Multi-Sensor Multi-Target Tracking Systems
Daniel J. Clancy · 2025
The objective of this chapter was to introduce the concepts, components, mathematics, and algorithms necessary to understand and possibly perform MSMTT. The discussion focused on discrete-time, SSSMs composed of an SSDM and a SOM, and typical target dynamics models and observation models used for radar target tracking were presented. Gaussian distributions and the estimation of second-order statistics underpinned the mathematical foundation presented in this chapter. A hybrid system tracking and fusion architecture consisting of both centralized, M2T and decentralized, T2T processing components was presented, and some of the primary algorithms required by each type of fusion engine were presented in detail. The two primary components discussed were KSE and DA, along with the interplay between them. The KSE discussion for the M2TFS focused on four variants of the KF, namely, linear, iterated, extended, and sigma point, based on using the TSE for the first three, and the SPT for the last one. The IMM was also discussed in some detail. For the T2TFS, the ideal information filter track update was presented along with the CI algorithm. The versatility of the CI algorithm was discussed, and the behavior was compared and contrasted with the naïve fusion approach.