Modeling and understanding locomotion of pneumatic soft robots

Abstract

Imitating natural locomotion of biological systems (soft-bodied animals) opens the door to the development of a new class of machine, referring to soft robots. A variety of soft robots have been demonstrated by researchers and engineers through incorporating soft technologies into their designs. Yet computer modeling of locomotion of soft robots remains to be a challenging task, not merely because their intrinsic deformation is continuous, complex, and highly nonlinear compared to conventional rigid-bodied robots, but moreover because of the complicated contact problems encountered during locomotion of soft robotics. Herein, we present a combined analytical and numerical analysis of the locomotion of pneumatic network-based soft robots. Concerning a quadruped robot, two fundamental different gaits (undulation and crawling) were identified and numerically validated by two driving modes of pneumatic robots. Extracting ground reaction forces and centroid trajectory from the simulation throws a light on the underlying mechanism of locomotion of soft robots. Our efforts would enhance the understanding and facilitate the control, manipulation, and trajectory optimization of bio-inspired soft robots.