Research on Multi-high-speed and High-Maneuverability Target Tracking Method Based on GM-PHD Filter

For the problem of multi-target positioning and tracking of high-speed and highly maneuverable targets, a Probability Hypothesis Density (PHD) filter algorithm, which can simultaneously estimate the number of targets and their positions, is commonly used. Among PHD filter algorithms, the Generalized Multi-Object Probability Hypothesis Density (GM-PHD) filter is the most widely adopted due to its computational simplicity. However, the currently used GM-PHD filter has its limitations, and when applied to long-distance positioning of high-speed and highly maneuverable targets, the original GM-PHD filter algorithm may introduce new errors. To address this issue, this paper proposes improvements to the GM-PHD filter algorithm by eliminating the merging step, making it more suitable for the positioning and tracking of high-speed and highly maneuverable targets. Simulation results demonstrate that the improved filtering algorithm reduces the tracking error to 96% of the original error.