Development of robot semi-autonomous control method for teleoperation support in nuclear power plants

In this work, a robot semi-autonomous control method has been developed to support the sequence of processes whereby teleoperation robots used in decommissioning work travel to their work area and undertake their designated tasks, the aim being to complete work more quickly by reducing operator workloads. The proposed method combines a traveling control, learning-based object recognition, and manipulator positioning control. The operator only needs to start and stop robot operation and decide whether particular tasks need to be performed. The control method was implemented on the prototype of a teleoperation sysytem using a hydraulic robot and trialed on the sort of radioactive waste removal work that would be required in practice. A comparison of the time taken to perform tasks autonomously and by an operator using manual operation found that use of the proposed method could shorten work times by approximately 25%.