Journal of Xidian University

Previous Articles     Next Articles

Multiple maneuvering extended targets tracking with Gaussian process regression

LI Cuiyun1;WANG Jingyi1,2;JI Hongbing1   

  1. (1. School of Electronic Engineering, Xidian Univ., Xi'an 710071, China;
    2. Unit 95980, PLA, Xiangyang 441000, China)
  • Received:2016-12-02 Online:2017-12-20 Published:2018-01-18

Abstract:

In view of the complexity of estimating the shape of the extended target and the low accuracy in multiple maneuvering extended targets tracking in the clutters, a multiple maneuvering extended targets tracking algorithm with Gaussian Process Regression is proposed. First, the extension of targets is modeled as a star-convex model. Then, the concept of weights used in the multiple targets tracking algorithm is introduced to the single maneuvering extended target tracking algorithm to realize multi-targets tracking. Finally, the Gaussian Process Regression is used to estimate the shape for the extended target. Simulation shows that the proposed algorithm is capable of tracking multiple maneuvering extended targets in the same scene with different shapes, and outperforms the traditional non-ellipsoidal extended target tracking algorithm in the estimation precision and computing speed.

Key words: multiple maneuvering extended targets, star-convex models, Gaussian processes regression, shape estimation