Spoofing Attack And Denial Of Service Attack Detection Model For Quadrotor

The quadrotor has made great progress in recent years.In the future,it will be widely used in many fields such as government affairs,commercial logistics,and entertainment.It can promote the development of the economic industry in a more three-dimensional space.However,with the gradual popularization of civilian quadrotor,it has been the target of many attackers.If the quadrotor needs to be widely put into commercial application,solving the security problem is a significant issue.Considering with the structural characteristics of the flight control system,spoofing attacks and denial-of-service attacks are the most common means in several kinds of attacks,which happens in the actuator control process or the feedback process of sensors.These two attacks can directly invade the quadrotor.In order to reject the attack of these two types of attacks on the quadrotor and improve its security,the intrusion detection technology is selected to research in this paper,which can protect actively.The intrusion detection models are built in the information domain and the physical domain.The following research work has been carried out in detail:(1)Firstly,the survey on the current drone security research situation and the intrusion detection method in the cyber ph ysical system is carried out.Here analyzes the similarities and differences between the security of cyber physical system and the security of IT.Combined with the information transmission characteristics and flight dynamics equations of the quadrotor,the information domain and physical domain of the system are modeled and analyzed respectively.The characteristics of the countermeasure target spoofing attack and denial of service attack are analyzed separately in this study.(2)At present,the intrusion detection system attaches importance to the characteristics of information content and ignores the characteristics of timestamp.An intrusion detection model based on graphical features is proposed in the information domain.By visualizing the time interval and time series of information sending time as the original feature,the vertex coordinates and various graphic areas of the graph are used to form a new feature dataset,and the classification model of intrusion detection is obtained by training.The model contains the scatter graph sub-model and the polygon graph sub-model,the former aiming at the spoofing attack,while the latter is effectively for the denial of service attack.(3)The main purpose of the attack is to interfere with the normal flight state of the quadrotor,causing the mission to fail.It is an idea to judge the attack by estimating the flight state.In order to make the intrusion detection system more rigorous,here want to play the effect of collaborative detection between information domain and physical domain.In the physical domain,the dynamic mathematical model of the quadrotor is built as the research obj ect,and the particle filter algorithm is applied to it.The flight state of the system is predicted,and the abnormality of quadrotor is detected at the first time by the threshold.The intrusion detection process is implemented in the physical domain.(4)Considering the limited resources of civilian low-cost situation,the attacker and defender are used as game participants to establish a non-cooperative complete information static game model.By discussing the pure strategy Nash equilibrium and the hybrid strategy Nash equilibrium,the defender obtain an optimal strategy.If a resource-constrained situation is encountered,defender would select an information intrusion detection model or physical model in accordance with the probability recommended by the optimal strategy,so as to achieve the goal of balancing detection performance and cost.According to the common principle of civilian quadrotor,the research object was built in MATLAB/Simulink software.The simulation and experimental verification of the algorithm and model discussed in the paper were carried out.The feasibility and effectiveness of the intrusion detection model were proved by experimental tests.