Enhancing interaction performance of soft pneumatic-networks grippers by skeleton topology optimization

The inherent compliance of soft materials imbues robots, generally referred to as soft robots, with particular advantages in producing adaptive and safe interactions. However, the mainstream design paradigms of soft robots have been focused on pursuing large free motions only, usually at the expense of greatly decreased stiffness, leading to limited capability of withstanding external loads in interactive scenarios. There is a pressing need to incorporate the interaction specifications at the design stage to embody soft robots with not only proper deformability but equally importantly, considerable stiffness to perform complex tasks in practical applications. Here, inspired by the dexterity of human hands, we propose a computational design framework for soft grippers with a focus on improving their interaction performance in power grasping or precision grasping mode. The design paradigm rests on attaching a relatively stiffer skeleton layer to the parametric pneumatic networks based actuator which is widely used due to the geometric advantage, and the skeleton layout is designed for customized interaction conditions by a level set based topology optimization approach. As expected, the optimized skeleton layouts exhibit specified structural features highly relevant to the predefined concentrated loads for precision grip or distributed loads for power grip, which physically implies the compromise between deformability and stiffness. Since the interaction forces are difficult to measure in situ, we devise power and precision grasping scenarios and evaluate the critical actuation pressure of the object's falling instead. The experiments qualitatively demonstrate the superiority of each specified design. This work represents an initial step toward the rational design for interaction in soft robots.

Automated summary

Soft robots have advantages in adaptive and safe interactions, but also require proper stiffness to withstand external loads. This work proposes a computational design framework for soft grippers inspired by human hand dexterity, optimizing skeleton layouts for improved interaction performance in different scenarios.