Interactive Method for Robot Control and Its Application to Deburring

摘要

An interactive robot control method is developed in this dissertation. The211u001econtrol method is studied in two main domains, human-robot interactive control 211u001eand robot-environment interactive control. A human passive and robot active 211u001einteractive control scheme is proposed. A new robot task description method and a 211u001esystem structure are presented to achieve the control scheme. Uncertain 211u001einformation within the robot working environment is considered in the task 211u001edescription. Human-robot interactive control and robot-environment interactive 211u001econtrol are integrated hierarchically by the structure. A new conceptual human-211u001erobot interaction model is proposed. Task allocation between human and robot is 211u001epresented based on the model. Task oriented language is developed using the task 211u001edescription method. Robot level control and task level control are employed to 211u001eachieve the robot-environment interactive control. A novel fuzzy-PI control 211u001ealgorithm is proposed to implement the robot level control, while a new self-211u001etuning algorithm is proposed to implement the robot level control, while a new 211u001eself-tuning algorithm is used in the algorithm. The task level control is based 211u001eon a task knowledge, thus a task level control models is proposed. Human skills 211u001eare acquired and used by an adaptive network based fuzzy, inference system 211u001e(ANFIS) method. The proposed robot interactive control methods are demonstrated 211u001eby a deburring application, while the human-robot interaction is demonstrated by 211u001ethe deburring job plan. The task level control method is utilized to achieve 211u001erobot-environment interaction. Experimental results show that the method is 211u001eapplicable.