针对自主驾驶车辆长时间导航精度要求难以满足的问题,建立了GPS与微惯性导航系统的组合导航滤波模型,在位置观测的同时引入姿态信息,提高了导航精度。在此基础上提出了基于权值矩阵的模糊自适应卡尔曼滤波算法,该算法通过模糊控制器自适应地改变每个观测量的权值,得到权值矩阵引入卡尔曼滤波器实现自适应滤波。仿真和实验结果表明,所提出的权值矩阵模糊卡尔曼滤波性能优于衰减因子自适应卡尔曼滤波,特别是在GPS信号失真及噪声先验统计特性不可知的情况下,其定位精度能够保证在1m之内。
The model of the integrated INS/GPS navigation system was established to provide long-term highaccuracy navigation information. The navigation accuracy was improved greatly by importing the attitude information to the observation. Besides, the fuzzy adaptive Kalman filtering algorithm based on weighted matrixes was proposed. The algorithm adaptively changes the corresponding weighted factor via fuzzy logic for every observable, and the weighted-matrixes, instead of the single factor, was used to adjust the Kalman filter realizing adaptive filter. The simulation and experiment results show that the performance of the proposed algorithm is better than that of the fading memory adaptive Kalman filter, especially at the situation that the information of GPS is lost or the prior statistics is known insufficiently. It can ensure the navigation error within 1 m, which is small enough for the autonomous driving.