Design of rolling capsule-type miniature robot (RCMR) for pipe inspection

Abstract

Study on the dynamics of activation in real-time and autonomous position control of the target point of a miniature bio-robotic device is a challenging research domain. Miniature robotics involves components that come in compact sizes. Manufacturing these intricate components through the conventional route and getting the right functional output for the desired application is always formidable. The work reports successful CAD design, kinematic simulation and static analysis of a tethered miniature compliant rolling device in Fusion 360, Solidworks and FEASTSMT respectively. This is followed by hardware experimentation, prototyping and fabrication using 3D printing technology. The average device dimension of all the design variations is approximately 20 x 40 mm and thus the component is capable of rolling into a pipeline with a diameter as small as 50 mm. The back lid of the rolling component can act as a mound for a small charged coupled device (CCD) camera which will in turn be used for internal inspection of pipelines and tubes. The work further sheds light on deep learning techniques such as multilabel image segmentation and classification that can be implemented on the images captured by the mounted camera for offline image based defect detection in order to classify and localize internal pipe defects for maintenance.