In-vehicle network intrusion detection method and system based on capsule neural network

Abstract

The invention belongs to the technical field of automotive electronics and discloses an in-vehicle network intrusion detection method and system based on a capsule neural network. Real-time dynamic data of vehicle CAN and MOST buses are used as packet frequency in original data. Sequence information is used as features. Vehicle driving state information is used as features, and related special methods are carried out to convert the data into a feature matrix for processing. The capsule neural network can perform high-order modeling on the correlation between the features. A model based on thecapsule neural network is introduced. The structural relation between the feature data is mined, the accuracy of a traditional neural network intrusion detection method is improved, the safety of vehicle driving is enhanced, and meanwhile the model is more universal and better in practicability.

Description

technical field [0001] The invention belongs to the technical field of automotive electronics, and in particular relates to a method and system for intrusion detection of an in-vehicle network based on a capsule neural network. Background technique [0002] Currently, the closest prior art: [0003] Intelligentization and networking have led to a rapid increase in the number of electronic devices inside automobiles, and the electronic control system has become increasingly complex. There are more and more information interactions between these vehicle-mounted electronic devices and electronic control units and the outside world, and most of these vehicle-mounted devices and electronic control units are connected to the bus network inside the car. The interface penetrates into the key vehicle bus network system. Hackers can use security loopholes to steal information and attack vehicles. Once a vehicle is maliciously overridden, it will pose a serious threat to people's liv...

AI Technical Summary

Problems solved by technology

[0006] (1) In the prior art, using the existing Internet or Ethernet intrusion detection method has poor applicability to the internal network of the vehicle;

[0007] (2) Some methods for the intra-vehicle network only rely on a certain type of bus data, and it is difficult to detect the possible threats inside the entire vehicle; the existing methods do not combine the driving status information of the vehicle itself for analysis, which increases false positives Rate;

[0008] (3) At the same time, most of the existing methods do not consider the correlation between different feature types of data, and it is difficult to perform high-dimensional modeling of feature relationships only by simple neural network methods, which reduces the accurate detection rate of the method

Images