Robotic manipulators used in space applications are often lightweight, exhibit significant joint flexibility, and have limited joint torques. Many applications of space manipulation involve the transport of massive objects, e.g., a captured satellite, well beyond the 1-g capability of these robots. This may cause challenges during manipulation, such as imprecise trajectory following during transport, oscillatory behavior, and difficulty berthing the object. Motivated by the space application, a fixed-base scenario is considered and a methodology is proposed for massive object manipulation via flexible joint manipulators in this paper. The manipulation task includes the following phases: tracking and capture, path and trajectory planning, trajectory execution, and berthing. Visual servoing and compliance control are used to track and capture the object. Motion planning accounts for kinematic and dynamic constraints on joint stops, path smoothness, actuator limits, end effector slippage, and braking distance. Berthing is achieved through the use of compliance control. The efficacy of this methodology is experimentally investigated with a flexible joint Rethink Baxter robot arm and each manipulation phase is completed with a heavy load beyond the 1-g limit of the arm on an air bearing table.
AIAA Journal on Guidance, Dynamics, and Control, published onlin, January 2021.