Font Size: a A A

Research On Key Technologies Of Localization And Mapping In Dual Steering Wheel AGV

Posted on:2020-03-05Degree:MasterType:Thesis
Country:ChinaCandidate:R D MeiFull Text:PDF
GTID:2428330599459392Subject:Power Machinery and Engineering
Abstract/Summary:PDF Full Text Request
Automated Guided Vehicle is developing rapidly in intelligent manufacturing industry.Positioning and mapping were one of the core issues in implementing AGV systems.Based on the actual project,this paper built a double steering wheel AGV handling platform in the actual factory,and conducted in-depth research on the positioning and mapping technology based on the double steering wheel AGV.The main research work includes:The chassis kinematics model and the navigation sensor were compared and selected,then the camera model was constructed,the camera was calibrated,the image was extracted and matched,and the ICP algorithm was implemented.Finally,the odometer was built.The graph optimization algorithm was studied in detail,and the experimental comparison was made to construct the map of the plant.By analyzed the advantages and disadvantages of the graph optimization algorithm,It was proved that the graph optimization algorithm has high precision,but the real-time performance was not good.Particle filter algorithm was used in Real-time positioning.The Monte Carlo method and Bayesian filter algorithm were introduced into the particle filter.Then the particle filter algorithm was implemented step by step by importance sampling,sequential importance sampling and resampling algorithm.Finally,the visual odometer was added to the program to tested the car in real time,confirming the feasibility of the positioning scheme.Experimental comparison between visual odometer and encoder.The positioning accuracy and robustness of the two were analyzed under different road conditions.The results showed that whether it was on a flat road or on a rugged road,the positioning accuracy of the visual odometer was higher.The results showed that the the problem of encoder failure for double-rudder wheel AGV in the case of rough terrain and tire slippage.This paper used visual odometer and lidar for data fusion,which avoided the problem of excessive cumulative error caused by direct kinematics modeling of double steering wheel AGV.The used of RGB-D cameras for visual odometers for dual-rudder wheels AGV has good practical application value.
Keywords/Search Tags:AGV, double steering wheel, simultaneous positioning and mapping, graph optimization, particle filter
PDF Full Text Request
Related items