作 者：
刘进一;杜岳峰;张硕;朱忠祥;毛恩荣;陈雨
单 位：
中国农业大学现代农业装备优化设计北京市重点实验室
关键词：
农业机械;导航;信息融合;全球导航卫星系统;微机械惯性单元;扩展卡尔曼滤波
摘 要：
为解决实际农田环境中农业机械自动导航系统的位姿信息易受跟踪卫星数动态变化、天线遮挡、数据传输错误等因素影响,造成定位精度和稳定性变差的问题,采用全球导航卫星系统(GNSS)、微机械惯性测量单元(MIMU)及航位推算(DR)相融合的组合导航定位系统,建立了农业机械两轮运动学定位模型,针对系统状态噪声不稳定的因素,提出了一种根据当前估计值与预测值的偏差自适应地调节系统状态协方差阵的扩展卡尔曼滤波算法,并分别基于静止状态和直线导轨运动状态进行了试验验证。试验结果表明,在静止状态和直线导轨运动状态下,采用自适应滤波算法的航向角平均值绝对偏差分别为0.001 4°、0.024 5°,标准差分别为0.047 4°、0.251 1°;位置距离平均偏差分别为0.003 7 m、0.007 6 m,标准差分别为0.001 0 m、0.004 4 m,与固定值滤波算法相比,自适应滤波算法的各项评价指标得到了明显改善,进一步验证了算法改进的必要性和优越性。
译 名：
Automatic Navigation Method for Agricultural Machinery Based on GNSS/MIMU/DR Information Fusion
作 者：
Liu Jinyi;Du Yuefeng;Zhang Shuo;Zhu Zhongxiang;Mao Enrong;Chen Yu;Beijing Key Laboratory of Optimized Design for Modern Agricultural Equipment,China Agricultural University;
关键词：
agricultural machinery;;navigation;;information fusion;;global navigation satellite system;;miniature inertial measurement unit;;extended Kalman filtering algorithm
摘 要：
The real-time accurate position update technology of agricultural machinery was one of the most important studies in agricultural machinery automatic navigation system,and it has a very important significance to improve the efficiency of intelligent agricultural production. However, in the field operation of agricultural machinery automatic navigation system,the situations that the number of satellite was unstable,the GPS signal was blocked,and data transmission was wrong would cause low location precision and poor stability. In order to solve the above problems,a two dimensional kinematic model of agricultural machinery was built,and an adaptive extended Kalman filtering algorithm which adjusted the system's state covariance was carried out based on the integrated navigation system of GNSS / MIMU / DR.The algorithm was used to compute the difference of present estimate and predictive value. When the difference became greater,it showed that the system state had changed greatly,so as to make appropriate adjustments to the system state covariance matrix for better filtering. The algorithm was verified by static test and linear guide rail test to accurately evaluate the accuracy of the integrated navigation and positioning system. The tests indicated that: in static state,the average value deviation of heading was0. 001 4°,the maximum deviation was 0. 099 8°,the standard deviation was 0. 047 4°,and the position average deviation was 0. 003 7 m,the maximum deviation was 0. 008 1 m,the standard deviation was0. 001 0 m; in straight rail state,the average value deviation of heading was 0. 024 5°,the maximum deviation was 0. 432 4°,the standard deviation was 0. 251 1°,and the position average deviation was0. 007 6 m,the maximum deviation was 0. 018 6 m,the standard deviation was 0. 004 4 m. All the different evaluations proved that the adjusted filtering was superior to the traditional filtering,which indicated the necessity and superiority of the proposed algorithm. At the same time,it is proved that the method can satisfy the accuracy and stability requirements of the agricultural machinery navigation and positioning system.
