Decision Sequencing and Distributed Data Association

Multi-sensor multi-target tracking requires the solution to a challenging data association problem. While the processing architecture may be dictated by sensor network characteristics, distributed processing often provides benefits in both performance and robustness with respect to centralized processing. We cast the selection of fusion architecture as one of decision sequencing and study the benefits of competing architectures with respect to a probability of correct association criterion. We use a similar approach to study decision sequencing in forensic surveillance problems. Additionally, we propose a generalization to the Mori-Chang-Chong model to account for multiple-hypothesis association and discuss the impact of measurement latency.