Research On Binocular Vision-based Navigation System Of A Wall-Cleaning Robot

Mobile robot navigation technology is an important research direction and a key technology of intelligent robot. In the last few years, along with rapid development of the image process technology and the computer operation ability, and the merits of vision signal of wide searching range and integrated information, the vision-based navigation becomes one of the primary development directions for robot navigation. The main purpose of the paper is to design vision-based navigation system for the wall-cleaning robot, so that the robot can complete cleaning task independently.The wall-cleaning robot vision-based navigation system vests in machine vision, while is also the researching focus of artificial intelligence. At present, the machine vision is mostly established on PC-BASED system or PLC-BASED system. Because requests of software and hardware is too high, the volume is too bulky, and the cost is too expensive, the application is limited. However, the embedded system which applied in controlling field has merits of small volume, low power, high capability-to-price ratio, flexible utilization, reliable working and so on, so obtained the widespread application in various fields. Therefore, if can the machine vision established in the embedded system, not only can it be developed application scope of the machine vision, but also is the technological development tendency. Therefore, the paper tries to unify the machine vision and the embedded system, and research robot embedded vision-based navigation.Firstly the topic background of the wall-cleaning robot vision-based navigation is elaborated, then the developing process and the domestic and foreign research present situation are introduced. After the working condition characteristic of the wall-cleaning robot is analyzed, an embedded system of vision-based navigation system is designed. In order to insure the robot can move quickly and the system can work reliably, the questions of real-time and reliability are researched. The parallel binocular vision system is established, the obstacle avoiding algorithm is studied. An integrated software module is designed, and the MATLAB programs of obstacle-avoiding are compiled, and then the validity of the algorithm is confirmed through experiments. The model of two-dimensional Descartes rectangle grid for glass-frame wall of buildings is established, the path planning method based on the model is put forward, and the principle is detailedly discussed.