Calibration-free optical waveguide bending sensor for soft robots

Abstract

Soft robots have become increasingly popular due to their compliance, deformability, and adaptability. Soft sensors, particularly bending sensors, play a crucial role in providing essential posture and position information for these robots. However, current soft bending sensors encounter difficulties in accurately measuring joint bending angles and directions under different curvatures. To address these challenges, we propose a novel dual-colored layer structured (DCLS) bending sensor based on the optical soft waveguide. The DCLS sensor is constructed using polydimethylsiloxane (PDMS) as the clear core, with red and blue layers on each side. The sensor’s performance is evaluated through experiments involving bending, compression, and impact conditions. The DCLS bending sensor exhibits excellent calibration-free properties, allowing it to effectively monitor bending angles and directions of joints of varying sizes without requiring any additional calibration. The sensor is successfully integrated into various soft robots, including a fruit sorting robot, a fish-inspired robot, and a hand orthotic exoskeleton robot, showcasing its versatility and potential for different applications.