Expanded quasi-static models to predict the performance of robotic skins on soft cylinders

Abstract

Listen

Translate article

Robotic skins with embedded sensors and actuators are designed to wrap around soft, passive objects to control those objects from their surface. Prior state estimation and control models relied on specific actuator and sensor placement in robotic skins wrapped around soft cylinders, as well as used simplistic assumptions based on geometry and an ideal connection between the robotic skin and underlying structure. Such assumptions limit model fidelity and affect its utility in the design and control of surface-actuated systems. In this work, we relax prior assumptions and present a new quasi-static model with mechanics, controls, state estimation, and kinematic sub-models, or modules, for robotic skins placed around cylindrical structures. The kinematics module is used post-process to analyze the performance of the other three modules. We test the utility of the model on two robotic skin designs and compare the performance against a previous model and physical experiments. We demonstrate that the mechanics, controls, and state estimation modules presented herein outperform the previous model and the mechanics module can be used to predict the behavior of new robotic skin designs. The accuracy of the model increases as the stiffness of the host body material increases. This expanded theory could be utilized to reduce fabrication costs and speed up the design process and could be further extended to include system dynamics and model systems with multiple robotic skins.