Liu Yu, Liu Jun, Li Gang, Qi Lin, Li Yaowen, He You
School of Electronic and Information Engineering, Beihang University, Beijing 100191, China.
Research Institute of Information Fusion, Naval Aeronautical and Astronautical University, Yantai 264001, China.
Sensors (Basel). 2017 Nov 5;17(11):2546. doi: 10.3390/s17112546.
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.
本文聚焦于非线性杂波环境下多传感器多目标的跟踪问题。为避免雅可比矩阵计算和缩放参数调整，提高数值稳定性，并获得更准确的集中式非线性跟踪估计结果，提出了一种新颖的集中式多传感器平方根容积联合概率数据关联算法（CMSCJPDA）。首先，将多传感器跟踪问题分解为若干单传感器多目标跟踪问题，在估计过程中依次进行处理。然后，在每个传感器中，基于联合概率数据关联（JPDA）完成其测量值到目标轨迹的分配，并利用带有容积卡尔曼滤波器平方根版本（SRCKF）的加权概率融合方法来估计目标状态。在处理完所有传感器的测量值后，得出CMSCJPDA并获得全局估计状态。实验结果表明，CMSCJPDA在跟踪精度、数值稳定性和计算成本方面优于现有算法，为解决多传感器跟踪问题提供了新思路。
