Multisensor Random Finite Set Information Fusion: Advances, Challenges, and Opportunities

In this chapter, we provide an overview of cutting-edge approaches and remaining challenges in multisensor multitarget information fusion based on the random finite set (RFS) framework. The fusion that plays a fundamental role in multisensor collaboration is classified into data-level multitarget measurement fusion and estimate-level multitarget density fusion, which share and fuse local measurements (typically the corresponding likelihood function) and posterior densities (including both cardinality distribution and multitarget localization distribution) between sensors, respectively. In particular, two optimization-oriented density-averaging approaches, namely arithmetic-average fusion and geometric average fusion, are addressed in detail for various RFSs. Important properties of each fusion rule including the optimality, sub-optimality, conservativeness, advantages and disadvantages are presented. Remaining challenges and emerging research topics such as the continuous-time trajectory modeling and estimation, and heterogeneous fusion are also highlighted.