无色卡尔曼滤波目标跟踪算法的尺度参数调整

Adaptation of Scaling Parameter for Unscented Kalman Filter Target Tracking Algorithm

利用无色卡尔曼滤波算法来研究目标跟踪过程中非线性随机系统的状态估计问题。在无色卡尔曼滤波算法过程中,目标状态的估计依赖于两个设计参数——尺度参数和方差矩阵式。针对尺度参数的不同选择将会影响整个目标状态估计的性能质量。为此论文提出构造四种不同的最优化性能准则函数,通过对此准则函数的最小化来迭代地自适应选择尺度参数。这四种准则函数从所使用观测信息的不同和尺度参数优化选择的计算复杂度来体现其各自的特性。最后通过仿真算例来验证论文所提出的无色卡尔曼滤波目标跟踪算法的尺度参数自适应调整策略。

The state estimation problem of the nonlinear stochastic systems is studied by means of the unscented Kalman filter algorithm from target tracking process.In the unscented Kalman filter algorithm,the state estimation is influenced by two design parameters—the scaling parameter and a covariance matrix.Because the choice of scaling parameter may lead to the increased quality of the state estimation.So here the four different criterion functions are constructed,and the scaling parameter is chosen adaptively by minimizing one criterion function.The property of each four criterion functions is shown from their own different observed information and computation complexity.Finally,the efficiency and possibility of the adaptation of scaling parameter for unscented Kalman filter target tracking algorithm is confirmed by the simulation example results.