为提高GPS/DR组合导航系统航向估计精度,在卡尔曼滤波基础上提出了一种实时调节状态误差方差阵Q、观测误差方差阵R的自适应滤波方法;利用多重判决条件和规则约束,对惯性器件信息和滤波后输出的航向信息进行数据筛选。将此方法与在卡尔曼滤波中常用的Q,R取经验值法和Sage-Husa法进行比较,经实验仿真结果表明该方法能明显提高系统输出的航向精度,可获得更为准确的航向估计结果。

In order to improve heading estimation accuracy of the GPS/DR integrated navigation system,an adaptive filtering method was proposed based on Kalman Filtering Method,which could adjust state error variance matrix Q and the observation error variance matrix R in real time dynamically.On this basis,the data about inertial device and the output heading information after filtering were screened by use of multiple judging conditions and rule constraints.The experimental simulation results indicated that compared with conventional Q and R valued from the experiences and Sage-Husa adaptive filtering method,the method proposed here can improve the accuracy of heading estimation significantly for obtaining a more precise course estimation result.