Design of a soft tactile sensor with a broad tactile bandwidth for improvement of robotic grasping ability

Abstract

Deploying robots in the human workspace requires that the robots can physically interact with the environment in a safe and dexterous manner. In humans, the tactile sensing through the skin plays an important role in achieving a safe and dexterous physical interaction with the environment. The two important characteristics of the human skin that allow humans such safe and dexterous physical interaction are its inherent softness and its ability to transduce mechanical contacts of different frequencies into tactile signals through mechanoreceptors. Due to the importance of inherent compliance, soft and elastic tactile sensors have extensively been researched, using soft transducers of different mechanisms, such as piezoresistive, capacitive and piezoelectric. In terms of the detectable bandwidth, the similarities between different transducers and mechanoreceptors have been pointed out in previous researches. However, there are limited research on combining two or more different types of soft transducers into a single tactile sensor to build a soft tactile sensor with a broad tactile bandwidth, similar to that of the human skin. In this work, a soft tactile sensor that is comprised of two components is proposed. The two components transduce mechanical stimuli of different frequency bands, each providing the utilizing system with tactile information of different tactile bandwidth, similar to mechanoreceptors in human skin. This allows for separate encoding of the tactile inputs based on their frequencies and furthermore, allows the robotic system to identify contact events that discretize action phases in an object manipulation task, as is done in human object manipulation control schemes. The former part of this work, the mechanism, structure and analysis of the two components, low and high frequency transducer, of the proposed soft tactile sensor is introduced. Capacitive and piezoelectric mechanisms are utilized for the transduction of the mechanical stimuli of low and high frequencies, respectively, and both are fabricated using soft and flexible materials to provide compliance to the tactile sensor. In the latter part of this paper, the proposed sensor will be integrated with an existing commercial robotic hand. Object grasping tasks will be performed with the integrated robotic hand system using tactile signals to specify the start and end of each action phase to achieve a grasping control similar to that of humans.