Review of Target Tracking Algorithms Based on Bayes Filtering
Download as PDF
DOI: 10.23977/csic2022.015
Corresponding Author
Yu Yang
ABSTRACT
Target tracking technology is widely used in the military, traffic, medical, and other fields. Target tracking can be divided into single target tracking and multi-target tracking according to the number of tracks. This paper reviews the research on common algorithms involved in target tracking from the above two aspects. One of the cores of the research is to effectively filter the single target tracking by measurement, and accurately estimate the state of the target in real-time. This paper introduces the related algorithms based on Bayesian filtering that can solve this problem. For example, the Kalman filter is applied to a linear system, the extended Kalman filter is applied to the nonlinear system, and the particle filter. The core of this paper also includes the most important steps in multi-target tracking to solve the uncertain factors in observation samples and complete the correlation between observation samples and trajectory sets. Four algorithms are introduced to solve data correlation problems, such as the global nearest neighbor, joint probability data association, multivariate hypothesis tracking, probability hypothesis density filtering, etc. Finally, this paper summarizes various algorithms to solve target tracking in different situations, discusses their advantages and disadvantages, and puts forward the future problems of data association research.
KEYWORDS
target tracking, data association, Bayes filter, multi-target tracking