Based On Image Recognition And Motion Control Technology AGV Car Design And Development

AGV (Automatic/Autonomous Guided Vehicle) is an autonomous system integrated with technologies in varied disciplines such as mechanical engineering, electric and electronic engineering, computer science and informatics. As the development of automatic and intelligent conceptions in modern society, AGVs will definitely become more and more useful in manufacturing processes, logistics and family life.In this dissertation, starting from the foundational theory of wheeled mobile robotics with the reference to widely applied present AGV systems in industries; we analyze those pivotal problems and research status. The outlines as followed:(1) By analyzing the configuration of chassis based on the WMR (Wheeled Mobile Robot) theory, a variation of four-wheel chassis is selected to design the vision-based AGV. Three-dimensional model is created in Pro/E.(2) In terms of control system including hardware and software, a dual-core system using ARM Cortex Series microcontrollers is designed and implemented, in order to meet the demand of vision-based AGV.(3) By analyzing widely used feature detecting and matching methods in image-processing areas, an improved algorithm using advanced edge detection is used to recognize the route information. Some optimal steps are implemented to adapt to the specific platform, and the results indicate a good performance on real-time and anti-interference.(4) With the research on optimal path tracker with multi-step prediction and path tracker based on fuzzy control theory, we prove that for AGV using steering chassis, the former one is best in accuracy and the latter is best in adaptability.Finally, we developed a prototype for experiments, and the results indicated that the AGV system and control algorithms in this dissertation are feasible and effective, which offer some experience of applying vision-based AGV into industries.