Anomaly Detection Method For In-vehicle CAN Bus Based On Renyi Information Entropy

In recent years,with the development of the automobile industry,the Internet has gradually entered this traditional field.It not only brings the new vitality of the car,but also brings in the concept of Internet of Vehicles(IOV).The function of remote control car,such as opening the door by mobile phone,adjusting the temperature of air conditioning,switching on or off windows,embedded automobile network etc.All of these features have greatly facilitated people’s driving experience.However,these features exposed more safety loopholes to some extent,that is to say,the more functions the vehicle have,the higher the criminals will find the safety loopholes.Thus,IOV’s security has been closely related with our driving security,how to protect vehicles from the criminals attack in a better way has become a very important subject.CAN bus is the most widely used in-vehicle bus network at present while people did not take into account that the vehicle network would no longer confined to invehicle network one day as they proposed CAN bus technology,which caused the attack protection capabilities from the outside world has been very poor.For this reason,the criminals take this advantage to steal car information by eavesdropping attack or replay attack which will inevitably put the vehicle and personal safety under threat.Therefore,how to take appropriate defensive measures according to the characteristics of the vehicles has become an urgent issue.At the moment,people have grasped some means of protection for the information security of CAN bus,such as message authentication as well as encryption means.But the in-vehicle network is different from traditional Internet;it has its own limitations,such as hardware computing power,memory constraints,vehicle real-time demand and so on.Anomaly detection method can make a detection of network behavior to alarm the abnormal behavior,which will make up for the inadequacy of encryption authentication.The concept of entropy originates from physics to describe the degree of disorder of a thermodynamic system.After Shannon’s introduction of the information theory,it represents a measurement of uncertainty.Through the analysis of the thesis,it can be learned that Renyi information entropy(the expansion of information entropy)has a good adaptability with in-vehicle CAN bus.Hence,people can use anomaly detection for vehicle safety protection.Based on the above,this thesis proposes a method of vehicle protection: anomaly detection method for in-vehicle CAN bus based on Renyi information entropy.The main work and research contents including:(1)Introduction about the status and threat of IOV.(2)Introduction about characteristics and deficiencies of CAN bus;summarization about the possible attack methods for CAN bus;analysis and explanation about the current research status of CAN bus safety;illustration of the characteristics,classification of anomaly detection method as well as the analysis of its environmental adaptability with CAN bus.(3)Introduction about the concept and nature of entropy meanwhile the author put forward anomaly detection method for in-vehicle CAN bus based on Renyi information entropy.Through the analysis and statistics of network message of CAN bus,calculating the divergence of Renyi information entropy and Renyi to establish anomaly detection model.(4)According to anomaly detection model proposed in the thesis,experimental verification is made to test the feasibility of the method.Then the author made a calibration about the length of the time slice.On the one hand,to make sure the normal detection interval in order to draw the conclusion of the experiment,on the other hand,to make an improvement about the detection model.