Mechanically compliant soft robots are highly efficient in terms of human-machine interaction. To further excel the interactive power, high performance tactile and pressure sensors have to be embedded into the thin layer beneath the surface or mounted directly on surface. Artificial skin is comprised of individual sensor arrays with high stretchability, sensitivity and long-term operability. The sensor prototypes derived from commercial rubbers are ultrastretchable and multiple stimuli responsive (strain, pressure, touch) with high spatiotemporal resolution. Highly sensitive hybrid sensors can be ideal for localized sensing; however, they may be inaccurate in dynamic sensing environment, i.e., in soft robotic segments. As robotic control is not as precisive as human motion, sensors optimized for physiological motion detection may not be not able to process shear amount of information from external stimuli. An e-skin array sensor can process large amount of information in a dynamic sensing mode. In addition, a stretchable, micropatterned e-skin can cover larger surface of a soft robot and can record complex texture information by producing varied electrical voltage signal as a function of vibration, touch or force. Transfer-printed and 3d-printed sensors are highly reproducible in terms of performance, compared to composite nanostructures prepared by conventional chemical fabrication.
The main goal of this project is to design functional elastomeric substrates by incorporating additive manufacturing methods, self-healable properties and transfer-printing techniques for highly sensitive triboelectric and piezoresistive sensors. Furthermore, a triboelectric tactile module can also harvest mechanical energy that could function as a nanogenerator. Our objective will be to design hybrid piezoresistive-triboelectric sensor arrays that are multifunctional and therefore, mimic diverse receptors atypical to human skin. An array of hybrid sensors should be able to map the touch, pressure and strain on robot skin more effectively. However, the sensors made from commercial rubbers are incompatible in terms of adhesion on silicone-based robotic components. The objective of this work will also be to investigate the surface and adhesion chemistry of the e-skin arrays developed using commercial rubbers. As e-skin based sensors have progressed enormously in the last few years, our goal will to be to at least or, nearly mimic the performance of robot-integrated sensor arrays with that of biological systems. A collaborative development of machine learning algorithms for hybrid sensor-based e-skins will also be presented during the project tenure.
