The invention provides a rotor unmanned aerial vehicle inertial vision landing navigation method which comprises the following specific steps: 1, cooperative target coordinate system error modeling: calculating a displacement amount on a geographic coordinate system by an inertial navigation system, converting the displacement amount into a cooperative target coordinate system, performing conversion by adopting a Cna attitude transfer matrix, and then performing cooperative target coordinate system error modeling; error modeling is carried out under the cooperative target coordinate system; 2, fusing the inertial navigation information and the visual navigation information: fusing the inertial navigation information and the visual navigation information through a Kalman filtering algorithm; 3, a system measurement noise matrix fitting method: calculating a relative distance between the rotor unmanned aerial vehicle and the cooperative target, and fitting a measurement noise matrix Rk; 4, position error correction: correcting various errors of inertial navigation; the position information of the cooperative target does not need to be acquired in advance, the technical problem that the visual navigation precision changes along with the distance between the unmanned aerial vehicle and the cooperative target is solved, and the visual navigation precision of the unmanned aerial vehicle is improved.
本发明提供一种旋翼无人机惯性视觉着陆导航方法,该导航方法的具体步骤如下:第一步,合作目标坐标系误差建模:将惯性导航系统解算地理坐标系上的位移量转换至合作目标坐标系中,采用Cna姿态转移矩阵进行转换,然后,在合作目标坐标系下进行误差建模;第二步,对惯性导航与视觉导航信息进行融合:通过卡尔曼滤波算法对惯性导航与视觉导航信息进行融合;第三步,系统量测噪声阵拟合方法:计算旋翼无人机与合作目标间的相对距离,对量测噪声矩阵Rk进行拟合;第四步,位置误差修正:即对惯性导航各项误差进行修正;本发明实现了不需要提前获取合作目标位置信息,解决了视觉导航精度随无人机相对合作目标距离变化的技术问题,提高了无人机视觉导航精度。
Rotor unmanned aerial vehicle inertial vision landing navigation method
一种旋翼无人机惯性视觉着陆导航方法
2024-05-14
Patent
Elektronische Ressource
Chinesisch