基于Bayesian-Fisher 混合模型的强跟踪当前统计模型

针对当前统计模型(CS)不能自适应调节机动参数,导致对弱机动以及强机动目标跟踪性能下降的问题,提出了一种基于Bayesian-Fisher 混合模型的新方法。首先,通过引入Bayesian-Fisher 混合模型,将机动加速度均值作为未知的确定性输入增广到状态变量中,实现了对加速度均值的在线自适应估计;其次,根据强跟踪滤波器(STF)的思想,引入时变渐消因子,增强算法对突变状态的适应能力。仿真结果表明,该算法不仅提高了对弱机动和强机动目标的跟踪精度,也削弱了对初始机动参数的依赖。

Strong Tracking Current Statistical Model Filter Based on Mixed Bayesian-Fisher Model

The current statistical (CS) model has poor performance for tracking weak and high maneuvering targets due to its short-coming in adjusting the maneuvering parameters. To solve this problem, a new method based on mixed Bayesian-Fisher model is proposed. By introducing the mixed Bayesian-Fisher model, the maneuvering acceleration is adaptively estimated as an additive input term in the corresponding state equation. Secondly, adopting the idea of the strong tracking filter (STF), the new algorithm introduces a multiple fading factor to enhance the stability to the sudden change state. The new algorithm not only improves the accuracy of weak and high maneuvering targets, but also eliminates the dependence on initial maneuvering parameters. The simulations test and verify the effectiveness of the algorithm.