RESEARCH

The number of actuators is directly related to the functionality of a robotic glove. With more actuators, the robotic glove can move a greater number of fingers and realize a wider variety of motions, thereby enabling more tasks. However, increasing the number of actuators also adds weight and reduces compactness, making the system less portable. To address this trade-off, I devised a slack-based tendon-driven mechanism that allows multiple motions to be executed in an intended sequence with only a single actuator. Using this approach, I developed a glove capable of grasping objects in various postures with just one actuator.

When developing the robot with soft materials, a key consideration is that, unlike rigid materials, the body itself can deform. For example, in the silicone-based Exo-Glove Poly II, friction-induced asymmetry of tendon tension in the finger body often led to distortion of the finger body structure, preventing the glove from performing as designed. To address this issue, I applied topology optimization and size optimization methods to design the finger body in a way that minimizes distortion (functionality) while also tailoring the tensile stiffness to match user preference (usability).

The control method of the robotic glove is directly tied to its usability. No matter how many motions the glove can perform, if control is difficult or complex, it cannot be effectively applied in daily living. To address this issue, I leveraged the Exo-Glove Poly III, which employs slack-based passive sequencing to realize two contrasting grasping strategies with a single actuator. Based on this functionality, I utilized vision data and deep learning algorithm to interpret contextual situations. The algorithm decides which grasping strategy should be used and sends this information to the actuation system. This approach makes the robot’s functionality more practical and applicable to real-world daily tasks.