基于有效预测区域的模糊数据关联

多目标多传感器跟踪系统由数据关联和目标状态估计两部分组成,数据关联是多目标跟踪系统研究的核心。数据关联和目标状态估计两部分既有一定的独立性又有密切的联系,而将两部分合理地结合对提高跟踪系统的性能是重要的。该文以跟踪目标的有效预测区域为依据,利用基于Mahalanobis距离的模糊均值聚类方法解决数据关联问题,在一定程度上将数据关联和目标状态估计两个不同的过程相结合,仿真计算说明了其有效性。

Fuzzy Data Correlation Based on the Predicted Validation Region

Multitarget-multisensor tracking systems consist of data correlation and state estimation. The multitarget tracking is made interesting by the data association problem. The data correlation and state estimation are both certainly independent and closely relative, but the performance of tracking systems can be improved by suitable incorporating the two components. In this paper, a fuzzy correlation approach is presented based on fuzzy clustering means algorithm with Mahalanobis distance. The approach, in a sense, fuses two different procedures of data correlation and state estimation. The simulation result using Monte Carlo method is given to demonstrate the efficiency of the new approach.