论文部分内容阅读
针对传统Kalman滤波算法中噪声协方差取值固定容易制约滤波性能的缺点,提出一种基于差分进化(DE)算法的Kalman(DE-Kalman)滤波方法。该方法将估计误差的均方根作为适应度函数,利用差分进化算法选择滤波过程中最优的噪声协方差。针对尾矿库浸润线的动态特性,建立了常速和常加速两种机动目标跟踪中常用的系统状态模型,分别用Kalman和DE-Kalman滤波算法进行状态估计,实验结果证明了DE-Kalman滤波算法在浸润线监测信息处理中的有效性。
Aiming at the shortcomings of the traditional Kalman filtering algorithm which is easy to restrict the filtering performance of the noise covariance, a Kalman (DE-Kalman) filtering method based on differential evolution (DE) algorithm is proposed. The method uses the root mean square of the estimation error as a fitness function and selects the best noise covariance in the filtering process by using the differential evolution algorithm. Aimed at the dynamic characteristics of the wetting line of tailing pond, a common system state model for two kinds of maneuvering target tracking under constant speed and constant speed is established. The Kalman and DE-Kalman filtering algorithms are respectively used to estimate the state. The experimental results show that DE-Kalman filter The effectiveness of the algorithm in monitoring wetting line information.