Industrial control system intrusion detection architecture and method based on DCGAN (Distributed Control Generation Area Network) under side cloud collaboration

Abstract

The invention discloses an industrial control system intrusion detection architecture and method based on a DCGAN under side cloud cooperation, and belongs to the field of industrial control system security defense. According to the method, an edge-cloud cooperation scheme is adopted, offline training of an intrusion detection model is realized by utilizing the high-speed computing capability, mass storage space and global analysis characteristics of a cloud platform, real-time intrusion detection is realized by cooperating with the low-delay characteristic of edge computing, and real-time and global active safety protection is provided for an industrial control system. Meanwhile, the DCGAN is applied to intrusion detection of the industrial control system, the early-stage data sample marking process can be omitted, the workload is reduced, and compared with a complex deep learning detection model, the DCGAN model is simple in structure, high in detection precision and capable of improving the detection efficiency; and the detection model can perform intrusion detection on the communication network layer data and the field layer data at the same time, so that double-closed-loop protection of the communication network layer and the field layer is realized.

Description

technical field [0001] The invention belongs to the field of security defense of industrial control systems, and more specifically relates to a framework and method for intrusion detection of industrial control systems based on DCGAN under edge-cloud collaboration. Background technique [0002] Industrial Control System (ICS) is widely used in military industry, chemical industry, water plant, power plant and other safety-critical systems related to national economy and people's livelihood. Accelerate, the Internet, cloud computing and other emerging technologies are combined with traditional systems, industrial control systems are transformed and upgraded from traditional industries to digital, networked and intelligent, and gradually become the brain and central nervous system of national key infrastructure and various industrial production. But at the same time, the security threat of industrial control systems being attacked by cyber attacks is also spreading day by day....

AI Technical Summary

Problems solved by technology

Existing industrial control system intrusion detection technologies mostly use supervised classification models. However, industrial control systems have the characteristics of large data volume and high dimensionality. Supervised classification methods need to mark samples in the early stage, and the workload is heavy and complicated.

Moreover, most industrial control systems work in a normal environment, resulting in too few attack data samples and unbalanced data samples. Even if the attack data is expanded, it is difficult to achieve the effect of actual attack samples.

Images