Robotics-Assisted 3D Scanning of Aircraft

Abstract

3D digital modeling of aircraft is important for aircraft maintenance, repair, and overhaul (MRO). However, it is very labor intensive and time consuming to generate an accurate digital outline surface model of an aircraft or even a large portion of it because the scanning work still must be done manually. We propose a solution of automated 3D scanning of an aircraft by applying the latest robotics, artificial intelligence, and sensing technologies. A robotic system, consisting of an unmanned aerial vehicle (UAV) with a RGB-D camera, an unmanned ground vehicle (UGV) and a manipulator equipped with a high-precision and close-range 3D scanner, is used to perform automated scanning for acquiring 3D model of an aircraft by following an optimal scanning trajectory computed by using reinforcement learning technique. The resulting 3D model can be analyzed for identifying widespread defects, such as dents and cracks, on the surface of aircraft fuselage, wings, and other components. Such a solution makes the scanning process much more efficient and of higher quality. Further, the solution also removes the tedious, boring, and sometimes even risky human labor from the scanning work. The focus of this paper is using Monte Carlo Tree Search algorithm to learn an optimal scanning trajectory based on a low-resolution point cloud model of an aircraft.