Debris mitigation will require robust robotic grasping. Currently, most approaches to tactility in robotic grasping rely on either significant processing resources or soft robotics. There is little doubt that robust and adaptable robotic grasping would be a boon in both orbital debris removal and on-orbit servicing. A three-layered approach is proposed to grasping: (i) hybrid force/impedance control through feedback is the traditional approach but delays in the feedback cycle can generate instabilities; (ii) the addition of a feedforward predictive capability partially circumvents this problem by emulating the function of the human cerebellum (essentially, a Kalman filter-based neural network); (iii) the implementation of a software-based viscoelastic response to the electric motors emulates muscle/tendon behaviour to provide robustness to grappling errors. All three approaches operate in parallel to generate robust and adaptive manipulation of forces.
