移动机器人自适应视觉伺服镇定控制

Adaptive visual servo regulation of mobile robots

中文关键词  单目视觉  移动机器人  自适应镇定控制  未知外参数

英文关键词  monocular vision  mobile robot  adaptive regulation control  unknown extrinsic parameters

中文摘要

      对有单目视觉的移动机器人系统, 提出了一种自适应视觉伺服镇定控制算法; 在缺乏深度信息传感器并且摄像机外参数未知的情况下, 该算法利用视觉反馈实现了移动机器人位置和姿态的渐近稳定. 由于机器人坐标系与摄像机坐标系之间的平移外参数(手眼参数)是未知的, 本文利用静态特征点的位姿变化特性, 建立移动机器人在摄像机坐标系下的运动学模型. 然后, 利用单应矩阵分解的方法得到了可测的角度误差信号, 并结合2维图像误差 信号, 通过一组坐标变换, 得到了系统的开环误差方程. 在此基础之上, 基于Lyapunov稳定性理论设计了一种自适应镇定控制算法. 理论分析、仿真与实验结果均证明了本文所设计的单目视觉控制器在摄像机外参数未知的情况下,可以使移动机器人渐近稳定到期望的位姿.

英文摘要

      For a monocular camera-based mobile robot system, we propose an adaptive visual servo regulation controller. This controller asymptotically drives the robot to its desired position/orientation, even when the range sensors are not available and the camera extrinsic parameters are unknown. Because the translational parameters between the robot frame and the camera frame are unknown, the kinematical model of a monocular camera-based mobile robot system is developed based on a static point in the camera coordinate system. The orientation error is extracted from the image by decomposing the homography matrix, which is then incorporated with the image error signals to form the open-loop error system by a coordinate transformation. On this basis, an adaptive regulation controller is proposed by using Lyapunov techniques. The performance of the controller is validated by both the theoretical stability analysis and the experiment results, showing that the controller can perform a rapid asymptotic regulation even when the extrinsic camera parameters are unknown.