基于CKF的特征辅助数据关联多目标跟踪

Multiple Target Tracking Based on CKF with Feature Aided Data Association

针对密集杂波环境下的多目标近距跟踪问题,提出了一种基于容积卡尔曼滤波(CKF)和特征辅助数据关联的多目标跟踪算法(FADA-CKF)。通过特征信息来对传统量测进行扩维,利用扩维后的量测对关联概率进行修正,将特征信息辅助技术融入到联合概率数据关联中,再利用容积卡尔曼滤波(CKF)处理非线性观测量,对目标状态进行估计。将FADA-CKF算法用于近距多目标跟踪场景中,仿真结果表明,改进算法在跟踪精度和误跟率方面要优于传统的JPDA跟踪算法。

For solving the problem of tracking multiple targets which are spaced closely under high dense clutter, an Featured Aided Data Association （FADA） tracking algorithm was proposed. The measurement was augmented by feature information, and the feature aid technology was integrated into Joint Probabilistic Data Association （JPDA） . Then nonlinear measurements were processed and the states of targets were updated by Cubature Kalman Filter （CKF）. Multiple targets tracking algorithm was proposed based on CKF and feature aided data association. The ex- perimental results show that the proposed algorithm improves the tracking accuracy and reduces the rate of lost target in comparison with traditional JPDA algorithm.