Tailoring the in-plane and out-of-plane stiffness of soft fingers by endoskeleton topology optimization for stable grasping

The intrinsic compliance of soft materials endows soft robots with great advantages to achieve large deformation and adaptive interactions in grasping tasks. However, current soft grippers usually focus on the in-plane large deformation and load capacity but ignore the effect of out-of-plane external loads, which may lead to instability in practical scenarios. This problem calls for stiffness design along multiple directions to withstand not only in-plane interacting forces with objects, but also unexpected out-of-plane loads. In this paper, we design a new type of soft finger by embedding an endoskeleton inside the widely-used Pneu-Nets actuator, and the endoskeleton layout is optimized to achieve a remarkable bending deflection and limited lateral deflection under combined external in-plane and out-of-plane loads. Based on the multi-objective topology optimization approach, the key structural features of the optimized endoskeleton are extracted and parameterized. The multi-material soft fingers are fabricated by the silicone compound mold method. Static and dynamic experiment results validate that the soft gripper with endoskeleton embedded exhibits remarkably improved out-of-plane stiffness, without sacrificing the in-plane bending flexibility, and leads to more stable grasping.