Multiple Extended Target Tracking Algorithm with Adaptive Target Birth Intensity

A Multiple Extended Target Tracking (METT) algorithm using the Star-convex Random Hypersurface Model (SRHM) in cooperation with Adaptive Target Birth Intensity (ATBI) is proposed in this paper to solve the problem that it is difficult to obtain the prior of the extended target newborn intensity in complex environments. Firstly, the Cardinalized Probability Hypothesis Density (CPHD) filter is used to establish a Bayesian filter framework for multiple extended targets. Then the SRHM is used to model the measurement distribution of the extended target, and the CPHD filter flow is embedded by the Unscented Transform (UT). In addition, based on the ATBI algorithm, the intensity of the new target is reversely derived from the measurement data, and the filter no longer depends on the prior information of the new target. The numerical experiment results verified that the SRHM-CPHD filter based on ATBI algorithm has better performance than the traditional Gaussian Mixture (GM) model for METT in the unknown newborn target intensity scene, and is more sensitive to the phenomenon of target newborn.