Centralized Multi-Sensor Square Root Cubature Joint Probabilistic Data Association

Sensors

卷:17

Yaowen Li¹  Gang Li¹  Jun Liu²  Lin Qi²  You He²  Yu Liu³ 

[1] Department of Electronic Engineering, Tsinghua University, Beijing 100084, China;

[2] Research Institute of Information Fusion, Naval Aeronautical and Astronautical University, Yantai 264001, China;

[3] School of Electronic and Information Engineering, Beihang University, Beijing 100191, China;

关键词: multi-sensor tracking;    data association;    cubature Kalman filter;    state estimation;    centralized filtering;   

DOI  :  10.3390/s17112546

来源: DOAJ

【 摘 要 】

This paper focuses on the tracking problem of multiple targets with multiple sensors in a nonlinear cluttered environment. To avoid Jacobian matrix computation and scaling parameter adjustment, improve numerical stability, and acquire more accurate estimated results for centralized nonlinear tracking, a novel centralized multi-sensor square root cubature joint probabilistic data association algorithm (CMSCJPDA) is proposed. Firstly, the multi-sensor tracking problem is decomposed into several single-sensor multi-target tracking problems, which are sequentially processed during the estimation. Then, in each sensor, the assignment of its measurements to target tracks is accomplished on the basis of joint probabilistic data association (JPDA), and a weighted probability fusion method with square root version of a cubature Kalman filter (SRCKF) is utilized to estimate the targets’ state. With the measurements in all sensors processed CMSCJPDA is derived and the global estimated state is achieved. Experimental results show that CMSCJPDA is superior to the state-of-the-art algorithms in the aspects of tracking accuracy, numerical stability, and computational cost, which provides a new idea to solve multi-sensor tracking problems.

【 授权许可】

Unknown