Visual Navigation Of AGV Based On Image Processing

AGV is the abbreviation of unmanned automatic guided vehicle. It is an important equipment in modern logistics and workshop, serving as a carrier in the workshop along the specified path. With the development of technology, the navigation methods of AGV are developing in multiple aspects. Combined with fuzzy theory, neural networks and other intelligence technology, researchers harvest many scientific achievements. Visual navigation is a novel navigation method and becomes a hot issue in the international AGV field, due to the following three reasons: first, its path laying is simple and easy maintained; second, it can identify a variety of road signs; third, it has good flexibility.This article focuses on the research of visual navigation of AGV based on image processing. The digital images captured from camera equipment get processed and compared at first so as to determine more suitable image pre-processing method for visual navigation. Then, make use of the morphological knowledge to realize image edge detection to determine the location of the path center line and fit a straight line according to the location as the final result. Next, fuzzy controller is designed, whose two inputs are positional deviation and Angle deviation respectively. Its output, the voltage value, is used to drive the differential rate of the two-wheel motors and adjust the speed as well as direction to meet the demand.Fuzzy controller designed in this paper passes the simulation test in Matlab/Simulink environment, and could track the circular path with smaller error limitation . The experimental platform verifies this method, various performance of the AGV are tested, and ultimately the conclusion of its effectiveness and feasibility is obtained.