【摘要】 针对机器人携带充电枪进行充电任务中的寻孔过程,研究一种基于接触位置状态识别的机器人寻孔策略。将机器人寻孔时充电枪头与充电插座接触位置分为平面、倒角和内壁接触状态,分析不同接触位置状态的接触力和力矩特征,通过支持向量机(support vector machine,SVM)算法与投票法,提出了一种寻孔接触位置状态识别方法。针对平面接触状态、倒角接触状态和内壁接触状态,分别提出相应的寻孔轨迹规划方法,并采用力/位混合控制将寻孔轨迹规划与力跟踪导纳控制相结合。最后基于机器人充电实验平台,通过寻孔实验证明所提算法的有效性。更多还原

【Abstract】 Aiming at the hole finding process in the charging task of robot carrying charging gun, a robot hole finding strategy based on contact position state recognition is studied. The contact position between the charging gun head and the charging socket during robot hole finding is divided into plane, chamfer and inner wall contact state, the contact force and torque characteristics of different contact position states are analyzed, and a hole finding contact position state recognition method is proposed through support vector machine(SVM) algorithm and voting method.According to the plane contact state, chamfer contact state and inner wall contact state, the corresponding hole tracing trajectory planning methods are proposed, and the force/position hybrid control is used to combine the hole tracing trajectory planning with force tracking admittance control. Finally, based on the robot charging experimental platform, the effectiveness of the proposed algorithm is proved by hole finding experiments.更多还原