Discriminative touch from pressure sensors

Abstract

Touch is an important sensory pathway for exploring the world, but most robotic systems either have no sense of touch, use simple binary bump switches, or require expensive custom sensors. In this work we investigate the use of low-cost sensors to acquire more discriminative representations of touch sensations. We show that using two pressure sensors in a 3Dprinted housing we can determine the location of a touch along a one dimensional axis. Furthermore, we can distinguish between different types of touches by the profile of the sensor response.