Non-Linear, Large Deformation Modeling of All-Elastomer MEMS Tactile Sensors with Bionic and Robotic Applications

Abstract

The study is motivated by the need to develop highly sensitive tactile sensors for both robotic and bionic applications. The ability to predict the response of an elastomeric layer under severe pressure conditions is a key to the development of highly sensitive capacitive tactile sensors capable of detecting the location and magnitude of applied forces over a broad range of contact severity and layer depression. Thus, in this work, a large deformation Mooney-Rivlin (M-R) material model is employed in developing empirical-analytical models capable of predicting the response of tactile unit-sensors under the application of normal forces under large strain, large deformation conditions. Initially, the non-linear analytical model associated with uniform compression of an infinitely long elastomeric layer of finite thickness is developed (infinite punchinfinite layer assumption). Then, the aforementioned model is combined with an enhanced capacitance model, accounting for fringe field and saturation effects, to predict the response of capacitive tactile unit-sensors under the application of normal forces. Furthermore, to account for the infinite contact length assumption, related calibration factors are introduced. The calibration factors are obtained through model comparisons to the related experimental results. Due to the lack of appropriate mechanical testing the M-R material parameters are estimated using elastic moduli and material incompressibility condition. To further improve the tactile sensor modeling, based on broad parametric studies carried out using method of finite elements (FE), a semi-analytical model for a large, finite, flat indentation of an infinitely long elastomeric layer of finite thickness is developed. By considering finite punch instead of infinite punch (i.e., indentation instead of uniform compression), the new semi-analytical model is capable of predicting the contact forces without any calibration factor. The indentation model is then integrated with the capacitance model to predict the capacitive tactile unit-sensor response under the application of normal forces. Inverse analyses, combining FE simulations and non-linear numerical optimization, is employed to obtain the correct M-R material parameters. Remarkable agreement are found to exist between the experimental measurements and the model predictions.