一种新的多渐消因子容积卡尔曼滤波

A Novel Multiple Fading Factors Cubature Kalman Filter

摘要:

将强跟踪思想引入容积卡尔曼滤波（cubature Kalman filter，CKF），建立强跟踪CKF能有效克服CKF在模型不确定、状态突变等情况下，滤波性能下降的问题。通过分析现有多渐消因子计算方法，发现它们均只利用了协方差矩阵的对角线元素，并没有考虑各个状态之间的相关性，不能充分发挥多渐消因子的优势。为此，本文提出渐消因子矩阵，基于正交原理推导渐消因子矩阵的求解方法，提出多渐消因子强跟踪CKF算法。多渐消因子强跟踪CKF算法突破了传统多渐消因子为向量的限制，也不再要求渐消因子取值要大于1。仿真验证了算法具有更好的滤波精度何鲁棒性，能更好的满足工程应用的要求。

Abstract:

The strong tracking idea is introduced into the cubature Kalman filter (CKF), and the strong tracking CKF can effectively overcome the performance degradation problem of CKF under the condition of model uncertainty and state mutation. By analyzing the existing multiple fading factor calculation methods, it is found that they only use the diagonal elements of the covariance matrix, and do not consider the correlation between the states, and cannot give full play to the advantages of multiple fading factors. Aiming to solve this disadvantage, this paper proposes a novel fading factor matrix, based on the orthogonal principle to derive the solution method of the fading factor matrix, and proposes a multiple fading factors strong tracking CKF algorithm. The multiple fading factors strong tracking CKF algorithm breaks through the limitation of the traditional multiple fading factor as a vector, and does not require the value of the fading factor to be greater than 1. The simulation verifies that the algorithm has better filtering accuracy and robustness, and can better meet the requirements of engineering applications.