一类基于改进的当前统计模型的目标跟踪算法研究

提出了一种基于改进当前统计模型的交互式多模型算法。在交互式多模型中,利用滤波过程中的新息和新息协方差的变化,对当前统计模型(CS)的机动频率自适应调整,使模型更适应实际。经理论分析Singer,CS模型的先验假设条件,提出了一种改进的匀速运动(CV)模型,以适应目标弱机动的情形,弥补CS模型在目标弱机动时跟踪精度不高的缺点。理论分析和仿真结果验证了提出算法的有效性。

Improved Current Statistical Model Algorithm for Maneuvering Target Tracking

An improved adaptive tracking algorithm for maneuvering target based on current statistical model was presented in this paper. In the interacting multiple model, the maneuvering frequency of the target of the current statistical model was adaptivdy adjusted by taking the advantage of the change of the bias information in the process of the filtering, which made the model more close to the fact. Based on the theory analysis on the priori assumptions of Singer and CS model, an improved constant velocity model was proposed to make up the weakness of the single model, which could offset the disadvantages of low tracking accuracy. The theory and Monte-Carlo simulation results showed that the method was valid.