Data Association as a Core Technology in Multi-Target Tracking

Data association is a critical technology in multi-target tracking. While JPDA is widely recognized as a high-performance algorithm assuming one-to-one measurement-to-target associations, real-world scenarios often involve many-to-many relationships. This paper introduces the Generalized Probability Data Association (GPDA) algorithm to address these complex cases. Theoretically analyzes both algorithms' performance and conducts comparative simulations using Monte Carlo techniques, demonstrating GPDA's superior handling of complex association scenarios.