基于递推加权最小二乘法的多目标跟踪算法

Multitarget Tracking Algorithm on the Recursively Weighted Least Square Rule

为了对多个机动目标进行跟踪,一些文献提出了交互多模型-概率数据关联算法,但这种方法必须建立大量的目标状态模型,如果目标的模型无法确定或建立的模型不准确,将影响这种方法的使用效果。为此,提出用最小二乘法拟合出目标的运动轨迹,避免建立大量的目标状态模型,导出了基于递推加权最小二乘法的JPDA算法的计算公式并进行了仿真验证,理论分析、MontoCarlo仿真以及在实际系统中应用表明了该算法的有效性。

The interacting multiple models probabilistic data association algorithm has been proposed to track multiple maneuvering targets, but, many state models have to be set up in this method. To solve the case of multiple maneuvering targets tracking in no exact models, the calculate formula of JPDA on the Recursively Weighted least Square Rule is set up and the tracks of targets will be estimated by the Recursively Least Square, substituting for multiple models. Theoretical analysis and Monte Carlo simulation show that this method is efficient.