Development and commercialization of a visual learning system for robot programming by demonstration

该项目旨在通过示范将机器人编程的视觉学习系统商业化。学习系统采用视觉系统来捕捉由人类专家演示的任务。然后，将记录下来的演示动作进行处理并直接转换为机器人程序，用于机器人执行演示任务。视觉学习过程的方法是由申请人和他在瑞尔森大学机器人、机电一体化和自动化实验室的研究团队开发的。

The proposed project aims at commercializing a visual learning system for robot programming by demonstration. The leaning system employs a vision system for capturing tasks demonstrated by a human expert. Afterwards, the recorded demonstrated motions are processed and directly transferred into robot programs, which are utilized for execution of the demonstrated tasks by the robot. The methodology for the visual learning procedure has been developed by the applicant and his research team in the Robotics, Mechatronics and Automation Laboratory at Ryerson University. The proposed system is going to be either licensed to a robot manufacturer or commercialized through a start-up company, namely Visual Learning Robotics Inc. The project objectives encompass development of a prototype system for visual robotic learning, protection of the intellectual properties rights by submitting patent application, technology promotion and transfer to the receptors. The outcome of the project is a software package for visual learning that can be used by industrial companies employing robotic systems. The technology reduces the costs for development of robotic applications, and more important, it enables quick and intuitive programming of robots by end-users that do not necessarily have programming expertise. The technology receptors are manufacturing companies that employ robotic technology in their daily production, as well as robot manufacturers and software developing companies. The visual learning technology addresses the growing demand for automation in the manufacturing sector, as well as it advances the level of core competency of Canadian companies and streamlines their competitiveness in the global market.