DEVELOPMENT OF VISION-BASED CONTROL SYSTEM FOR MOBILE 3D PRINTER

The 3D printing technologies can produce objects with a very complex shape or geometry nowadays thanks to the advanced researches in their precision, repeatability, material ranges, etc. The size of 3D printed objects also varies. The Oak Ridge National Lab printed a 17.5 feet long, 5.5 feet wide and 1.5 feet tall tool for Boeing that weights 1,650 lbs. On the other hand, using Two Photon Polymerization, sub-mu m structures can be produced. The majority of current 3D printers design resembles the traditional FDM 3D printer. The printer is stationary during the printing period, limiting the print zone by the size of the robots, regardless as to the printer designs (Cartesian, polar, delta or articulated robots). In addition, all current 3D printers work as stand-alone equipment, which prevents the possibility of further speeding up the fabrication by using multiple collaborating 3D printers. Using mobile robots as a 3D printer could eliminate the size limit of the print zone and enable collaboration among different mobile robots to speed up the printing process. However, major problems are remaining unsolved in the current mobile 3D printer research, such as the precise localization of the robot, material slipping, accumulative printing error, etc. In this paper, a vision-based feedback control system is presented as a solution to these problems in mobile 3D printers. The system is equipped a single camera as sensory input. Using Simultaneous Localization and Mapping (SLAM) methods, the mobile 3D printer could potentially achieve sub-millimeter accuracy for localization. The slipping and accumulative error could also be mediated using image processing and object recognition. The system also enables the possibility for multiple 3D printers to work simultaneously. It is believed that mobile 3D printers equipped with vision-based feedback control system could have a great potential in the future.