Tendon Driven Assistive Hand for Mirror Based Robotic Therapy

Abstract

Soft robotic hands, particularly those utilizing tendon-driven mechanisms, offer significant advantages in precision and adaptability for handling delicate objects. These soft robotic hands are designed to mimic the dexterity and flexibility of the human hand by employing variable stiffness materials. Despite recent advancements in tendon-driven soft robotic hands, there remains a gap in the development of simplified variable stiffness mechanisms with integrated sensors for real-time adaptability. This paper presents the development of a modular, bio-inspired, tendon-driven soft robotic hand, fabricated using 3D printing. Flexible material NinjaFlex TPU85A is used for flexible joint mechanism whereas rigid PLA is used for fingers and rest of the hand for structural support. The combination of these materials allows the soft robotic hands to dynamically adjust their stiffness, striking a balance between flexibility and strength. Following the muscle contraction model, the hand is actuated by servo motors while optimized tendon paths reduce friction for smooth motion. Additionally, force and flex sensors provide real-time feedback, allowing precise control of finger movements and grip pressures. Experimental tests conducted using both glove hand simulation and robotic prototypes confirmed the system's ability to replicate glove finger movements accurately. The presented design holds potential for applications in prosthetics, industrial automation, and delicate object manipulation, where precise, flexible handling is essential.