Geometric Modeling and Estimation of Robotic Fin Shape with Bending Actuators and Passive Elements

One possible design approach for bio-inspired robots is to employ structures that facilitate their shape estimation to enable control that mimics the animal's locomotion. We propose a discrete surface model for surface-shaped body parts of bio-inspired robots and a method for computing its shape. The surface model is constructed by alternately connecting bending actuators and passive elements. This study focused on a robotic stingray, and its fins were fabricated from cylindrical bending actuators with living hinges and fabric. The proposed computation takes about 10 ms, short enough to allow real-time computation. The shapes were estimated from the actuator's control value obtained in experiments, and the results were compared with the measurement results using an optical motion capture system. The differences between the estimated and measured point positions were within 15 mm, suggesting that a more accurate estimation is possible with robot system improvements. If actuators for a bio-inspired robot are surface-shaped, as in soft or bio-hybrid actuators, this fin model is expected to be a helpful reference for robotic design.