干扰
执行机构
刚度
计算机科学
工作区
弯曲
抗弯刚度
夹持器
软机器人
触觉技术
结构工程
适应性
模拟
自适应控制
法向力
仿生学
机器人
出处
期刊:Soft robotics
[Mary Ann Liebert, Inc.]
日期:2025-12-24
卷期号:: 21695172251405996-21695172251405996
标识
DOI:10.1177/21695172251405996
摘要
In this work, we introduce the Jamming Adjustable PneuNet Actuator (JAPA), a novel soft robotic finger that enables both stiffness modulation and tunable bending behavior through a flexible hybrid jamming approach. This method combines the high stiffness gain of layer jamming and the adaptability of granular jamming. By adjusting the effective length of the paper-based layer jamming using a magnetically positioned sliding mechanism, JAPA can dynamically reshape its bending profile. Meanwhile, the granular jamming element distributed throughout the finger can provide adaptive stiffness reinforcement across all bending configurations. The combination of adjustable stiffness and reconfigurable bending profile substantially enhances JAPA's multidirectional force control and dexterity. To evaluate its performance, we conducted a series of experiments to assess JAPA's stiffness modulation, pull-off and output forces, multidirectional force control, and workspace. Experimental results demonstrate that JAPA can achieve a maximum stiffness gain of up to 3.55×, with adjustable stiffness distribution contributing to a workspace expansion exceeding 200% and a more than 300% improvement in multidirectional force modulation. To visualize its multidirectional force control ability, we used a single JAPA unit to operate a computer cursor via a TrackPoint, dragging the cursor in different directions. To further validate its manipulation capability, we constructed a four-unit JAPA gripper capable of in-hand object rotation and safe handling of diverse objects, including delicate and irregularly shaped items. The proposed soft finger design holds promise for applications in assistive robotics, adaptive grasping, and human-interactive devices, where both safety and functional versatility are critical.
科研通智能强力驱动
Strongly Powered by AbleSci AI