自主空中加油多摄像机近距相对导航与控制  

Relative Navigation and Control of Multi-Camera Systems in Close Proximity for Autonomous Aerial Refueling

作　　者：郭军[1,2] 董新民[1] 王龙[1] 成万舟 

机构地区：[1]空军工程大学工程学院 [2]93716部队57分队 [3]93735部队司令部

出　　处：《信息与控制》2012年第1期95-101,共7页Information and Control

基　　金：航空科学基金资助项目(2008ZC01006)

摘　　要：针对伸缩套管方式下自主空中加油(AAR)对接与位置保持过程,提出了一个基于多摄像机系统(MCS)的相对导航与控制方案.该方案中的位姿估计算法通过结合3D-3D位姿估计技术和N点透视技术,能够充分利用多摄像机系统的宽视野和冗余测量信息.使用卡尔曼滤波算法对视觉估计数据进行滤波,并与速度测量数据进行融合,以提高视觉导航算法的精度和鲁棒性.为抑制外部风扰动,设计了一个基于动态对策理论的最优相对轨迹跟踪控制律.仿真结果表明提出的方案能够满足空中加油要求.A relative navigation and control scheme based on multi-camera systems （MCSs） is presented for the processes of docking and position holding in autonomous air refueling （AAR） mission using boom and receptacle system. By integrating 3D-3D pose estimation technique and perspective-N-point technique, the wide field of view and the redundant measured information of MCSs are fully utilized by the scheme. The data estimated by vision system are filtered by Kalman filtering algorithm and fused with velocity measurements to increase the accuracy and the robustness of vision navigation algorithms. To reduce the effect of wind turbulence, an optimal relative trajectory tracking control law is designed based on the theory of dynamic game. The results of simulation show that the proposed scheme can meet the requirements of aerial refueling.

关 键 词：多摄像机系统 位姿估计 自主空中加油 卡尔曼滤波 动态对策 

分 类 号：V249.1[航空宇航科学与技术—飞行器设计]