Real-time cooperative control of a dual-arm redundant manipulator system

The position, force, impedance control problem for single non-redundant and redundant robot manipulators has been addressed extensively in the last two decades. Robot manipulators have been widely used in industrial manufacturing, under-water exploration, hazardous environment operations and space application. While a single robot can handle many tasks today in place of human beings, more sophisticated tasks can only be performed safely and efficiently by multiple robots. Multiple robots performing tasks together in a cooperative manner can have a significant advantage over a single robot just as a human being using two arms has an advantage over one using a single arm and multiple humans have an advantage over a single human. This becomes obvious in tasks such as handling massive payloads or non-rigid payloads—tasks such as those required of space manipulators like the Special Purpose Dexterous Manipulator (SPDM) on the International Space Station. All these applications normally involve the robot system coming in contact with its environment. In such scenarios, the contact force has to be regulated in order to prevent physical damage to the robot and/or the environment. An impedance behavior has to be controlled between the robot and the environment. When multiple robots are involved in handling a common object, the internal forces which do not contribute to the motion of the object but can cause damage to the objects or the robots during manipulation must be regulated. Most conventional single-robot control approaches and recent master-slave cooperative multiple manipulator control schemes will not be able to perform safely or efficiently the tasks mentioned above.; In this thesis, an impedance control based cooperative dual-arm controller is developed. This allows the jointly held object to be controlled such that a desired impedance is exhibited between the robot and the environment. The internal (squeezing) loading of the object caused by both robots is regulated to have another impedance behavior between the internal force tracking error and the object tracking trajectory. By doing this, we can balance the object tracking accuracy and the internal force tracking accuracy depending on the calibration status of the robot. Also, we gain high robustness to the system's kinematic and dynamic uncertainties. The proposed controller also solves the “transition” problem that occurs during the transition from independent open-chain control to closed-chain dual-arm control. The control scheme also takes advantage of the redundancy in the robots to achieve singularity avoidance in closed-chain operation and collision avoidance in open-chain operation.; The dual-arm control scheme has been designed, simulated, and experimentally verified. The scheme has been implemented in the Robotics Laboratory of the Department of Electrical and Computer Engineering at University of Western Ontario using a real-time operating system and provides real-time performance. Extensive experiments involving a number of strawman tasks have been performed using two redundant manipulators to demonstrate the capability of the dual-arm system in open-chain and closed-chain operation, and the robustness of the system to kinematic and dynamic uncertainties.