SkB-Hand：具複合韌帶骨架結構仿生手之機器手臂物件抓取任務

Abstract

仿生手的發展是機器人研究的關鍵方向，旨在實現類似人手的功能性、靈巧性、適應性。在結構仿生、靈活性與輕量化、抗衝擊性和低成本之間取得平衡，是當前研究挑戰之一。同時，在驅動自由度有限的條件下，如何盡可能提升仿生程度與抓握擬人化程度，也成為能否完成各類抓握類型任務的關鍵因素。 為此，本研究提出了具複合韌帶骨架結構仿生手 —— Skeleton Bionic Hand（SkB-Hand），每根手指整合兩種不同材質、由超彈性材料構成的韌帶，分別用於模擬人體手指伸肌腱（extensor tendon）與掌側板（volar plate），以模擬人體手指的伸展及一定程度的過伸功能與自我保護功能。 SkB-Hand 整體質量小於 600g、手部約120g，成本低於 200 美元, 其提供了一種輕量化且低成本的解決方案。為系統驗證其性能，本研究進行了一系列實驗，分為特性驗證（Characteristics Analysis and Validation）與應用場景（Application Scenarios）兩部分。特性驗證包括：三種回彈機制比較、關節彎曲順序分析、拇指能力評估、衝擊下的柔順性與保護性驗證、及各類抓握與基準測試。測試結果表明，SkB-Hand能在單自由度驅動下再現類人彎曲模式，並具備良好的抗衝擊與拇指活動能力。SkB-Hand在Kapandji test中得分為7/10，在GRASP Taxonomy系統測試中達成滿分（33/33），在基於資料手套的遙感操作下，SkB-Hand能順利執行焊槍操作、螺絲拾取等常見工程作業，同時展現了其優異的負載提拉能力。在應用場景中，SkB-Hand在與TM Robot，視覺系統基於ROS2整合下，完成了抽紙，抓取乒乓球等特定任務，展示了其的實際整合應用之潛力。 本研究實驗結果表明，SkB-Hand 不僅實現了預期的抓握目標，也展現出其在協作型機器人抓取、人機協作及與視覺感知系統集成等實際應用中的廣闊潛力。

The development of bionic hands is a critical focus in robotics, aiming to replicate the functionality, dexterity, and adaptability of human hands. Balancing structural biomimicry, flexibility, lightweight design, impact resistance and low cost remains a key challenge. Meanwhile, how to replicate the natural motion patterns of a human hand under limited degrees of freedom has become key to enabling diverse grasping tasks. This study presents the Skeleton Bionic Hand (SkB-Hand), which features a composite ligament-bone structure. Each finger integrates two hyperelastic ligaments simulating human extensor tendons and volar plates, enabling extension, hyperextension, and self-protection. With a total weight under 600g (hand ~120g) and cost below $200, the SkB-Hand offers a lightweight, low-cost solution. To validate its performance, a series of validation tests—including rebound mechanisms, joint flexion sequencing, thumb dexterity, impact resistance, and benchmark grasp evaluations—confirmed its anthropomorphic motion and robustness. The SkB-Hand scored 7/10 in the Kapandji test and 33/33 in GRASP Taxonomy. It successfully performed soldering and screw handling via data-glove teleoperation, while demonstrating strong lifting capability. Integrated with a TM Robot and RealSense camera in a ROS2 environment, it completed tasks such as tissue paper extraction and ping pong ball grasping. These results indicate that the SkB-Hand not only achieves the expected grasping performance but also shows significant potential for collaborative robotic, human-robot interaction, and vision-integrated tasks.