基于卡尔曼滤波的数据融合算法与应用研究

针对无人机飞控系统对输入的多传感器信息融合时传统卡尔曼滤波算法容易出现滤波发散,滤波精度和系统的实时性降低的问题,研究了一种改进的自适应滤波算法,可以让数据融合后的信息精度更高,实时性更强。改进的算法是在Sage-Husa滤波的基础上引入滤波收敛性判据,并提出了基于改进的Sage-Husa滤波算法的联邦卡尔曼滤波器的设计,可以抑制滤波发散,提高滤波精度和稳定性。同时引入强跟踪滤波算法的思想,调整增益矩阵,改进滤波算法,提高系统突变情况下的滤波处理能力。最后,通过对特定的自主避障系统用改进后的算法与传统卡尔曼滤波算法进行MATLAB仿真比较,仿真结果显示改进的自适应滤波算法在系统模型参数失配和实变噪声情况未知时,可以较好地保持滤波的精度和实时性。

Research on Data Fusion Algorithm and Application Based on Kalman Filter

For the multi-sensor information fusion of the UAV flight control system,the traditional Kalman filter algorithm is prone to filter divergence,filtering accuracy and real-time degradation of the system. In this paper,an improved adaptive filtering algorithm is studied,which can make the information after data fusion more accurate and real-time. The improved algorithm introduces filter convergence criterion on the basis of Sage-Husa filter. The federated Kalman filter based on the improved Sage-Husa filter algorithm is proposed to restrain filter divergence and improve filter accuracy and stability. At the same time,the idea of strong tracking filter algorithm is introduced, the gain matrix is adjusted,and the filtering algorithm is improved,so as to improve the filtering processing capability under the condition of system mutation. Finally,the MATLAB simulation comparison between the improved algorithm and the traditional Kalman filtering algorithm for a specific autonomous obstacle avoidance system shows that the improved adaptive filtering algorithm can better maintain the filtering accuracy and real-time performance when the system model parameter mismatch and real-time noise are unknown.