A Collaborative Sensor Fusion Algorithm for Multi-Object Tracking Using a Gaussian Mixture Probability Hypothesis Density Filter
This paper presents a method for collaborative tracking of multiple vehicles that extends a Gaussian Mixture Probability Hypothesis Density (GM-PHD) filter with a collaborative fusion algorithm. Measurements are preprocessed in a detect-before-track fashion, and cars are tracked using a rectangular shape model. The proposed method successfully mitigates clutter and occlusion problems. In order to extend the field of view of individual vehicles and increase the estimation confidence in the areas where a target is observable by multiple vehicles, PHD intensities are exchanged between vehicles and fused in the Collaborative GM-PHD filter using a novel algorithm based on the Generalized Covariance Intersection. The method is extensively evaluated using a calibrated, high-fidelity simulator in scenarios where vehicles exhibit both straight and curved motion at different speeds.