This week, Dr. Chris Macnab from the University of Calgary is visiting our department in MPI. She gave a talk yesterday in the theme of “Control Systems for a Surgical Robot on the Space Station”. I showed my electro-adhesion demonstration to him, and we had a fruitful brainstorming chat for the underlying physics behind this phenomena.
Abstract: As part of a proposed design for a surgical robot on the space station, my research group has been asked to look at controls that can provide literally surgical precision. Due to excessive time delay, we envision a system with a local model being controlled by a surgeon while the remote system on the space station follows along in a safe manner. Two of the major design considerations that come into play for the low-level feedback loops on the remote side are 1) the harmonic drives in a robot will cause excessive vibrations in a micro-gravity environment unless active damping strategies are employed and 2) when interacting with a human tissue environment the robot must apply smooth control signals that result in precise positions and forces. Thus, we envision intelligent strategies that utilize nonlinear, adaptive, neural network, and/or fuzzy control theory as the most suitable. However, space agencies, or their engineering sub-contractors, typically provide gain and phase margin characteristics as requirements to the engineers involved in a control system design, which are normally associated with PID or other traditional linear control schemes. We are currently endeavoring to create intelligent controls that have guaranteed gain and phase margins using the Cerebellar Model Articulation Controller.