improve measurement selection accuracy. Monte-Carlo simulations demonstrate that MITS achieves approximately 91% reduction in computational complexity and a 25% improvement in false track discrimination (FTD) compared to standard ITS and JITS algorithms, particularly in scenarios with multiple cross-over targets and measurement occlusion.