论文部分内容阅读
针对家庭环境中服务机器人物品的抓取问题,提出一种改进的基于位置的视觉伺服抓取算法。首先,利用Naomark标签完成对物体的快速识别,并通过世界平面单应矩阵分解对物体的位姿进行估计;然后,对NAO机器人的机械臂进行运动学建模,并分别设计单臂和双臂抓取的视觉伺服控制律;最后,为进一步提高抓取的稳定性和鲁棒性,对末端执行器进行路径规划。实验结果表明,本方法能够快速、稳定地抓取目标物品。
In order to solve the problem of service robots in home environment, this paper proposes an improved location-based visual servo capture algorithm. Firstly, the Naomark label is used to quickly recognize the object and the pose of the object is estimated by the plane plane homography decomposition of the world. Then, the kinematics of the NAO robot manipulator is modeled, and one arm and two arms are designed separately Finally, in order to further improve the stability and robustness of crawling, the end-effector path planning. Experimental results show that this method can quickly and stably grab the target object.