基于多信息加权融合的降维航迹关联算法

Low Dimension Track Correlation Algorithm Based on Multi-Source Information Weighted Fusion

摘要:

针对分布式 3个传感器多目标的航迹相关算法如果直接计算时间和花费都比较高这一问题,提出降维航迹关联算法。该算法先利用 2个传感器的目标位置估计点构造航迹相关代价矩阵,求出最优解,再利用这个最优解与第 3个传感器的目标位置估计点构建航迹相关代价矩阵,进一步得到三维航迹相关配对。针对单信息系统不稳定这一问题,提出了融合多个特征信息的加权算法。该算法利用熵权法赋予各种不同信息的权重进行加权融合,转化为单信息问题。仿真结果说明本文所给出的新算法不仅减小了目标跟踪误差而且其时间花费较少,因此, 新算法是可行的。

Abstract:

Low dimension track correlation algorithm was proposed to solve the high cost and long time problem of trackcorrelation algorithm which aimed at distributed three sensors multi-target. This algorithm used, first of all, the estimatedpoint of the two sensors’target position to construct the cost matrix of track correlation, to obtain the optimal solution; thenreused this optimal solution and the estimated point of the third sensor’s target position to construct the cost matrix oftrack correlation, to further obtain the three-dimension track correlation pairing. To solve the instability problem of singleinformation source, in this paper, the algorithm of multi-information weighted fusion. This algorithm used the entropyweight method giving weight to information of various weighted fusion was proposed, transforming multi-information issueinto single information issue. The simulation results showed that the new algorithm proposed in this paper not only reducedthe tracking error of the target but also spent less time, demonstrating the effectiveness of the new algorithm.