Network attack injection semi-physical simulation platform for unmanned aerial vehicle cluster

Abstract

According to the network attack injection semi-physical simulation platform for the unmanned aerial vehicle cluster, provided by the invention, the network attack and defense scene of the unmanned aerial vehicle cluster can be simulated and the network attack resistance of the unmanned aerial vehicle cluster can be verified based on the semi-physical simulation platform, so that powerful support is provided for design, test and optimization of related functions of network security attack and defense of the unmanned aerial vehicle. The simulation platform comprises a main control computer, a three-dimensional simulation system, an unmanned aerial vehicle cluster autonomous flight system and a network attack injection system, state information of the unmanned aerial vehicle cluster is displayed and a control instruction is issued through the main control computer, and the three-dimensional simulation system constructs a simulation environment of the unmanned aerial vehicle cluster; the unmanned aerial vehicle cluster autonomous flight system receives information from the three-dimensional simulation system and the main control computer and realizes an unmanned aerial vehicle cluster task; the network attack injection system tests and evaluates the network security of the unmanned aerial vehicle through data acquisition, preprocessing, vulnerability utilization analysis and network attack injection.

Description

Technical Field

[0001] The present application relates to the technical field of drone semi-physical simulation, and in particular to a network attack injection semi-physical simulation platform for drone clusters. Background Art

[0002] Drone swarms, with their characteristics of "low, slow, small, and group", strong flexibility, and "1+1>2" collaborative combat effectiveness, have become a key link in gaining asymmetric technological advantages in future wars. The confrontation and game between the two sides of the conflict over drones fully demonstrate that drone warfare and anti-drone warfare have become the main means of modern warfare. Whether in the military or civilian fields, the research on countering drones with cyber attacks and drones' resistance to cyber attacks is of great significance.

[0003] The semi-physical simulation system adds physical hardware to the digital simulation link, which can build a test scenario similar to the actual environment while reducing costs and risks, overcoming the limitations of field experiments due to many factors. However, existing technologies cannot evaluate the network security attack and defense performance of drone clusters.

[0004] Therefore, how to simulate the network attack and defense scenarios of drone swarms based on semi-physical simulation systems and conduct experimental tests to verify the ability of drone swarms to resist network attacks is a problem that needs to be solved at present. Summary of the invention

[0005] The embodiment of the present application provides a semi-physical simulation platform for network attack injection for drone clusters, which can simulate drone cluster network attack and defense scenarios and verify the ability of drone clusters to resist network attacks based on the semi-physical simulation platform, thereby providing strong support for the design, testing and optimization of drone network security attack and defense related functions.

[0006] In order to achieve the above objectives, this application adopts the following technical solutions:

[0007] In a first aspect, the present application provides a semi-physical simulation platform for network attack injection for drone clusters, the simulation platform comprising: a main control computer, a three-dimensional simulation system, a drone cluster autonomous flight system and a network attack injection system;

[0008] The main control computer is used to display the status information of the drone cluster and issue control instructions;

[0009] A three-dimensional simulation system, including a simulation model of a drone swarm, an environment model, and a sensor model, to build a simulation environment for a drone swarm;

[0010] The UAV swarm autonomous flight system is used to receive control instructions from the main control computer and sensor data sent from the three-dimensional simulation system, and extract and estimate the current state information and environmental information of the UAV swarm based on the sensor data, and then obtain the control instructions of the UAV swarm at the next moment in the simulation environment based on the current state information and environmental information of the UAV swarm and the control instructions, and send the control instructions at the next moment to the three-dimensional simulation system;

[0011] The network attack injection system is used to collect and pre-process data related to the drone cluster, analyze potential vulnerabilities based on the pre-processed data, and then conduct network security attacks on the drone cluster based on the potential vulnerabilities to verify the drone cluster's ability to resist network security attacks.

[0012] A possible design scheme is that the platform of the first aspect also includes that the main control computer, the three-dimensional simulation system, the drone cluster autonomous flight system and the network attack injection system communicate through a public API interface.

[0013] A possible design scheme, the platform of the first aspect also includes that the status information of the drone cluster includes: location information, speed information and attitude information of the drone cluster.

[0014] A possible design scheme, the platform of the first aspect also includes that the three-dimensional simulation system is modeled using AirSim.

[0015] In a possible design scheme, the platform of the first aspect further includes: the UAV cluster autonomous flight system includes a posture estimation module and a UAV flight controller, wherein:

[0016] The pose estimation module is used to receive control instructions from the main control computer and sensor data sent from the three-dimensional simulation system, and extract and estimate the current state information and environmental information of the drone cluster based on the sensor data;

[0017] The UAV flight controller is used to receive the current state information and environmental information of the UAV cluster from the posture estimation module, and receive the control instructions from the main control computer, and then obtain the control instructions of the UAV cluster in the simulation environment at the previous moment based on the current state information and environmental information of the UAV cluster and the control instructions, and send the control instructions at the next moment to the three-dimensional simulation system.

[0018] In a possible design scheme, the platform of the first aspect further includes: the network attack injection unit includes a data acquisition module, a data preprocessing module, a vulnerability analysis module and a network attack injection module, wherein:

[0019] Data collection module, used to collect log data, device information, controller data, sensor data, actuator data and network topology of drone clusters;

[0020] Data preprocessing module, used to denoise and extract features from collected data and identify attack surface elements;

[0021] Vulnerability analysis module, which is used to evaluate the network security weaknesses of drone clusters based on attack surface elements and determine potential network attack paths;

[0022] The network attack injection module injects network attacks into the 3D simulation system according to potential network attack paths.

[0023] A possible design scheme, the platform of the first aspect also includes that the network attacks include: denial of service attacks and false data injection.

[0024] A possible design scheme, the platform of the first aspect also includes that the denial of service attack is to obtain the IP address that has not been used by the drone cluster wireless network through the collected data, use the IP address to obtain the access right of the network, and then send a large amount of invalid information to block the network to achieve the denial of service attack.

[0025] A possible design scheme, the platform of the first aspect also includes, after the false data injection is invaded into the communication network, part of the data sent between the main control computer, the three-dimensional simulation system and the drone cluster autonomous flight system is extracted and injected with false data.

[0026] In a possible design scheme, the platform of the first aspect also includes that the false data includes random information or erroneous information.

[0027] In the embodiment of the present application, the main control computer displays the status information of the drone cluster and issues control instructions, and the three-dimensional simulation system constructs the simulation environment of the drone cluster; the drone cluster autonomous flight system receives information from the three-dimensional simulation system and the main control computer to realize the drone cluster task; the network attack injection system tests and evaluates the network security of the drone through data collection, preprocessing, vulnerability exploitation analysis, and network attack injection. The simulation platform can simulate the network attack and defense scenarios of the drone cluster and verify the ability of the drone cluster to resist network attacks based on the semi-physical simulation platform, thereby providing strong support for the design, testing and optimization of the related functions of the drone network security attack and defense.

[0028] Other features and advantages of the present application will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0030] Figure 1 A schematic diagram of a semi-physical simulation platform for network attack injection into drone clusters provided in an embodiment of the present application;

[0031] Figure 2 A schematic diagram of a network attack injection system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application. At the same time, in the description of the embodiments of the present application, the terms "first", "second", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more features. In the description of the embodiments of the present application, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0033] Figure 1 A schematic diagram of a semi-physical simulation platform for network attack injection for drone clusters provided in an embodiment of the present application. Based on this simulation platform, the ability of drone clusters to resist network attacks can be verified.

[0034] like Figure 1 As shown, the semi-physical simulation platform includes: a main control computer, a three-dimensional simulation system, a drone cluster autonomous flight system, and a network attack injection system. Among them, the three-dimensional simulation system, the drone cluster autonomous flight system, and the network attack injection system are each on a computer or a virtual machine, or other possible methods, which are not limited here.

[0035] In addition, the main control computer, three-dimensional simulation system, drone swarm autonomous flight system and network attack injection system communicate through a public API interface.

[0036] The main control computer is used to display the status information of the drone cluster and issue control instructions.

[0037] It can also be said that the main control computer is the core management unit of the drone cluster system, responsible for real-time monitoring of the status, position, speed, attitude, task execution and operation status of each drone of the entire cluster. Through its display interface, the operator can view the flight status of each drone in real time, including key parameters such as flight altitude, speed, battery power, and communication signal strength. In addition, the main control computer is also responsible for issuing control instructions to the drone cluster. According to the task requirements of the cluster, the main control computer will send precise flight instructions or collaborative task arrangements to each drone through the ground control system.

[0038] The three-dimensional simulation system includes a simulation model of the drone cluster, an environment model and a sensor model to construct a simulation environment for the drone cluster.

[0039] The 3D simulation system is modeled using AirSim, and the semi-physical simulation LAN is networked based on the UDP communication protocol. The IP address of the computer (or host) where the 3D simulation system is located is 192.168.1.1, and the port number is 14550. It communicates with other systems or computers on the platform by identifying the IP address and port number.

[0040] For the simulation model, taking the quadcopter model as an example, the model includes parameters such as mass, linear drag coefficient, thrust coefficient and power coefficient. The simulation model calculates the force and torque generated by each propeller according to the control instructions sent by the drone cluster autonomous flight system at the next moment, and then calculates the kinematic state of each drone at the next moment.

[0041] The quadrotor drone can be defined as a set of 4 vertices, with the control input u i The propeller rotation speed at the four vertices is determined. Calculate the force F generated by the propeller i and torque τ i :

[0042]

[0043] Among them, C T and C pow are thrust and power coefficients respectively; ρ is the air density; D is the diameter of the propeller; ω max is the maximum angular velocity per minute.

[0044] The kinematic state of the drone cluster can be expressed as the position P of k+1 k+1 , Linear speed v k+1 , linear acceleration a, angular velocity ω and angular acceleration a ω To express, through the above 5 variables, we can calculate:

[0045]

[0046]

[0047] τ d =∫ s r ds ·dF

[0048] F net =∑ i F i +F d

[0049] τ net =∑ i [τ i +r i ·F i ]+τ d

[0050]

[0051] a ω =I -1 ·(τ net -(ω·(I·ω)))

[0052] Among them, |F d | is the absolute value of linear resistance; |dF| is the absolute value of linear damping at ds; ds is the infinitesimal surface area of ​​the carrier surface, r ds is the radius of motion of the infinitesimal surface area; τ d is the resistance moment; F net is the net force, τ net is the torque; C lin is the linear air drag coefficient, A is the cross-sectional area of ​​the drone, and I is the inertia vector.

[0053] For the environmental model, this environmental model can construct a highly realistic virtual reality environment. The three-dimensional simulation system based on AirSim mainly includes gravity, magnetic field, air pressure and atmospheric density parameters, and the calculation results are:

[0054]

[0055] tanα=2cotθ m

[0056]

[0057] ρ=P / (R·T)

[0058] Where g is the acceleration due to gravity, R e is the radius of the earth, g0 is the surface gravitational constant; cosθ mis the common magnetic latitude, θ0 and It represents the latitude and longitude of the true magnetic North Pole, B is the magnetic field strength; ρ is the atmospheric density, P is the atmospheric pressure, T is the temperature, and R is a constant.

[0059] For the sensor model, GPS, monocular camera, binocular stereo camera, lidar, inertial measurement unit, barometer, magnetometer and other sensors are integrated to collect sensor data so that the current state information and environmental information of the drone cluster can be extracted and estimated based on the sensor data.

[0060] The UAV swarm autonomous flight system adopts the open source flight controller PX4, which is used to receive control instructions from the main control computer and sensor data sent from the three-dimensional simulation system. Based on the sensor data, it extracts and estimates the current state information and environmental information of the UAV swarm, and then obtains the control instructions of the UAV swarm in the simulation environment at the next moment based on the current state information and environmental information of the UAV swarm and the control instructions, and sends the control instructions at the next moment to the three-dimensional simulation system.

[0061] It should be noted that the UAV swarm autonomous flight system includes a posture estimation module and a UAV flight controller.

[0062] The posture estimation module is used to receive control instructions from the main control computer and sensor data sent from the three-dimensional simulation system, and extract and estimate the current state information and environmental information of the drone cluster based on the sensor data.

[0063] The UAV flight controller is used to receive the current state information and environmental information of the UAV cluster from the posture estimation module, and receive the control instructions from the main control computer, and then obtain the control instructions of the UAV cluster in the simulation environment at the previous moment based on the current state information and environmental information of the UAV cluster and the control instructions, and send the control instructions at the next moment to the three-dimensional simulation system.

[0064] That is, based on the current state information and environmental information of the drone cluster and the control instructions, the drone flight controller plans a continuous motion trajectory from the starting point to the target position that meets the constraints (including no collision, shortest path length, lowest energy consumption, etc.) based on the mission planning algorithm and the given map and target position, generates the control instructions for the drone cluster at one moment in the simulation environment, and sends the control instructions for the next moment to the three-dimensional simulation system.

[0065] Afterwards, the simulation model in the three-dimensional simulation system can calculate the next moment's motion state of the UAV in the simulation environment according to the control instructions at the next moment.

[0066] It can be understood that the UAV flight controller obtains the control instructions of the UAV cluster at the previous moment in the simulation environment based on the current state information and environmental information of the UAV cluster and the control instructions, and sends the control instructions at the next moment to the three-dimensional simulation system. The simulation model in the three-dimensional simulation system can calculate the motion state of the UAV in the simulation environment at the next moment according to the control instructions at the next moment, and move. The sensor model then generates sensor data and sends it to the posture estimation module, which is forwarded to the UAV flight controller. The above steps are repeated to complete the simulation of the UAV cluster.

[0067] The network attack injection system is used to collect and pre-process data related to the drone cluster, analyze potential vulnerabilities based on the pre-processed data, and then conduct network security attacks on the drone cluster based on the potential vulnerabilities to verify the drone cluster's ability to resist network security attacks.

[0068] The network attack injection unit includes a data acquisition module, a data preprocessing module, a vulnerability analysis module, and a network attack injection module. For details, please refer to Figure 2 understand.

[0069] The data collection module is used to collect drone-related information such as log data, device information, controller data, sensor data, actuator data, and network topology of the drone cluster, without specific restrictions here.

[0070] It can be understood that the flight log data of the open source flight controller PX4, related conventional GPS, optical flow, barometer and other sensors, and actuators such as motors are collected. Communication protocols such as UDP / TCP, serial communication protocols, CAN protocols and Wi-Fi are collected to monitor data flow, event logs and network device status in the network. There is no specific restriction here.

[0071] The data preprocessing module is used to denoise and extract features from the collected data and identify attack surface elements.

[0072] Among them, attack surface elements refer to potential vulnerabilities, entry points or exposure points in a system, application or network that can be exploited by attackers. In short, it represents all the paths and interfaces that attackers can use to enter or affect the target system. When performing tasks, drone swarms usually rely on multiple communication links and control systems, which may become potential attack surfaces. Attack surface elements mainly include hardware, software, communication protocols, network interfaces, etc., which are all weak links that hackers may exploit.

[0073] The vulnerability analysis module is used to evaluate the network security weaknesses of the drone cluster based on the attack surface elements, and then determine the potential network attack paths.

[0074] This step usually includes a series of integrated analysis tools and algorithms, which are designed to comprehensively identify and evaluate possible security vulnerabilities in the cluster. Each drone in the cluster is analyzed one by one to identify potential vulnerabilities in its hardware and software architecture. For example, the embedded control system of the drone may have firmware vulnerabilities, or the communication module may use outdated or insecure encryption protocols that are easily cracked by attackers. In addition, the communication network between clusters is also an important attack surface, and it is necessary to evaluate whether there is a risk of data leakage or interruption in the communication link. For example, if the wireless communication link between the drone and the main control computer does not have sufficient encryption protection, attackers can intercept or tamper with flight instructions through man-in-the-middle attacks, etc. Through a comprehensive assessment, potential security weaknesses of the drone cluster can be identified, and possible network attack paths can be derived.

[0075] The network attack injection module injects network attacks into the 3D simulation system according to potential network attack paths.

[0076] Among them, network attacks include: denial of service attacks and false data injection.

[0077] A denial of service attack is to obtain an IP address that has not been used by the drone cluster wireless network through the collected data, use the IP address to obtain network access rights, and then send a large amount of invalid information to block the network to achieve a denial of service attack.

[0078] False data injection is to inject false data into the data sent between the main control computer, the 3D simulation system and the autonomous flight system of the UAV cluster after invading the communication network. The false data includes random information or error information.

[0079] For example, the sensor data in the 3D simulation system is X0. Considering the error E X (This error follows Gaussian distribution). When calculating the autonomous flight system of the drone cluster, according to X1 = X0 + E X Perform calculations.

[0080] Subsequently, when extracting part of the data sent between the main control computer, the 3D simulation system and the autonomous flight system of the UAV cluster and injecting false data, the sensor data is tampered with in this part and the attack vector f0 is injected into the sensor data. Then, when the autonomous flight system of the UAV cluster calculates, it follows X f =X1+f0 for calculation.

[0081] For example, when extracting part of the data sent between the main control computer, the 3D simulation system and the autonomous flight system of the UAV cluster and injecting false data, the force F generated by the propeller is tampered with in that part. i , the force generated by the propeller is injected with the attack vector f1, then the force generated by the propeller becomes F if, F if =F i +f1.

[0082] The above two examples are two ways of injecting virtual data, but because the various parts and links are interconnected, they can disrupt the normal operation of the drone cluster.

[0083] In summary, in the embodiment of the present application, the semi-physical simulation platform for network attack injection for drone clusters displays the status information of the drone cluster and issues control instructions through the main control computer, and the three-dimensional simulation system builds the simulation environment of the drone cluster; the drone cluster autonomous flight system receives information from the three-dimensional simulation system and the main control computer to realize the drone cluster task; the network attack injection system tests and evaluates the network security of drones through data collection, preprocessing, vulnerability exploitation analysis, and network attack injection. The simulation platform can simulate the network attack and defense scenarios of drone clusters and verify the ability of drone clusters to resist network attacks based on the semi-physical simulation platform, thereby providing strong support for the design, testing and optimization of related functions of drone network security attack and defense.

[0084] The various embodiments in the present application are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other. Each embodiment focuses on the differences between it and other embodiments.

[0085] The above describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the specific order or continuous order shown in the process depicted in the drawings is not necessarily required to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0086] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A network attack injection semi-physical simulation platform for drone clusters, characterized by: The semi-physical simulation platform includes: a main control computer, a three-dimensional simulation system, a UAV cluster autonomous flight system and a network attack injection system; The main control computer is used to display the status information of the drone cluster and issue control instructions; The three-dimensional simulation system includes a simulation model of the drone cluster, an environment model and a sensor model to construct a simulation environment of the drone cluster; The drone cluster autonomous flight system is used to receive control instructions from the main control computer and sensor data sent from the three-dimensional simulation system, and extract and estimate the current state information and environmental information of the drone cluster according to the sensor data, and then obtain the control instructions of the drone cluster at the next moment in the simulation environment according to the current state information and environmental information of the drone cluster and the control instructions, and send the control instructions at the next moment to the three-dimensional simulation system; The network attack injection system is used to collect and pre-process data related to the drone cluster, analyze potential vulnerabilities based on the pre-processed data, and then conduct a network security attack on the drone cluster based on the potential vulnerabilities to verify the ability of the drone cluster to resist network security attacks.

2. The network attack injection semi-physical simulation platform for drone clusters according to claim 1 is characterized in that: The main control computer, the three-dimensional simulation system, the drone cluster autonomous flight system and the network attack injection system communicate with each other through a public API interface.

3. The network attack injection semi-physical simulation platform for drone clusters according to claim 1 is characterized in that: The status information of the drone cluster includes: location information, speed information and attitude information of the drone cluster.

4. The network attack injection semi-physical simulation platform for drone clusters according to claim 1 is characterized in that: The three-dimensional simulation system is modeled using AirSim.

5. The network attack injection semi-physical simulation platform for drone clusters according to claim 1 is characterized in that: The UAV cluster autonomous flight system includes a posture estimation module and a UAV flight controller, wherein: The posture estimation module is used to receive control instructions from the main control computer and sensor data sent from the three-dimensional simulation system, and extract and estimate the current state information and environmental information of the drone cluster based on the sensor data; The UAV flight controller is used to receive the current state information and environmental information of the UAV cluster from the posture estimation module, and receive the control instructions from the main control computer, and then obtain the control instructions of the UAV cluster at the next moment in the simulation environment based on the current state information and environmental information of the UAV cluster and the control instructions, and send the control instructions at the next moment to the three-dimensional simulation system.

6. The network attack injection semi-physical simulation platform for drone clusters according to claim 1 is characterized in that: The network attack injection unit includes a data acquisition module, a data preprocessing module, a vulnerability analysis module and a network attack injection module, wherein: The data acquisition module is used to collect log data, device information, controller data, sensor data, actuator data and network topology of the drone cluster; The data preprocessing module is used to denoise and extract features from the collected data, and identify attack surface elements; The vulnerability analysis module is used to evaluate the network security weaknesses of the drone cluster based on attack surface elements, and then determine potential network attack paths; The network attack injection module injects a network attack into the three-dimensional simulation system according to the potential network attack path.

7. The network attack injection semi-physical simulation platform for drone clusters according to claim 6 is characterized in that: The network attacks include: denial of service attacks and false data injection.

8. The network attack injection semi-physical simulation platform for drone clusters according to claim 7 is characterized in that: The denial of service attack is to obtain an IP address that has not been used by the drone cluster wireless network through the collected data, use the IP address to obtain network access rights, and then send a large amount of invalid information to block the network to achieve a denial of service attack.

9. The network attack injection semi-physical simulation platform for drone clusters according to claim 7 is characterized in that: The false data injection is to extract part of the data sent between the main control computer, the three-dimensional simulation system and the UAV cluster autonomous flight system and inject false data after invading the communication network.

10. The network attack injection semi-physical simulation platform for drone clusters according to claim 9 is characterized in that: The false data includes random information or erroneous information.

Images