基于多传感器融合的高精度无人机空中对接方法
A High-Precision Aerial Docking Method for Drones Based on Multi-sensor Fusion
向玉龙 1李嘉诚 1陈芊 1盛汉霖1
作者信息
- 1. 南京航空航天大学能源与动力学院,南京 210016
- 折叠
摘要
多旋翼无人机空中对接技术可以解决多旋翼无人机续航能力有限、载荷小等问题,对提高多旋翼无人机的续航能力和任务效率有重要意义.但是当前多旋翼无人机空中对接方案成本高昂且对接精度差,易受环境限制.本文提出了一种基于单目视觉与惯性测量单元(Inertial measuring unit,IMU)融合定位并结合最优轨迹控制的空中对接方案,在保证对接精度的同时大幅降低成本.首先相机通过AprilTag对目标进行识别定位,再通过非线性优化对重投影误差进行最小化优化,并借助状态误差卡尔曼滤波器(Error state Kalman filter,ESKF)更准确地消除IMU的零偏,提高融合定位的精度;然后根据线性二次型控制器(Linear quadratic regulator,LQR)设计了最优对接控制器,并结合无人机的动力学模型,直接输出多旋翼无人机的姿态控制量,实现更加安全的空中对接;最后在四旋翼无人机平台上进行实验验证,顺利完成了空中对接任务,证明了该方法的可行性.
Abstract
The aerial docking technology for multirotor drones can address the issues such as limited endurance and small payload capacity,which is of great importance for extending the endurance and improving the task efficiency of multirotor drones.However,the current airborne docking solutions for multirotor drones suffer from high cost and poor docking precision,and are easily affected by environmental constraints.Therefore,this paper proposes an airborne docking solution that integrates monocular vision with an inertial measurement unit(IMU)for positioning,combined with optimal trajectory control,significantly reducing costs while ensuring docking precision.First,the camera identifies and locates the target by AprilTag.Second,it minimizes the reprojection error through nonlinear optimization and employs the error state Kalman filter(ESKF)to more accurately eliminate the IMU bias,improving the accuracy of the fusion positioning.Third,an optimal docking controller is designed based on the linear quadratic regulator(LQR).Combining the drone's dynamic model,it directly outputs the attitude control commands for the multirotor drone,achieving safer aerial docking.Finally,experiments are conducted on a quadrotor drone platform and successfully complete the aerial docking task,demonstrating the feasibility of this method.
关键词
多旋翼无人机/视觉定位/状态误差卡尔曼滤波器融合/空中对接/线性二次型控制器Key words
multirotor drones/visual positioning/error state Kalman filter(ESKF)fusion/aerial docking/linear quadratic regulator(LQR)controller引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据