视觉辅助的无人机自主空中加油建模与仿真  

Visual-aided modeling and simulation of unmanned aerial vehicle autonomous aerial refueling

作　　者：王旭峰[1] 董新民[1] 孔星炜[1] 王龙[2] 程建锋[1] 

机构地区：[1]空军工程大学航空航天工程学院 [2]中国人民解放军94638部队

出　　处：《飞行力学》2013年第4期331-335,共5页Flight Dynamics

摘　　要：为准确获得插头锥套式空中加油过程中无人机与锥套的相对位姿信息,提出了机器视觉辅助的插头锥套式无人机自主空中加油仿真方案,开发了基于Open CV,Simulink及Vega Prime视景仿真技术的闭环仿真系统。研究了机器视觉识别跟踪锥套的图像处理算法,利用Kalman滤波算法估计无人机与锥套的相对位姿。仿真结果表明,机器视觉算法可以准确识别跟踪锥套,滤波器状态估计的收敛速度较快,无人机姿态角变化范围较小,保证了插头锥套式无人机自主空中加油的平稳安全对接。In order to precisely obtain the relative pose of unmanned aerial vehicle（UAV） during aerial refueling docking,a visual-aided simulation scheme of probe-drogue UAV autonomous aerial refueling（PD-UAV-AAR） is proposed.A detailed closed loop simulation system based on Open CV,Simulink and Vega Prime has been presented.Analyze the tanker vortex and refueling drogue moving characteristics.Investigate the refueling drogue recognition and tracking algorithm of machine vision.By using of Kalman filter algorithm,the relative pose of UAV and refueling drogue is estimated.The experimental result shows that the machine vision algorithm can recognize and track refueling drogue precisely and the Kalman filter algorithm can estimate the relative pose of UAV and refueling drogue well which guarantees the stable and secure docking of PD-UAV-AAR.

关 键 词：自主空中加油 机器视觉 锥套识别 相对位姿 虚拟仿真 

分 类 号：V249.1[航空宇航科学与技术—飞行器设计]