High-performance combination of low resolution tactile images using a bit-based representation

Abstract

Since the seminal work of Harmon[1] through to more recent reviews such as that of Dahiya et al.[2], the focus in the field of tactile sensing has been on creating sensors with resolutions as low as 1mm 2. Such specifications are justified by the simplicity of manufacturing a single resolution of skin sensor for all possible applications. Unfortunately, this approach requires not only that a robot has some means of connecting to potentially millions of sensors, but more significantly that it has some means of processing the commensurately large volume of data. The work presented here reduces the volume of data that must be processed by allowing the requirement for high resolution sensors to be relaxed and by the novel use of the bitboard data structure that has been used to represent board games for computers as well as solve complex graph problems [3]. © 2012 Springer-Verlag.