Research On Intelligent Recognition Of Tracks Tatus Based On ZPW-2000R Track Circuit

As a ground-vehicle information monitoring system for monitoring whether the track is idle or occupied,the track circuit plays a key role in ensuring the safety of trains and improving the operational efficiency of the urban rail transit system.However,the current track condition monitoring signal judgment based on track circuit is still in the stage of manual identification,which not only has low identification efficiency and huge consumption of manpower and material resources,but also easily leads to misjudgment and missed judgment due to the existence of unreliability due to human factors.This leads to the occurrence of rail traffic safety accidents,and even affects the normal operation of other lines in the network.Therefore,this paper mainly studies the intelligent identification of track state based on ZPW-2000 R track circuit signal.The research contents include:（1）Aiming at the application scenario of ZPW-2000 R track circuit with host tuning access,an intelligent track state identification method based on image scaling and convolutional neural network is proposed.First,the image scaling process is performed on the voltage image to reduce the image size,thereby improving the operation efficiency of convolutional neural network image recognition.Then,a convolutional neural network is used to intelligently identify the track state on the scaled image.In order to verify the effectiveness of the method in this chapter,the method in this chapter is applied to the actual track circuit data,and the results are compared with the method of numerical analysis + BP neural network.The experimental results show that the track state intelligent recognition method based on image scaling and convolutional neural network has a 100% accuracy rate for different track states,and the recognition accuracy of the two track occupancy states is higher than that of the comparison method,which reflects this paper.Advantages of the research method.（2）Aiming at the application scenario of ZPW-2000 R track circuit with no host modulation access,an intelligent track state identification method based on grayscale atlas and convolutional neural network is proposed.A grayscale map is constructed from the monitoring data of multiple devices in the track circuit to accurately describe the idle and occupied status of the track,and the grayscale map is used to intelligently identify the track status by using the convolutional neural network.In order to verify the effectiveness of the method in this chapter,the method in this chapter is applied to the actual track circuit data,and the results are compared with the method of numerical analysis + BP neural network.The experimental results show that the track state intelligent identification method based on gray-scale atlas and convolutional neural network has 100% accuracy in identifying different states of the track,and the recognition accuracy of the two track occupancy states is higher than that of the comparison method,which reflects the Advantages of the method studied in this paper.（3）Development of the track state intelligent identification prototype system based on ZPW-2000 R track circuit signal.Combined with the above research results,based on the visualization tool Pyqt5 framework,Pycharm integration framework and Anaconda environment,a parameter-configurable track state intelligent identification prototype system is designed and developed.The prototype system includes a graphical user interface,which can realize the functions of ZPW-2000 R track circuit signal data preprocessing and track status intelligent identification.The innovation of this paper is that the research object of the track circuit signal is changed from the track circuit device state to the track state,and the ability of the convolutional neural network to quickly extract the image features is used to establish a ZPW-2000 R track circuit for two different application scenarios.The convolutional neural network model uses intelligent algorithms to identify orbital states to replace the existing manual identification methods,and expands and supplements existing achievements.The research results can not only effectively improve the track state identification efficiency based on the ZPW-2000 R track circuit signal,thereby reducing the operating cost of enterprises,and because it can fundamentally avoid the misjudgment behavior of manual identification,it is also useful for improving the safety and reliability of rail transit operations.important social significance.