Extended Target-Tracking Algorithm Based on Variational Bayes and Axis Estimation Theory

Aiming at resolving the problem of low tracking accuracy for maneuvering extended target in lidar systems, an interactive multi-model variational Bayes independent axis estimation (IMM-VB-IAE) algorithm is proposed in this paper. Firstly, the algorithm utilizes IMM to adaptively select the appropriate model to track the target according to the changes of the target’s motion state, thereby improving the overall tracking performance. Secondly, the algorithm uses variational Bayes theory to approximate the posteriori closed expression, which simplifies the solution process by transforming the original complex inference problem into parameter optimization problem. Finally, the principle of eigen-decomposition is utilized to quadratically estimate the ellipse axes lengths, which improves the estimation accuracy of the ellipse parameters. The final simulation and experimental results demonstrate that the proposed algorithm outperforms several other algorithms in the accuracy of tracking maneuvering extended targets, with average OSPA and Gaussian Wasserstein distances reduced by at least 55.5% and 56.1%, respectively.