Artificial Tactile Sensing of Position and Slip Speed by Exploiting Geometrical Features

Abstract

Rich information about artificial grasps can be obtained using tactile sensing arrays. However, the complexity of the integration and computation inherent to tactile sensor arrays limits their applicability for prosthetic manipulators. We present an artificial ridged skin that detects the position and speed of a slipping object using a single force sensor. The artificial skin features parallel ridges arranged in a nonuniform configuration. An evolutionary algorithm generates distributions of ridges evaluated by the accuracy and resolution of detecting the position and speed of a slipping object. Slip experiments on real skins featuring ridge distributions generated by the evolutionary algorithm show that ridge arrangement is critical for an improved sensing of object position and slip speed. We report ridge patterns that detect object position and speed with errors as low as 10% at slip speeds of up to 60 mm/s. The artificial skin has an average speed sensing resolution of 10 mm/s, an average position sensing resolution of 15 mm, and is robust to various grip conditions, e.g., speed variation, object weight, and contact area. By exploiting the geometrical features of the artificial skin, enhanced tactile information is acquired. The concept opens a promising avenue for robust and energy-efficient tactile sensing systems.