论文部分内容阅读
针对点航关联在多目标跟踪中精度与实时性难兼顾的问题,提出了一种基于最小二乘拟合的点航关联算法。首先采用滑窗将历史航迹截断,采用最小二乘法在不同维度分别拟合、外推融合航迹历史信息条件下的航迹点,增加外推点的多样性及信息量。同时定义了5种全概率关联事件,提取传统滤波方法的预测点,将拟合外推点与滤波预测点融合,使归属判决更加准确。最后分别推导了不同事件发生时的状态更新方程与误差协方差更新方程,给出了其中参数的确定方法。经仿真数据验证,与经典的最近邻域法和联合概率数据互联算法相比,所提算法能够更好地兼顾精度与实时性,且计算复杂度较低,易于工程实现。
Aiming at the problem that the accuracy and real-time performance of point-ship association in multi-target tracking are both unattainable, a point-by-point algorithm based on least squares fitting is proposed. Firstly, the historical track is truncated by sliding window, and the least square method is used to fit and extrapolate the track points under the historical information of the track respectively in different dimensions to increase the diversity and information of the extrapolated points. At the same time, five kinds of total probability related events are defined, the prediction points of the traditional filtering methods are extracted, the fitting extrapolation points and the filtering prediction points are fused, and the vesting decision is more accurate. Finally, the state update equation and the error covariance update equation of different events are deduced respectively, and the method of determining the parameters is given. The simulation results show that compared with the classical nearest neighbor method and joint probability data interconnection method, the proposed algorithm can achieve better accuracy and real-time performance with lower computational complexity and easy implementation.