Structure Design And Navigation Algorithm Research Of AGV Base On Machine Vision

Based on the study of AGV related technology, and combined with the performance requirements, this paper stated the structure design of AGV. And it presented a form of machine vision navigation on which AGV runs at a certain distance along the road marking line. With the navigation algorithm research, this paper explored the new application in the field of AGV.In the aspect of structure design, it consists of the driver module design, the traction module design, the vehicle frame design, the panels design and the components selection. The motors are arranged in a staggered way in the driver module. This way can realize AGV differential motion well. Linear actuator is used as the driving mechanism of traction module. This design can simplify the structure validly. The vehicle frame is welded by equilateral angle steel according to the device layout. The panels are designed on the basis of vehicle contour and function required. The main components are selected with the requirement of AGV.In the aspect of navigation algorithm research, it includes the extraction algorithm of navigation deviate parameter, the identification algorithm of navigation identifier and the fuzzy control algorithm. The extraction algorithm consists of the binary processing of gray image, the morphological processing which can extract the line edge, the inverse perspective transform which can make the line edge mapped into the world coordinate system, the Hof transform which can obtain navigation deviate parameter by the equation of line edge. The identification algorithm includes the tilt correction, positioning and segmentation processing of navigation identifier, the feature extraction processing which can transform identifier into LED digital image, and the template matching method which can get the final identify results. The fuzzy control algorithm is designed according to the motion and control law of AGV. And the algorithm is analyzed with simulation in MATLAB.Finally, on the basis of AGV structure and navigation algorithm, this paper built the AGV experimental platform to do experiments in which AGV runs long the marking line and identifies navigation identifier. Experimental results verify the feasibility of the navigation method and the correctness of navigation algorithm.