Research On Navigation Technology Of Intelligent Car Based On Ultrasonic And Visual Positioning

With the development of industrial automation,automated guided vehicle(AGV)has been the key equipment in workshop logistics system.Currently,the most common used AGV relies on magnetic conductor to handle materials.It has poor motion flexibility and can't adapt to complex work environment.Flexible AGV working under free path is the current development trend.This paper takes intelligent as similar object of AGV and does research on indoor positioning,path planning and obstacle avoidance.To realize self-localization of the intelligent car,an indoor ultrasonic positioning system is built with 1 emitter and 4 receivers.Ultrasonic and radio-frequency(RF)signals are both used to achieve time difference of arrival(TDOA)ranging.And improved trilateral algorithm is proposed and applied to realize wireless positioning of the intelligent car.On the basis of analyzing the advantages and disadvantages of different modeling methods,grid method is chosen to describe known environment.And multilevel optimization method is used to complete global path planning.From comparison between A*algorithm and ant colony algorithm to elite ant colony algorithm and to parameter optimization of ant colony model,the optimum solution is obtained step by step.Simulation results on Matlab show the path length is the shortest and the convergence rate is fast under the optimal parameters.The intelligent car hardware platform is built and monocular vision localization algorithm and obstacle algorithm are researched.The parameters of camera are calibrated by Matlab toolbox.By constrain z-coordinate of the barrier in world coordinate system,monocular vision positioning model is obtained.In terms of obstacle avoidance algorithm,obstacle avoidance direction can be obtained by relationship between barrier boundary and position of intelligent car and optimum heading angle can be obtained by evaluation function.Processing is applied to develop the upper software program.The real-time position and movement of the intelligent car can be tracked on two main interfaces.Experimental results show that positioning accuracy and navigation accuracy both meet standard.The algorithms researched in this paper provide reference for navigation design of AGV.