Master-Assisted Cooperative Tracking Control of Human and Robot. Model Reference Cooperative Method Using a Neural Network.

A cooperative control approach between human and robot is important for carrying out various tasks in a hazardous environment, including space. In this case, the robot is operated based on a cooperation of direct human control and autonomous robot control. In this work, a process of cooperation between human control and autonomous robot control using a neural network for optimization of the degree of cooperation between human and robot is proposed. The proposed control system complements the judgment ability of the human operator to fuse the recognition ability of the human operator and the sensing functions of the robot. The degree of participation of the human operator in the control is determined based on a reference cooperative model which expresses the desired human and robot cooperative form. In experiments, contacting tasks for various object walls were performed by a two degrees of freedom Cartesian robot. The results indicate that the proposed cooperative method can be used for successful cooperation of human and robot.