一种适于快速实现的卡尔曼滤波算法的改进

卡尔曼滤波算法具有收敛平稳、速度快、精度高的优点,在信号处理中受到广泛应用.但由于卡尔曼滤波算法不仅是一种迭代算法,而且还包含矩阵的连乘运算和求逆运算,这使得该算法在MCU和DSP这类数字平台上的实现面l临很大的困难.本文介绍了一种数字保护中卡尔曼滤波算法的改进,针对数字保护的特点,对算法的结构进行适当的变换,将迭代过程中大量与采样点数据无关的矩阵运算和求逆运算分离出来进行离线运算,从而避开难以实现的矩阵求逆运算和运算量巨大的矩阵连乘运算,大大简化了迭代的过程.由于离线运算不受运算量、运算时间和位宽精度的限制,因而改进后的算法不仅克服了原先迭代过程的复杂性,而且进一步提高了运算精度,可以在MCU和DSP这类数字平台上实现.

The Improvement of Kalman Filtering Algorithm in Fast Implementation

The advantage of Kalman filtering algorithm is shown by the stabilization,the speed and the accuracy of convergence.It is applied widely in signal processing.However,Kalman filtering algorithm not only includes iterative computation but also consecutivematrixes-multiplication and inverse matrixes computation.So some very great difficulties are faced in digital implementation of the algorithm based on MCU or DSP system.An improvement of Kalman filtering algorithm in digital protection is introduced in this p...