分布式多传感器融合跟踪

Distributed Multi-sensor Fusion Tracking

针对传感器观测的非线性问题,引入了无迹卡尔曼滤波算法,它在稳定性和精确性等方面均高于卡尔曼滤波,进一步,针对其弱可观测性,采用多个传感器分布式融合跟踪策略。而协方差交集算法,它不需要考虑每个传感器之间的相关度,根据实际要求选择好合适的权值后,可以直接进行融合,并且达到很好的精度,于是,提出了基于协方差交集算法的分布式多传感器融合跟踪算法,最后分别以2个和4个被动站跟踪为例进行仿真研究,结果表明所提出的分布式融合算法能够达到很好的跟踪精度。

According to the nonlinear problem for the sensor system, Unscented Kalman filter （UKF） is introduced,the areas of the UKF such as stability and accuracy are higher than the Kalman filter. On the target tracking algorithm, multi-sensor fusion scheme is adopted to deal with the difficulty of observability problem for single sensor system, and the Covariance Intersection Algorithm （CI） yields consistent estimations irrespective of the actual correlations, and has good precision. Thus, this article presents distributed multi-sensor fusion tracking algorithm based on the CI, then an analysis of the experiment results is made and proved that the algorithm can reach a good tracking accuracy.