Posterior Cramér-Rao Lower Bounds for Extended Target Tracking with Gaussian Process PMHT

摘要

In practical target tracking scenarios with high-resolution sensors, targets often appear as extended targets with irregular and arbitrary shapes. In this paper, the posterior Cramér-Rao lower bounds (PCRLB) for extended target tracking with a Gaussian Process (GP) measurement model is derived to quantify the achievable accuracy of estimates of multiple extended target states within the Probabilistic Multi-Hypothesis Tracker (PMHT) framework in scenarios with clutter. Simulation results verify the effectiveness of the proposed PCRLB.