A Tunable Magnet-based Tactile Sensor Framework

Abstract

Tactile sensing enables controllable interactions as robots enter unknown and unstructured environments. However, existing tactile sensors suffer from limited form factors and durability, lack of dynamic range, and are not cost-effective. Magnet-elastomer-based tactile sensors offer a potential solution to compensate for these deficiencies. In this paper, we present a generalizable design approach, a first-order model based on the magnetic field and elastomer fundamentals, and an automated calibration routine to create custom tactile sensors. To demonstrate the performance and versatility of the proposed sensor architecture, we selected two sample design configurations: a small form factor that was optimized for tumor localization (Pinky) and a larger form factor for robust contact estimation on legged robots (Foot). These two unique cases demonstrate the breadth of sensors that can be designed using this approach as well as how changes in sensor parameters can be used to tune the range and resolution for different applications.