A Hybrid Control Strategy for Dual-arm Object Manipulation Using Fused Force/Position Errors and Iterative Learning

Abstract

We report on the development of a force/position hybrid control strategy for collaborative object-handling tasks in a dual-arm system composed of two position-controlled commercial manipulators. A Kalman filter (KF) is used to fuse the measured position and force errors to generate the fused force error as the input to a force PID controller that generates the associated position command that is accepted by the embedded controller of the commercial manipulators. A spring -inerter-damper model is also embedded within the KF, simulating the interface dynamics and providing the smooth force interaction between the manipulator hand and the grasped object. In addition, the iterative learning control that uses the fused position error serves as the additive position-based compensation of the manipulator. To verify the proposed strategy, the system is validated experimentally. The experimental results show that the proposed control strategy can properly alter the balance between the force and position errors. The results also show that adding the KF can prevent severe trembling of the robot's arms.