Research On Intrusion Detection System For Subway Communication Network Based On Fastgrnn

Real-time Ethernet technology has been used more and more widely in subway communication network due to its advantages of low cost and high speed.However,the openness of Ethernet also increases the possibility of malicious attacks and intrusions on the network.In order to ensure the safety of passengers,the subway communication network system needs to establish an efficient and accurate network intrusion detection system.In this paper,relying on the horizontal development subject of the research group based on the subway passenger information system,combined with the structural characteristics of the subway communication network,aiming at low deployment cost,high detection accuracy and low detection delay,an intrusion detection system suitable for the subway communication network is proposed.The subway communication network intrusion detection system proposed in this paper includes the intrusion detection nodes located in each carriage and the master control equipment located in the driver’s cab.The intrusion detection node performs intrusion detection on the traffic data flowing through the carriage,and sends the detection results to the master control device in the driver’s cab.The master control device summarizes the network abnormalities in each car,and prompts and displays the results.After research and experimental verification of various machine learning models suitable for intrusion detection tasks,especially models based on recurrent neural networks,this paper chooses the Fast GRNN（Fast Gated Recurrent Neural Network）model as the intrusion detection model.The Fast GRNN model is improved from the recurrent neural network.The theoretical analysis and experimental results in this paper show that the Fast GRNN model maintains the same or even better performance than the gated recurrent unit（GRU）and the long short-term memory network（LSTM）,which are also based on the recurrent neural network.While achieving excellent detection accuracy,memory usage and detection delay are significantly reduced.With the help of the low hardware resource occupation of Fast GRNN,the intrusion detection model is depolyed directly on the existing equipment of the subway communication network,which further reduces the difficulty and cost of deployment,and introduces a layer of protection by the intrusion detection system into the communication network without making major changes to the network topology.In this paper,the guide screen system is selected as the hardware device for deploying the intrusion detection nodes of each carriage.In order to verify the validity and reliability of the proposed system,this paper selects Qt as the GUI framework,and implements various subsystems,including the intrusion detection system,the guide screen system and the master control screen system,and completes the virtual machine test on the PC.The system is deployed on the STM32MP1 embedded platform.The transplantation and performance test results show that the intrusion detection system of the subway communication network proposed in this paper has a low system occupancy.Even if it is deployed on the existing hardware in the subway car,it can ensure that the intrusion detection and the original tasks run smoothly.