一种改进UKF算法在超视距雷达中的应用

Application of Improvement UKF algorithm in Over-the-Horizon Radar

在雷达目标跟踪精度优化问题的研究中,针对传统的UKF算法在对天波超视距雷达目标跟踪中精度不高和存在发散等问题,提出了一种改进的UKF算法,在Sigma采样中引入调节因子,保证预测方差矩阵的半正定性,并且在滤波更新后,采用整体最小二乘法得到下一次更新的预测值,比最小二乘法的拟合误差更小,从而使重新经过观测模型利用卡尔曼滤波机制得到的滤波值优于未经过迭代的滤波值,以达到提高滤波跟踪精度的效果。仿真结果表明,通过径向距离误差和方位角误差的比较,改进的UKF算法在处理目标跟踪中,既可提高滤波跟踪的精度,又能有效地加快跟踪系统的收敛速度。

To resolve the problem of slow convergence and divergence in traditional UKF algorithm in target tracking, this paper put forward an improved UKF algorithm. In the improved method for Sigma sampling, we used the adjustment factor to ensure the prediction of positive semi definite covariance matrix, and after the filter was updated, we used the total least squares method whose fitting error is smaller than the least square method. In order to improve the tracking accuracy, the next update values were predicted. Based on the comparison of radial distance error and angle error, the simulation results show that the improved UKF algorithm can not only improve the accuracy of filtering tracking, but also speed up the convergence rate of the tracking system effectively in dealing with target tracking.