Research On Intrusion Detection Algorithm Based On Vehicle FlexRay Network

With the rapid development of the Internet,the age of the Internet of Things(Io T)has also come to people.At the same time,the Internet has gradually penetrated into the automotive industry.The rapid development of in-vehicle network technology has made the car more and more intelligent and digital.Automotive Bus is the Basis for Digital Networking.For example,electronically controlled windows,brake control,and airbags are used for data transmission via the bus.The access to more networked features also exposes the car's own security holes,such as when the car network is attacked,it will directly threaten the safety of drivers and other vehicles on the road.Although the CAN bus is the most widely-used vehicle-mounted bus in today's automobiles,the CAN bus is ineffective in certain automotive systems that require high real-time performance and transmission speed.The FlexRay bus combines time triggering and event triggering.The transmission speed is ten times that of the CAN bus and is bound to become the next-generation car backbone network.Under this background,the automotive industry puts forward higher requirements for the safety of the vehicle-mounted FlexRay bus.In the in-vehicle network,the traditional technologies such as encryption and decryption and identity authentication on the bus can no longer fully meet current security requirements.In recent years,intrusion detection technology has attracted much attention because of its good protection effect.Therefore,how to use reasonable protection measures according to the different characteristics of different vehicle networks is currently a problem worthy of study.In this paper,we first study the algorithms that belong to different categories in intrusion detection,and according to the characteristics of each algorithm,after further screening,we finally select the support vector machine to construct the intrusion detection model of this paper and carry out follow-up experiments,at the same time,analyze the test results.Among many intrusion detection algorithms,thesupport vector machine originates from machine learning and has good learning ability.At the same time,the introduction of the kernel function also solves the "dimension disaster" problem.Combining the characteristics of message data in the Flex Ray network,this paper introduces a Gaussian kernel design SVM intrusion detection model for vehicle-mounted Flex Ray network.This can be used for intrusion detection to protect the safety of car communications.The following is the main work and research content of this article:(1)Four kinds of common vehicle buses are introduced and the significance of studying the Flex Ray bus is described.Further study the FlexRay bus protocol,and lay a solid theoretical foundation for the establishment of communication networks in subsequent experiments.At the same time,it outlines the intrusion detection method,which analyzes the characteristics of several common intrusion detection algorithms,and then selects the appropriate algorithm for establishing the detection model and completes its intrusion detection under the FlexRay network.(2)A SVM intrusion detection model for Flex Ray network is designed and a Gaussian kernel is introduced into the model.Each module function in the model is closely connected.By collecting,classifying,preprocessing,and training the SVM model of the content of the message in the FlexRay bus network,the optimal intrusion detection model in this environment is obtained.(3)For the designed intrusion detection model,use CANoe and other equipment to complete specific experiments and elaborate the experimental process to verify the accuracy of the model detection.Select the model with higher accuracy as the final intrusion detection model,verify it again,and draw the experimental conclusion.