Research On Attack And Defense Strategy Of Cyber-physical System Based On Denial-of-service Attack And Detection

Cyber-Physical System(CPS)is a large-scale complex system with high integration and collaborative planning between information space and physical space.CPS has been widely used in various fields due to its highly integrated and highly networked characteristics,which are different from traditional control systems.The open sharing of network layer not only brings higher efficiency to the practical application of CPS,but also increases the risk suffered by CPS.Among these potential risks,network attacks have become the primary security problem to be solved by CPS because of its great destructive power.Denial-of-service(DoS)attacks are the most common and most effective types of network attacks,which can cause the system to refuse to send or receive normal data by occupying network resources of communication channels.Among them,Denial-of-Service(DoS)attack is the most common and influential network attack.It can cause the system to reject the normal data receiving and sending requests by occupying the network resources of the communication channel.From the perspective of control,this paper studies the attack and defense strategy of CPS based on DoS attack and detection,and the specific research contents are as follows:First of all,in order to obtain the optimal solution of the attacker,a DOS optimal attack strategy which can maximize the expected average error is constructed,and the calculation formula of the attacker's profit is given.In this paper,we consider a system in which the sensor processes the measured data and sends the data to the remote estimator through the wireless channel.The attacker has limited energy budget and decides whether to block the channel at each sampling time point to reduce the quality of remote state estimation.On this contrary,the worst attack strategy and profit calculation formula under DoS attack are also given.In order to reduce the benefit of attackers,an optimal attack strategy for intrusion detection system(IDS)based on remote state estimation is proposed,and the effectiveness of IDS is verified by simulation.Secondly,the problem of attack defense game with energy constraints between sensor and attacker is studied.In order to obtain the optimal solution of the game,the interactive decision-making process of when the sensor sends packets and when the attacker attacks is established,and a game theory framework is constructed.It is proved that the optimal strategies of both sides constitute a Nash equilibrium of zero sum game.This paper proposes a dynamic update algorithm based on online information for both sensor and attacker.In order to solve the problem of computational complexity,the method of using the average energy constraint instead of the whole energy constraint is proposed,and the corresponding solution is given by using Markov chain theory.The simulation results show that the new energy constraint is effective.Finally,we continue to study the optimal attack decision of DoS attack in wireless network control,and analyze the optimal attack strategy which can maximize the linear quadratic Gaussian(LQG)function under the energy constraint.From the perspective of control system performance,this paper studies how attackers optimize their attack plans to maximize the impact on system performance when they are lack of energy.At the same time,the stability of the system under the optimal attack strategy is considered,and the effectiveness of the proposed optimal attack strategy is verified by an example.