Model Identification of a Soft Robotic Eye Actuator for Safe Social Interactions

Abstract

This paper explores the model identification of a novel tendon-driven soft continuum actuator, intended as a functional joint for the social robot HARU. The actuator’s design is customized for integration into HARU’s eye joints, emphasizing safety in interactions with children, in accordance with UNICEF’s “Policy Guidance on AI for Children”. The performed experimental study assesses and compares the accuracy of various auto-regressive with exogenous inputs (ARX) modeling techniques—linear, nonlinear, and recursive—through motion data from dynamic experimental tests of the actuator under different orientations. The results provide insights into the efficiency of these modeling strategies in dynamic conditions with continuum actuators, thereby offering a basis for model selection in the integration of soft actuators into robotic systems for practical applications.