论文部分内容阅读
无人机自主着陆过程中需要实时获得高精度的导航信息,对自主性、实时性的要求较高。现有的导航方式都存在各自的不足,且在室内等新型环境中不能使用。针对这一问题,提出了一种视觉/惯性组合导航算法。首先建立了世界坐标系下惯性导航的数学模型,随后通过Kalman滤波实现位置、姿态匹配,其中位置匹配完成速度误差、加表零偏的估计;姿态匹配完成安装误差角、陀螺漂移的估计,并利用估计得到的安装误差角和视觉导航系统输出的姿态信息对惯导姿态进行修正。仿真结果表明,该算法具有一定的工程应用价值。
The UAV needs real-time access to high-precision navigation information in real-time, which requires high autonomy and real-time performance. Existing navigation methods have their own shortcomings, and can not be used in new indoor environments. To solve this problem, a visual / inertial integrated navigation algorithm is proposed. First, a mathematical model of inertial navigation in world coordinate system is established. Then Kalman filtering is used to achieve the position and attitude matching. The position error is matched with the velocity error and the estimated bias is added. The attitude matching is used to estimate the error angle and gyro drift. The inertial navigation attitude is corrected by using the estimated installation error angle and the attitude information output by the visual navigation system. Simulation results show that the algorithm has some engineering application value.