论文部分内容阅读
针对家庭环境中服务机器人的物品抓取问题,以NAO机器人为实验平台,提出了一种结合迭代学习控制的基于位置的视觉伺服抓取算法.结合Bumblebee2双目立体相机构建了视觉伺服系统,对NAO机器人的机械臂进行了运动学建模.针对静止物品和传递过程中运动物品的抓取问题,分别设计了李雅普诺夫渐进稳定的基于位置的视觉伺服(PBVS)抓取控制律,最后为克服机械误差和物品运动预测误差的影响,结合迭代学习控制对PBVS控制律进行了改进,通过迭代实验验证提出的方法能够快速校正系统误差,提高系统响应速度,使机器人能够快速、稳定地完成目标物品的抓取,实现相应的家庭服务.
In order to solve the problem of service robots in home environment, NAO robot is used as an experimental platform to propose a position-based vision servo capture algorithm combined with iterative learning control. A visual servo system is constructed based on Bumblebee2 binocular stereo camera. NAO robotic arm was kinematically modeled.Aiming at the problem of still objects and the crawling of moving objects in the process of transfer, Lyapunov’s gradual and stable position-based vision servo (PBVS) grasping control law was designed, and finally To overcome the influence of mechanical error and article motion prediction error, the control law of iterative learning control is improved to PBVS. The proposed method through iterative experimental verification can quickly correct system errors and improve the system response speed so that the robot can accomplish the target quickly and stably Catch items, to achieve the appropriate family services.