Task Realization by a Force-Based Variable Compliance Controller for Flexible Motion Control Systems

Abstract

This paper proposes a task realization method by using a force-based variable compliance controller for flexible motion control systems. In recent years, the robots working in human life space are desirable. Considering the robots working in human life space, they should achieve the safety motion. From this viewpoint, one of the key technologies is flexible motion control system. Of course, task realization instead of human beings is important. Therefore, it is necessary to propose the realization method of several tasks for the flexible motion control systems. In this paper, two robot tasks are defined: “approach task” and “pushing task.” The approach task is the motion for a robot to approach an environment and not to contact the environment. On the other hand, the pushing task is the motion for the robot to contact and push the environment in order to achieve the several tasks. For the realization of the several tasks to the environment, it is necessary to achieve both the position control during the approach task and the force control during the pushing task. Therefore, the controller has to be modified depending on the contact state, which means whether the robot is in contact with the environment or not. In order to modify the controller, the variable compliance gain which is varied according to the contact state is proposed. Focusing on the approach task, the position control which is equivalent to the conventional position-based compliance method is achieved by using the proposed method. On the other hand, the proposed controller is suitable for the pushing task compared with the conventional position-based compliance controller since the proposed controller is based on the force control. Therefore, several tasks which include the position tracking and the contact with the environment are actualized by using the proposed method. In addition, the performance analysis by a Bode diagram and stability analysis by root loci are conducted. The validity of the proposed method is confirmed from the experimental results.