基于灵巧手触觉信息的未知物体类人探索策略

An anthropomorphic exploration strategy of unknown object based on haptic information of dexterous robot hand

为增强机器人对环境的适应性,基于对人类进行未知物体触觉探索时行为的观察和分析,提出一种适用于机器人灵巧手自主进行未知物体触觉探索的策略。机器人触觉探索过程被划分为顶面探索和侧面探索两个阶段。在顶面探索阶段,根据所得到的触觉信息对物体的基本尺寸和姿态进行估计,并用包围盒近似物体;而在侧面探索阶段,依据分类判别不等式将物体按基本尺寸进行分类,对不同类别的物体设计不同的侧面探索策略,从而指导灵巧手对未知物体的信息进行采集。最后,在仿人机器人平台上完成了对未知物体的触觉探索实验。实验结果表明,应用所提出探索策略,机器人可以仅依靠触觉自主的探索未知物体并获取物体信息。

Based on the observation and analysis of human behaviors in exploring an unknown object by touch,a haptic exploration strategy was proposed to enhance robot adaptability to an environment which is suitable for a dexterous robot hand to explore the unknown object autonomously. The whole exploration process was divided into two stages which were the top exploration and the side exploration. In top exploration,the object was approximated by a bounding box and the basic dimensions and posture of object were roughly estimated according to obtained haptic information. In side exploration,the objects were classified according to basic dimensions by the classification inequalities ,and a different side exploration strategy was applied to each kind of objects to guide the robot to collect haptile iniormation of unknown objects. Finally,the haptic exploration experiment of the unknown object was completed by a humanoid robot platf orm. Experimental results show that the robot can use only tactile sensors to complete exploration of the unknown object autonomously through the proposed strategy.