基于无迹卡尔曼滤波的农用无人机定位研究

Research on the agricultural unmanned aerial vehicles positioning based on unscented Kalman filter

为提高农用无人机在作业时的定位精度,提出应用全球定位系统及惯性导航系统信息融合的方法实现无人机位姿状态的建模,得到导航系统测量信号与无人机状态间的非线性微分方程。对于系统中存在的非线性状态估计问题,创新性的提出采用适合于非线性系统的无迹卡尔曼滤波算法(Unscented Kalman Filter,UKF)进行处理,实现对基于多传感器信息融合的无人机姿态(翻滚角、俯仰角、偏航角)、速度和位置的准确而稳定的估计。现场试验采用改造升级后的极飞科技P102018植保无人机,配备改造升级后的机载电子和传感器系统。试验结果表明,与常见的扩展卡尔曼滤波器相比,UKF与多传感信息融合技术结合可实现对无人机位置信息(欧拉角)的高精度估计,其翻滚角、俯仰角和偏航角误差估算准确度分别提高30.3%、45.8%、70.2%,绝对值最大为0.57°。

In order to improve the positioning accuracy of Agricultural Unmanned Aerial Vehicles(UAVs),the multi-sensor information fusion with global positioning system and inertial navigation system is used to construct the model of UAV,and the nonlinear differential equations between the measurable signals of navigation systems and the states of UAV.The Unscented Kalman Filter which is suitable for nonlinear system is innovatively proposed to solve the nonlinear states estimation problem,and realize accurate estimation of UAV’s attitude,speed and position(roll angle,pitching angle and yaw angle).In the field test,the modified and upgraded Plant Protection UAV is equipped with the modified and upgraded airborne electronics and sensor system.The test results show that compared with the common extended Kalman filter,UKF combined with multi-sensor information fusion technology can realize the high-precision estimation of UAV position information(Euler angle),and the error estimation accuracy of roll angle,pitch angle and yaw angle can be improved by 30.3%,45.8%and 70.2%respectively,with the maximum absolute value of 0.57°.