Interaction-Aware Labeled Multi-Bernoulli Filter

Tracking multiple objects through time is an important part of an intelligent transportation system. Random finite set (RFS)-based filters are one of the emerging techniques for tracking multiple objects. In multi-object tracking (MOT), a common assumption is that each object is moving independent of its surroundings. However, in many real-world applications, objects interact with one another and the environment. Such interactions are rarely considered within the tracking process. In this paper, we present a novel approach to incorporate target interactions within the prediction step of a RFS-based multi-target filter, i.e. labeled multi-Bernoulli (LMB) filter. The proposed method is capable of explicitly incorporating the effect of target interactions within the filtering process, with little to no changes made for specific applications. We have developed the proposed filter for two practical applications, i.e. tracking a coordinated swarm and vehicles. The method has been tested for a complex vehicle tracking dataset and compared with the PHD filter and LMB filter through the OSPA metric, and with the LMB filter through OSPA(2) metric and cardinality error. The results demonstrate that the proposed interaction-aware method depicts considerable performance enhancement over the other methods in terms of the selected metrics.