Shape Control of a Dual-Segment Soft Robot using Depth Vision

Pneumatic soft robots outperform rigid in complex environments due to the high flexibility of their redundant configurations, and shape control is considered a prerequisite for applications unstructured environments. In this paper, we propose depth vision-based method two-segment robot, which uses binocular camera achieve 3D robot. A closed-loop algorithm based on vision designed compensation when subject its own non-linear responsiveness coupling by solving feature parameters used describe robot analytically modeling motion curved points. Experimental results show that position angle errors are less than 2 mm 1° respectively, curvature error 0.0001mm-1, has convergence performance L-type S-type reference shapes. This work provides general being able adjust without on-board sensors.