Study On High Precision Autonomous Positioning Technology Of Metro Intelligent Inspection Vehicle

Rapid detection of infrastructure is an important guarantee for the safety of urban rail transit operations.Aiming at the problem that the train positioning control system of the existing rail transit train has low positioning accuracy and can not meet the requirements of the automatic inspection vehicle,this paper designs a high-precision real-time positioning technology based on multi-sensor fusion.The centimeter-level positioning accuracy can be realized,which provides support for the automatic driving of the inspection vehicle and the high-precision positioning of the dynamic detection data.In order to correct the accumulated mileage error caused by the car body snake shape,idling and slipping,and to realize the time and space synchronization of the line data collected by the inspection vehicle inspection system to ensure the meaningful collection information,this paper designs a multi-sensor fusion.The positioning technology,then gives the overall design of the positioning system,including system hardware design and high-precision positioning algorithm.The system hardware design includes sensor selection,functional module composition and working principle.The algorithm design is based on multi-sensor fusion.It consists of initial positioning and real-time positioning.The positioning algorithm is also the main innovation of this paper,and it stipulates the expected system.Level positioning accuracy and requirements for real-time positioning of the tunnel site.In order to realize the real-time positioning of the inspection vehicle,this paper builds an efficient and reliable hardware platform capable of real-time acquisition and data processing.The sensor is optimized and added based on the existing positioning system in the laboratory.The axle counter and laser displacement are mainly designed.The data acquisition module of the device includes the interface design with the FPGA and the internal acquisition logic circuit design.Finally,the software driver corresponding to each sensor acquisition board and the detailed explanation of each function call are written.The high-precision positioning algorithm is divided into two parts:initial positioning and real-time positioning.The initial positioning is based on RANSAC algorithm and vehicle body attitude compensation to extract the tunnel section environmental characteristics,and then the FastSLAM problem is processed by particle filtering,including the establishment and importance of the motion model and the observation model.Weight calculation.The real-time positioning algorithm is mainly based on the group "window" idea,combined with the initial positioning of the kilometer and the track characteristics of the ballast/fastener,to achieve the correction matching of the reference curve and the real-time curve.In the algorithm,the cumulative error of the mileage caused by the snake body shape,idling,slipping,etc.,and the misalignment error caused by the missed detection and misdetection of the fastener are eliminated.Finally,in order to verify the accuracy and reliability of the positioning system,the simulation model was combined with the simulation model in the laboratory environment in the early stage,and the data verification was carried out in the Guangzhou Metro site in the later stage.The experimental results show that the initial positioning accuracy can reach±0.3m,and the real-time positioning accuracy can reach±10cm.The high-precision real-time positioning in the huge subway line database has been realized,and the expected requirements have been met.