改进的EKF算法在目标跟踪中的运用

过程噪声和测量噪声影响Kalman滤波的性能,通常很难得到它们准确的值。提出观测噪声和过程噪声实时估计的自适应算法。该算法可以用在非线性和机动目标跟踪问题中,不必预先知道准确的噪声方差。重新估测观测噪声方差矩阵,可以较好地消除由观测噪声带来的误差;建立一个简单的线性Kalman滤波器对过程噪声进行实时估计,这对于机动目标来说是必要的,因为原有的过程噪声将受到加速度影响,不能包含全部的信息。实验表明,该算法保证EKF稳定性,提高了跟踪性能。模拟实验300次后,X,Y方向位置均方误差分别为7.8099,9.6838。

Application of the improved EKF algorithm in target tracking

Knowledge of the process and measurement noise covariance matrix is important for the application of Kalman filter.Hoverer,it is usually a difficult to obtain an explicit matrix for nonlinear time-varying system.An adaptive algorithm for estimating measurement noise and process noise at real time is proposed.The algorithm can be applied to tracking nonlinear and maneuvering targets without knowledge of the certain noise covariance in advance.The error induced by measurement noise can be effectively eliminated by modificatory measurement noise variance matrix.A simple linear Kalman filter for real-time evaluation of process noise is constructed.This is necessary for maneuvering targets because the original process noise will be affected by acceleration and it cannot include all the information.Experiments show that EKF stability can be ensured with the algorithm and tracking performance can be improved.