Abstract:
This study presents a framework for stiffness sensing and adaptive grasping using soft pneumatic grippers (SPGs). The motivation was to imitate the capabilities of human hands. The challenge was the compliance of soft pneumatic actuators (SPAs). The study associated the behaviors of grasped objects with those on SPAs, introduced a new concept of SPA modeling, and proposed a method for stiffness sensing through SPG pincer grasping. Compression testing was conducted for validation. A technique for forecasting deformation and force on grasped objects based on their stiffness was established, and a control law was elaborated. The presented technique and control architecture were validated at different input pressure conditions. The proposed method yielded similar stiffness trends with slight deviations compared to compression testing and demonstrated the potential for manual classification of samples. The presented technique provided small deviations in deformation but significant errors in force. A major limitation was caused be the using of computer vision for inspecting SPA deformation. To overcome this limitation further studies on the direct sensing of SPA deformation is recommended Overall, this study contributes to the field of soft robotics and object classification by integrating stiffness sensing and SPG grasping in a single action, leading to the next level of adaptive SPG grasping.
