ScoopNet: 6DOF Pose Estimation pipeline for Origami-inspired Worm Robots

Abstract

Origami-inspired soft and flexible robots have drawn immense attention in recent years for their wide range of medicine and engineering applications. While the ability of shape morphing presents a significant advantage, the shape-invariant pose estimation techniques are still under-explored. Pose estimation and tracking are vital to study, control and automate the locomotion of origami robots. This paper proposes ScoopNet that performs semantic segmentation and 6DOF pose estimation of origami-inspired worm robots. A vision-based deep learning model that can estimate the pose of a shape-morphing origami worm robots warrant a vast amount of annotated dataset. To overcome this, the network is trained using a self-supervised approach on a domain randomized synthetic dataset to infer the robot’s pose in real-world data. This paper also introduces a simulation setup that can be utilized to generate a massive synthetic dataset with ease.