Research On Self-adaptation AGV Guidance Technology Based On Multi-source Information Fusion

With the continuous promotion of Internet plus initiative,the demand for the logistics industry is also increasing,emerging new requirements and requirements.Among these growing demands,the most important thing is to improve the efficiency of logistics.A new type of automatic guided vehicle(AGV)--Based on multi-source information fusion adaptive AGV will gradually get attention,and become the research focus of automated warehousing.Because it is the core of the development of intelligent storage and intelligent storage.The sensor system based on multi-source information fusion adaptive AGV is integrated by odometer,gyroscope and CCD vision sensor to collect more comprehensive and accurate external environmental information and improve the utilization of external environmental information.The multi-sensor fusion navigation system can coordinate the nonlinear problems of the AGV obstacle avoidance system,thereby enhancing the AGV's flexibility and operational efficiency and job accuracy.First of all,this paper analyzes and studies the adaptive tracking technology of AGV,and designs the adaptive tracing process.And in the process of AGV adaptive tracing,SLAM(Simultaneous Localization and Mapping)real-time positioning and map building concept is introduced.This is also the core of AGV's autonomous navigation,and the analysis and design of the SLAM navigation process of adaptive AGV.The SLAM method based on the extended Calman filter(EKF-SLAM)and localization method based on particle filter(PF-SLAM)to compare the twoalgorithm,analyze the advantages and disadvantages of the two algorithms,the EKF-SLAM method is more excellent in real-time and accuracy,so the selected EKF-SLAM method as the research methods.Secondly,according to the analysis and comparison of three kinds of information fusion structure,this paper adopts a hybrid information fusion structure as the structure of adaptive information fusion of AGV system,and the general T-S fuzzy neural network algorithm was improved,he selection of neural network clustering algorithm was developed for the number of membership degree in structure,the weights of various neural node through the model under the condition of parameter uncertainty and fusion of self-learning and adaptive to make their own choice,improves the intelligence and flexibility of the system of data fusion.Finally,the adaptive fuzzy PID control algorithm is used to control each drive motor of the AGV to realize the adaptive guidance of the AGV.Compared to the single wheel drive,two-wheel drive and four-wheel differential drive,six-wheel differential drive four drive modes,a contrastive analysis of Mecanum wheel and four round differential drive,dual wheel structure three driving structure.Combining the two analysis and contrast results,this paper proposes an adaptive AGV selection four-wheel drive mode based on multi-source information fusion.The overall technology structure of adaptive AGV based on multi-source information fusion is given:vehicle frame,ARM processor,embedded hardware design based on ARMv8,software development and transplantation with embedded Linux,visual sensor and gyroscope selection.And the MATLAB is used to simulate the EKF-SLAM algorithm used by the designed AGV.The experimental results show that the designed system can make AGV better complete the multi-sensor information collection and fusion and autonomous tracking function.