基于改进“当前”统计模型的非线性机动目标跟踪算法

A nonlinear maneuver-tracking algorithm based on modified current statistical model

中文关键词  机动目标跟踪  当前统计模型  活化函数  UKF

英文关键词  maneuvering target tracking  “current”statistical model  activate function  unscented Kalman filter

中文摘要

      “当前”统计模型算法对目标加速度极限值及机动频率的依赖性这一缺点, 导致该算法在跟踪弱机动目标及高机动目标时, 跟踪性能会明显恶化. 本文在当前统计(CS)模型的基础上, 通过加权一个以残差方差的迹为参数的活化函数, 对加速度方差和机动频率进行自适应处理, 实现了对“当前”统计模型的改进; 同时, 采用UKF(unscented Kalman filter)对非线性目标进行跟踪滤波, 构造出基于改进“当前”统计模型的非线性机动目标跟踪算法, 仿真结果表明, 改进的“当前”统计模型不仅保留了“当前”统计模型跟踪算法对一般机动目标跟踪精度高的特点, 而且对弱机动目标跟踪也具有更优越的性能, 大大扩大了机动跟踪范围.

英文摘要

      The ‘current’statistical model depends on both the marginal value of the target acceleration and the frequency of the maneuver; this leads to a poor performance in tracking targets of low maneuverability or higher maneuverability. To obtain an improved model, we introduce to the existing ‘current’statistical(CS) model an activate function with the trace of the residual error variance as parameter for modifying the error covariance between the acceleration and the frequency. This modified model is then combined with an unscented Kalman filter(UKF) to form the modified current statistic model for the nonlinear maneuver tracking algorithm. Simulation results indicate that the proposed algorithms not only keep equal performance level as the CS model in tracking general maneuvers, but provide excellent performance in tracking targets with low maneuverability; thus, extending the range of maneuverability in tracking targets.