Research On Vehicle System Of AGV Based On Visual Navigation

The paper obtains the funding of the national science foundation of China(51265039) and production projects of Guangdong province (2010B090400267).Along with the globalization of industrial production and material circulation, thelogistics is referred to as “the third profit source” and more and more people in thefield become focus on it. The industrial logistics system, with the automatic guidedvehicles (AGV) as the principal part, has become the focus of the relevant enterprises.The rapid development of machine vision and sensor technology, made it possible todevelop the intelligent AGV.Based on the research and conclusion on the basis of predecessors’achievements,this paper proposes a AGV vehicle system design scheme based on vision navigation,through integrate the machine vision technology and multi-sensor information fusiontechnology, upgrading the intelligent of traditional AGV and achieving the AGVvision navigation, autonomous obstacle avoidance, vehicle tracking, etc.This article main content is as follows:First, this paper discusses the navigate mode of AGV and compares with them.On the basis of the research and summarizes the AGV development present situationat home and abroad. This paper proposed the feasible design scheme of the AGVvehicle system based on visual navigation. By considering the cost, quality, efficiencyand the practical application environment, establish the overall decision framework toevaluate feasible scheme, optimization solutions are obtained.Secondly, set up a visual navigation hardware platform, and based on processing,analysis and identification the image sequence collected by the vision sensors obtainthe information of navigation mark line, location identification points and obstacles,and make the informations as the input of the AGV motion control. Then through theMATLAB/simulation verified the reliability of image recognition algorithm.Thirdly, improve the function of existing AGV based on visual navigation, addthe function of intelligent obstacle avoidance. Through synthesize the obstacleinformation collected by the front vision sensor of AGV and the ultrasonic sensorinstall around the bodywork, constitute the information basis of the intelligenceobstacle avoidance system, and through the fuzzy control method for obstacleavoidance. Then make some test to verify the accuracy of the sensor, and the verification results conform to the requirements for industrial application.Finally, the functionality of visual navigation module and intelligent obstacleavoidance module is verified by experiment, and the algorithm is verified with wellrobustness.The AGV vehicle system based on visual navigation can be considered as thefunctional upgrade of the traditional model, blended in machine vision and sensingtechnology, and has played a good role for the development and promotion ofintelligent AGV.