Actuation Reading Insights

Abstract

Unlike traditional robots, soft robots can navigate narrow and complex environments while interacting safely and compliantly with their surroundings. The potential of these abilities motivates the development of algorithms that can accurately assess environmental interactions, making soft robots more autonomous. There is growing interest in estimating external loads in soft robots by matching precise shape data with mechanics models. This approach aims to enhance the robots’ ability to adapt to unpredictable forces and environments, improving real-time decision-making and responses. In this article, we propose a novel approach for simultaneous shape and force estimation of tendon-driven slender soft robot-based solely on tendon displacements and tensions. Our approach introduces a kinetostatic model for slender soft robots based on the geometric variable-strain method, builds upon the Cosserat rod theory. The length of the threadlike actuators (tendons) is defined as a function of the deformation of the soft robot. By solving the inverse problem using this kinetostatic model and the length and tension of the tendons, we can estimate both the external force and the shape of the soft robot simultaneously. We validate our approach with an experimental prototype, achieving accurate force and shape estimation results. This method efficiently integrates actuation and sensing in a small and integrated form, which can be highly beneficial for applications requiring compact designs.