Touch Sensing for Space Robotics

We describe the design of a spaceflight-traceable tactile skin suitable for use with robotic manipulators and the algorithms necessary to make use of it. In order to perform robustly in the presence of orbital thermal and radiation extremes, the skin is based on relatively large discrete force sensors arranged sparsely on a rigid substrate. In order to compensate for the resulting lack of spatial resolution, we have developed a signal reconstruction algorithm based on Basis Pursuit Denoising (BPDN), a sparse reconstruction technique which is shown to provide an effective resolution approximately twenty-five times greater than the Nyquist-Shannon Sampling Theorem suggests is possible for a skin of this type. The signal processing algorithm is robust to white noise and simultaneously performs environmental contact classification.