A False Data Injection Attack Detection and System Control Method for Unmanned Surface Vessels

Abstract

The present invention belongs to the technical field of unmanned surface vessel system control and network security, and particularly relates to a false data injection attack detection and system control method for unmanned surface vessels. The method includes: obtaining the desired trajectory signal required by the unmanned surface vessel and the attack trajectory during the transmission of the desired trajectory signal required by the unmanned surface vessel, and establishing an FDI attack model; constructing a detection architecture and a kinematic model of the unmanned surface vessel based on the kinematic characteristics of the unmanned surface vessel; designing a certified signal observer to estimate the state of the unmanned surface vessel in real time, calculate the estimation error, and detect the FDI attack; designing an attack observer and a random interference observer for estimating the attack and interference, and compensating for the influence brought by the interference through the obtained estimated values.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned surface ship system control and network security, and in particular relates to a false data injection attack detection and system control method for unmanned surface ships. Background Art

[0002] Unmanned surface vessels (USVs) can provide surveillance, reconnaissance, and environmental monitoring, greatly improving naval combat capabilities and changing the mode of maritime operations. With the widespread application of unmanned vessels in military, scientific research, commercial transportation and other fields, their network security has become increasingly important. As a key automation and autonomous navigation technology, USVs are highly dependent on their sensor data and communication signals to achieve their missions.

[0003] In the research of unmanned surface ship systems, security issues are the most popular issues, especially network security issues. For network security issues, the main research direction is to detect, estimate and compensate for FDI attacks. The purpose is to timely identify and confirm whether there is an FDI attack in the system when the system is attacked by FDI. Through the detection mechanism, the integrity and authenticity of the system data are protected to prevent attackers from interfering with the normal operation of the system through false data. Then, by estimating the attack, analyzing the characteristics, attack sources, attack paths, etc. of the attack, in order to have a deeper understanding of the attack behavior, according to the estimation results, optimize and adjust the existing defense strategy to make it more adaptable to the current security threat environment and improve the system's defense capabilities. Finally, the compensation mechanism is used to restore the normal state of the system, ensure the continuity and stability of the system, repair the tampered or damaged data, and restore the integrity and accuracy of the data. In real network systems, malicious network attacks will cause great damage when they occur, so it is more meaningful to study FDI attack detection, estimation and compensation.

[0004] When a system is attacked by FDI, the attacker's goal is to directly tamper with the sensor data. Through physical access or network attack means, the attacker directly modifies the data of the sensor or measuring device to send false information to the control system. The attacker may also sneak into the system and modify the configuration file of the device to change the data transmission method or format, thereby injecting false data. When the attacked is attacked by FDI, it may cause serious adverse consequences. For example, the attacker tampers with the data of a drone that is performing a secret military escort mission, which may cause problems with the data of the attacked target and cause the escort mission to fail.

[0005] During the movement of an unmanned surface vessel system, after being attacked by a cyber-attack, the focus is on the impact of the attack on the vessel's movement. Generally speaking, the movement trajectory of an unmanned surface vessel system will change uncontrollably after being attacked and deviate from the preset movement trajectory. Based on this, the present invention proposes an FDI attack detection and estimation technology for unmanned surface vessels, which can achieve FDI attack detection and estimation of the unmanned surface vessel system by adjusting parameter settings. Summary of the Invention

[0006] In order to overcome the problems in the prior art, the present invention proposes a false data injection attack detection and system control method for unmanned surface vessels.

[0007] The technical solution of the present invention to solve the above technical problems is as follows:

[0008] The present invention provides the following steps:

[0009] Step 100: Obtain the desired trajectory signal required by the unmanned surface vessel and the attack trajectory during the transmission of the desired trajectory signal required by the unmanned surface vessel, and establish an FDI attack model;

[0010] Step 200: Based on the kinematic characteristics of the unmanned surface vessel, construct a detection architecture and a kinematic model of the unmanned surface vessel; design a certification signal observer to estimate the state of the unmanned surface vessel in real time, calculate the estimation error, and detect FDI attacks;

[0011] Step 300: Design an attack observer and a random interference observer to estimate the attack and interference, and compensate for the influence brought by the interference through the obtained estimated values.

[0012] Further, in the step 100, the FDI attack model includes a bias attack model, and the bias attack model is:

[0013]

[0014] In the above formula, represents a bias attack, γ a represents the instantaneous component participating in the bias attack, γ ∞ represents the expected deviation, represents a known normal number.

[0015] Further, in the step 100, the FDI attack model includes a periodic attack model, and the periodic attack model is:

[0016]

[0017] Where χ a describes the state of the external system, V a and W aA known matrix representing the appropriate dimension of the system; Denote the derivative of the periodic attack, ω a Denote the periodic attack with unknown amplitude.

[0018] Furthermore, in the step 200, based on the kinematic characteristics of the unmanned surface vessel, a detection architecture and an unmanned surface vessel kinematic model are constructed:

[0019]

[0020] In the above formula, Consists of the actual position (x, y) and the heading angle θ; J(ψ) represents the rotation matrix; m represents the ship mass; v v Denote the motion state vector, which are the longitudinal velocity, lateral velocity, and yaw angular velocity respectively; τ v Denote the external force input vector, including the control force and environmental disturbance; d represents the authentication signal described externally through the unmanned surface vessel system; v represents the actual speed of the USV; D(v) represents the nonlinear damping matrix; the vector η = [x, y, ψ] T Consists of the position (x, y) and the heading ψ. In the fixed coordinate system of the earth, v0 = [u, v s , r] T Denote the surge velocity u, sway velocity v of the ship in the body-fixed coordinate system s And the yaw angular velocity r, then Denote the random ocean disturbance vector that changes with time, following a first-order Markov random process:

[0021]

[0022] In the above formula, τ d (t) represents the random ocean disturbance vector that changes with time; Denote the derivative of the disturbance vector; Ψ represents the amplitude matrix; ξ(t) represents the Gaussian white noise vector; T d Denote the time constant positive matrix; V v And W v Denote matrices of appropriate dimensions; ω v Denote the state of the external unmanned ship system;

[0023] Among them, J(ψ) represents the rotation matrix. For the motion dynamics of surface ships, M is defined as the inertia matrix including hydrodynamics, and D(v) is defined as the nonlinear damping matrix, specifically as follows:

[0024]

[0025] Among them, I Z Denote the moment of inertia about the fixed AZ axis of the ship; x Grepresents the distance between the center of gravity of the ship and the origin of the ship's fixed coordinate system A along the AX axis; and represents the zero-frequency added mass caused by pitch, roll and yaw accelerations; u0 represents the cruising speed, u0≈0 in DP operation and u0>0 in the case of forward ship motion; X u 、Y v 、N v 、Y r and N r represent the hydrodynamic derivatives due to laminar skin friction and wave drift damping.

[0026] Furthermore, in the step 200, a certified signal observer is designed to estimate the state of the unmanned surface vehicle in real time and calculate the estimation error, including:

[0027]

[0028] where d represents the certified signal described externally through the unmanned ship system represents the estimation of d; represents the estimation error; z v represents the signal of the observer; L v represents the gain matrix of the observer; v1 represents the auxiliary variable; represents the estimated value of the observer signal; W v represents a matrix of appropriate dimension.

[0029] Furthermore, in the step 200, detecting FDI attacks includes:

[0030] Converging the estimation error to a bounded set and designing an attack detection logic using the bounded set;

[0031] Within the proposed detection mechanism, the actual speed v is used when calculating the estimation error , where If FDI does not occur, the dynamics of the estimation error are expressed as:

[0032]

[0033] Alarm represents the attack alarm, 1 represents an attack occurring, and 0 represents no attack occurring;

[0034]

[0035] where represents the estimated error; P' represents the average power of the received signal; P represents the average power of the attack signal.

[0036] Further, in the step 300, the attack observer includes a deviation attack observer, and the deviation attack observer is:

[0037]

[0038] In the above formula, represents the estimated value of F γa ; z γa represents the auxiliary state variable of the observer; L γa represents the deviation gain selected by calculation; v represents the actual speed; m represents the ship mass; τ represents the external force input vector; D(v) is the nonlinear damping matrix; represents the estimated value of the ocean random disturbance vector; represents the estimated value of F χa .

[0039] Further, in the step 300, the attack observer includes a periodic attack observer, and the periodic attack observer is:

[0040]

[0041] In the above formula, θ represents a positive constant obtained from the gain matrix; represents the estimated value of observing the periodic attack; represents the auxiliary state variable of the observer; represents the selectable periodic gain; represents; W a represents a known matrix of an appropriate dimension of the system.

[0042] Further, in the step 300, the random disturbance observer is:

[0043]

[0044] dq(t) = -(T -1 + K)(q(t) + K0v)d(t) - Kτdt + K(C(v)v + Dv)dt;

[0045] In the above formula, represents the disturbance estimation vector; q(t) represents the auxiliary vector of the observer; v represents the actual speed; K0 represents and K represents the design matrix satisfying K = K0M -1 ; τ represents the external force input vector; C(v) represents the Coriolis force and centrifugal force matrix; Dv represents the nonlinear damping matrix.

[0046] Further, in the step 300, the influence brought by interference is compensated by the obtained estimated value, and specifically, a feedforward compensation control method based on a stochastic disturbance observer is adopted. The feedforward compensation control method based on a stochastic disturbance observer includes:

[0047] The attack estimated value is fed forward to the required desired trajectory, and the desired trajectory is adjusted as follows:

[0048]

[0049] The attack signal superimposed on the desired trajectory is cancelled; wherein, the attack estimate γ to be compensated a and χ a are given by the following formula: and

[0050] The disturbance estimated value is fed forward to the control input, and the composite controller is modified to compensate for the disturbance; the modified composite controller is:

[0051]

[0052] wherein, p′ da = p′ d + χ a + γ a , e′ p = p′ da - p,

[0053] In the above formula, F′ represents the equivalent force input of the attack; m represents the mass of the ship; P τ ′ d represents the adjusted desired trajectory; represents the estimate of the compensated deviation attack; represents the estimate of the compensated periodic attack; represents the estimated value of the influence of the deviation attack on the controller; K v represents the gain of the controller; K p represents the gain of the controller; represents a known normal number; F′ represents the equivalent force input of the attack; m represents the mass of the ship; e′ p represents the position error; e v ′ represents the velocity error; represents the estimated value of the signal; τ′ represents the derivative of the external force input; e η ′ represents the position vector error; e′ ω represents the signal error; represents the estimate of the external environmental disturbance; K ω and K η represent the design matrix; γ a represents the deviation attack; χa Represents a periodic attack with an unknown phase.

[0054] Compared with the prior art, the present invention has the following technical effects:

[0055] (1) By establishing an FDI attack model, designing a detection architecture, a kinematic model of the unmanned surface vessel, and a certified signal observer, the present invention can detect and identify in real time the false data injection attack (FDI attack) that the unmanned surface vessel may suffer during data transmission, which helps to timely discover and respond to potential security threats, thereby improving the overall security of the unmanned surface vessel.

[0056] (2) The present invention not only considers the FDI attack, but also considers the influence of external environmental interference on the unmanned surface vessel. By designing an attack observer and a stochastic disturbance observer, the influence of the FDI attack and external environmental interference on the control system can be accurately estimated.

[0057] (3) Through the feedforward compensation control (DOBC) method based on the stochastic disturbance observer, the present invention can use the disturbance value estimated by the observer to correct the control input, thereby compensating for the influence brought by the disturbance. This helps to improve the accuracy and efficiency of the control system, enabling the unmanned surface vessel to more accurately track the desired trajectory and achieve a more stable navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0059] Figure 1 is a schematic flow chart of the present invention;

[0060] Figure 2 is a system diagram of the application of the attack detection and system control established by the present invention in the USV. DETAILED DESCRIPTION OF THE INVENTION

[0061] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will, in conjunction with the drawings and preferred embodiments, detail the specific implementation manners, structures, features, and effects of the technical solutions proposed according to the present invention. The specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form. Unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0062] In one embodiment of the present invention, referring to Figure 1 - Figure 2 , the following steps are provided:

[0063] Step 100: Obtain the desired trajectory signal required by the unmanned surface vessel, and the attack trajectory during the transmission of the desired trajectory signal required by the unmanned surface vessel, and establish an FDI attack model;

[0064] Step 200: Based on the kinematic characteristics of the unmanned surface vessel, construct a detection architecture and a kinematic model of the unmanned surface vessel; design a certification signal observer to estimate the state of the unmanned surface vessel in real time, calculate the estimation error, and detect FDI attacks;

[0065] Step 300: Design an attack observer and a random interference observer to estimate attacks and interferences, and compensate for the influence brought by interferences through the obtained estimated values.

[0066] The above steps will be elaborated in detail below:

[0067] Step 100: Obtain the desired trajectory signal required by the unmanned surface vessel, and the attack trajectory during the transmission of the desired trajectory signal required by the unmanned surface vessel, and establish an FDI attack model.

[0068] Consider a scenario where the desired trajectory signal transmitted by a wireless data link is subject to an FDI attack. The attacker injects false data during the transmission of the required trajectory signal, and the receiver can decode the damaged desired trajectory signal after despreading and demodulating the received signal.

[0069] Step 110: The ground control system transmits the desired trajectory signal required by the unmanned surface vessel through a wireless data link.

[0070] The wireless data link is mainly responsible for completing the transmission and processing of the signals of the ground control system to the unmanned surface vessel. Therefore, the remote control of the system only depends on the uplink of the wireless data. The uplink of the wireless data link is the target of the attacker.

[0071] Step 120: Define the attack trajectory based on the deviation attack and the periodic attack suffered by the desired trajectory signal during the transmission process.

[0072] According to the analysis in Step 110, in the presence of various forms of incorrect data, the desired trajectory may be affected by different attack models. The focus of this embodiment is on the deviation and periodic attacks during the transmission of the required ballistic signal.

[0073] The transmitted signal S(t) in the wireless data link is expressed as:

[0074]

[0075] In the above formula, p dDenote the desired trajectory signal, P a Denote the attacked desired trajectory signal, p d (t) denotes the desired trajectory signal code; n(t) denotes the pseudo-random spreading code; t denotes time.

[0076] The FDI signal and the desired trajectory signal have the same structure. To successfully initiate FDI, the following attack signal S can be generated a (t):

[0077]

[0078] where, P a Denote the power of the FDI signal, p da (t) denotes the code of the error data.

[0079] In the presence of different forms of error data, the desired trajectory may be affected by various attack models. The attack trajectory is defined as:

[0080]

[0081] In the above formula, Denote the interference brought by the external environment, γ a Denote the bias attack, including the steady-state component, the desired deviation component and the transient component, χ a Denote the periodic attack with unknown phase, periodic attack with unknown phase. For the transient component, the opponent using a low-pass filter may cause the data corruption to slowly converge to the steady-state value.

[0082] Step 130: Based on the bias attack, establish a bias attack model.

[0083] Study the false data injection scenario, whose goal is to inject a constant bias into the system without being detected. An attack model with progressive convergence is adopted, and the attack value slowly converging to the desired bias value can also ensure the concealment of the attack.

[0084]

[0085] In the above formula, γ a Denote the instantaneous component participating in the bias attack, γ ∞ Denote the expected deviation, Denote the known normal number. The attack model described under the formula is used to achieve the purpose of injecting a bias attack. In addition, the attack value slowly converging to the desired bias value can ensure the confidentiality of the attack.

[0086] Step 140: Based on the periodic attack, establish a periodic attack model.

[0087] In addition to the bias attack, periodic attacks are also considered.

[0088]

[0089] where χ a describes the state of the external system, V a and W a are known matrices of appropriate dimensions of the system; represents the derivative of the periodic attack, ω a represents a periodic attack with an unknown amplitude.

[0090] Step 200: Based on the kinematic characteristics of the unmanned surface vehicle, construct a detection architecture and a kinematic model of the unmanned surface vehicle; design a certification signal observer to estimate the state of the unmanned surface vehicle in real time, calculate the estimation error, and detect FDI attacks.

[0091] As an example, this step may include the following steps:

[0092] Step 210: Construct a detection architecture and a kinematic model of a three-degree-of-freedom unmanned surface vehicle;

[0093]

[0094] In the above formula, is composed of the actual position (x, y) and the heading angle θ; J(ψ) represents the rotation matrix; m represents the ship mass; v v represents the motion state vector, which are the longitudinal velocity, lateral velocity, and yaw angular velocity respectively; τ v represents the external force input vector, including the control force and environmental disturbance; d represents the certification signal described externally by the unmanned surface vehicle system; v represents the actual speed of the USV; D(v) represents the nonlinear damping matrix; the vector η = [x, y, ψ] T is composed of the position (x, y) and the heading ψ. In the fixed coordinate system of the earth, v0 = [u, v s , r] T represents the surge velocity u, sway velocity v s and yaw angular velocity r of the ship in the body-fixed coordinate system, then represents the random ocean disturbance vector that changes with time and follows a first-order Markov random process:

[0095]

[0096] In the above formula, τ d (t) represents the random ocean disturbance vector that changes with time; represents the derivative of the disturbance vector; Ψ represents the amplitude matrix; ξ(t) represents the Gaussian white noise vector; T d represents the time constant positive matrix; V v and Wv represents a matrix of appropriate dimension; ω v represents the state of the external unmanned ship system.

[0097] Among them, J(ψ) represents the rotation matrix. For the motion dynamics of surface ships, M is defined as the inertia matrix including hydrodynamics, which is a symmetric positive definite matrix, and D(v) is defined as the nonlinear damping matrix, specifically as follows:

[0098]

[0099] Among them, I Z represents the moment of inertia about the fixed AZ axis of the ship; x G represents the distance between the center of gravity of the ship and the origin of the ship-fixed coordinate system A along the AX axis; and represent the zero-frequency added mass caused by pitch, roll and yaw accelerations; u0 represents the cruising speed, u0≈0 in DP operation and u0>0 in the case of forward ship motion; X u 、Y v 、N v 、Y r and N r represent the hydrodynamic derivatives caused by laminar skin friction and wave drift damping.

[0100] In practice, the motion dynamics of ships are affected by ship loads, speeds and maneuvers. The model uncertainties ΔM and ΔD are regarded as model uncertainties respectively, and the model uncertainties of ship motion dynamics are described by the model uncertainties of M + ΔM and D(v) + ΔD.

[0101] Step 220: Design a certification signal observer for real-time estimating the state of the unmanned surface ship and calculating the estimated value:

[0102]

[0103] Among them, d represents the certification signal described externally through the unmanned ship system represents the estimate of d; represents the estimation error; z v represents the signal of the observer; L v represents the gain matrix of the observer; v1 represents the auxiliary variable; represents the estimated value of the observer signal; W v represents a matrix of appropriate dimension.

[0104] Step 230: Calculate the estimation error according to the obtained estimated value and the set true value.

[0105] Step 240: Converge the estimation error to a bounded set and design the attack detection logic using the bounded set.

[0106] Within the proposed detection mechanism, the actual speed v is used when calculating the estimation error where, If FDI does not occur, the dynamics of the estimation error can be expressed as:

[0107]

[0108] Alarm represents the attack alarm, 1 indicates an attack occurs, and 0 indicates no attack occurs;

[0109]

[0110] where, represents the estimated error; P' represents the average power of the received signal; P represents the average power of the attack signal.

[0111] Step 300: Design an attack observer and a random disturbance observer for estimating attacks and disturbances, and compensate for the influence brought by the disturbances through the obtained estimated values.

[0112] As an example, this step may include the following steps:

[0113] Step 310: Design an attack observer to estimate the impact of FDI on the control system to compensate for and reduce the impact of false data on the desired trajectory; wherein, the attack observer includes a bias attack observer and a periodic attack observer.

[0114] The bias attack observer is a controller adapted along the bias and is designed to estimate the impact of the bias attack on the controller. The design of the bias attack observer is as follows:

[0115]

[0116] In the above formula, represents the estimated value of F γa ; z γa represents the auxiliary state variable of the observer; L γa represents the bias gain selected through calculation; v represents the actual speed; m represents the ship mass; τ represents the external force input vector; D(v) is the nonlinear damping matrix; represents the estimated value of the ocean random disturbance vector; represents the estimated value of F χa .

[0117] The periodic attack observer aims to evaluate the impact of the periodic attack on the controller. The design of the periodic attack observer is as follows:

[0118]

[0119] In the above formula, θ represents a positive constant obtained from the gain matrix; represents the estimated value of the observation period attack; represents the auxiliary state variable of the observer; represents the selectable periodic gain; represents; W a represents a known matrix of appropriate dimension of the system.

[0120] Step 320: The random disturbance observer is a non-linear observer that uses the nominal ship model; by combining the control input, control output deviated due to the disturbance, and the information provided by the stochastic controlled object, the unknown random disturbance is eliminated without online estimating the disturbance.

[0121] The random disturbance observer is designed to adopt the following form:

[0122]

[0123] dq(t) = -(T -1 + K)(q(t)+K0v)d(t)-Kτdt+K(C(v)v+Dv)dt;

[0124] In the above formula, represents the disturbance estimation vector; q(t) represents the auxiliary vector of the observer; v represents the actual speed; K0 represents and K represents the design matrix satisfying K = K0M -1 ; τ represents the external force input vector; C(v) represents the Coriolis force and centrifugal force matrix; Dv represents the non-linear damping matrix.

[0125] Define the vector of the observer estimation error as:

[0126]

[0127] In the above formula, represents the disturbance estimation error vector; τ d represents the disturbance vector; represents the disturbance estimation vector.

[0128] According to the first-order Markov random process and the above-designed random disturbance observer, the disturbance estimation error can be obtained as follows:

[0129]

[0130] In the above formula, represents the perturbation estimation error; represents the perturbation vector; Indicates disturbance estimation.

[0131] Step 330: To compensate for the influence of disturbances, a feedforward compensation control method based on a stochastic disturbance observer (DOBC, Disturbance Observer-Based Control) is proposed.

[0132] The attack estimate is fed forward to the desired trajectory, and the attack signal superimposed on the desired trajectory can be cancelled. Among them, by adjusting the desired trajectory as follows:

[0133]

[0134] In the above formula, P τ ′ d Indicates the adjusted desired trajectory; Indicates the estimate of compensating for the bias attack; Indicates the estimate of compensating for the periodic attack; K v Indicates the gain of the controller; K p Indicates the gain of the controller; τ′ represents the external force input; Indicates a known normal number.

[0135] The disturbance estimate is fed forward to the control input, and the composite controller is modified to compensate for the disturbance. The modified composite controller can be constructed as:

[0136]

[0137] Among them, p′ da = p′ d + χ a + γ a , e′ p = p′ da - p,

[0138] In the above formula, F′ represents the equivalent force input of the attack; m represents the ship mass; P τ ′ d Indicates the adjusted desired trajectory; Indicates the estimate of compensating for the bias attack; Indicates the estimate of compensating for the periodic attack; Indicates the estimated value of the influence of the bias attack on the controller; K v Indicates the gain of the controller; K p Indicates the gain of the controller; Indicates a known normal number; F′ represents the equivalent force input of the attack; m represents the ship mass; e′ p Indicates the position error; e v ′ represents the velocity error; represents the estimated value of the signal; τ′ represents the derivative of the external force input; e η ′ represents the position vector error; e′ ω represents the signal error; represents the estimation of the external environmental disturbance; K ω and K η represents the design matrix; γ a represents the bias attack, χ a represents the periodic attack with an unknown phase.

[0139] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A false data injection attack detection and system control method for unmanned surface ships, characterized in that: The following steps are involved: Step 100: obtaining the desired trajectory signal required by the unmanned surface ship and the attack trajectory of the desired trajectory signal required by the unmanned surface ship during transmission, and establishing an FDI attack model; Step 200: constructing a detection framework and a kinematic model of the unmanned surface ship based on the kinematic characteristics of the unmanned surface ship; Design an authenticated signal observer to estimate the state of unmanned surface vessels in real time and calculate the estimation error, and detect FDI attacks; Step 300: Design an attack observer and a random interference observer to estimate the attack and interference, and compensate for the influence of the interference by the obtained estimated values; In step 300, the random interference observer is: ； ； In the above formula, represents the interference estimation vector; It represents the auxiliary vector of the observer; Indicates actual speed; Representation and It means that the design matrix satisfies ; represents the external force input vector; represents the Coriolis and centrifugal force matrices; represents the nonlinear damping matrix; In step 300, the influence of the disturbance is compensated by the obtained estimated value, specifically using a feedforward compensation control method based on a random disturbance observer. The feedforward compensation control method based on a random disturbance observer includes: The attack estimate is fed forward to the desired desired trajectory by adjusting the desired trajectory as follows: ； The attack signal superimposed on the desired trajectory is cancelled; where the attack estimate to be compensated and Given by: and ; The disturbance estimate is fed forward to the control input and the composite controller is modified to compensate for the disturbance; the modified composite controller is: in, , , ; In the above formula, represents the equivalent force input of the attack; Indicates the quality of the ship; represents the adjusted expected trajectory; represents an estimate of the compensation bias attack; represents the estimate of the compensation cycle attack; represents the estimated value of the impact of the deviation attack on the controller; represents the gain of the controller; represents the gain of the controller; represents a known normal number; represents the equivalent force input of the attack; Indicates the quality of the ship; Indicates position error; Indicates speed error; represents the estimated value of the signal; represents the derivative of the external force input; represents the position vector error; Indicates signal error; represents an estimate of the external environmental disturbance; and represents the design matrix; indicates deviation attack; Indicates a periodic attack of unknown phase.

2. According to claim 1, a false data injection attack detection and system control method for unmanned surface ships is characterized in that: In step 100, the FDI attack model includes a deviation attack model, and the deviation attack model is: ； In the above formula, Indicates bias attack, represents the instantaneous component participating in the deviation attack, represents the expected deviation, Represents a known normal number.

3. A false data injection attack detection and system control method for unmanned surface ships according to claim 2, characterized in that: In step 100, the FDI attack model includes a periodic attack model, and the periodic attack model is: ； in, Describe the state of external systems, and a known matrix of appropriate dimensions representing the system; represents the derivative of the periodic attack, Indicates a periodic attack of unknown magnitude.

4. A false data injection attack detection and system control method for unmanned surface ships according to claim 1 or 3, characterized in that: In step 200, based on the kinematic characteristics of the unmanned surface ship, a detection framework and a kinematic model of the unmanned surface ship are constructed: ； In the above formula, By actual location and heading angle composition; represents the rotation matrix; Indicates the quality of the ship; Represents the motion state vector, which are longitudinal velocity, lateral velocity and heading angular velocity; Represents the external force input vector, including control force and environmental disturbance; Represents an authentication signal externally described by an unmanned surface vessel system; Indicates the actual speed of the USV; represents the nonlinear damping matrix; the vector By Location and heading In the Earth's fixed coordinate system, Indicates the ship's sway speed in the fixed fuselage coordinate system , swing speed and yaw rate ,but represents a random ocean disturbance vector that varies with time and follows a first-order Markov random process: ； ； In the above formula, represents the time-varying random ocean disturbance vector; Expressed as the derivative of the perturbation vector; represents the amplitude matrix; represents a Gaussian white noise vector; represents the time constant positive matrix; and represents a matrix of appropriate dimension; Indicates the status of external unmanned ship systems; in, Represents the rotation matrix. For the dynamics of surface ship motion, is defined as the inertia matrix that includes fluid dynamics, and It is defined as a nonlinear damping matrix as follows: ； ； ； in, represents the moment of inertia about the fixed AZ axis of the ship; It represents the distance between the center of gravity of the ship and the origin of the ship's fixed coordinate system A along the AX axis; , and represents the zero-frequency added mass caused by pitch, roll, and yaw accelerations; Indicates the cruising speed, In DP operation it is , when the ship is moving forward ; , , , and represents the hydrodynamic derivative due to laminar surface friction and wave drift damping.

5. A false data injection attack detection and system control method for unmanned surface ships according to claim 4, characterized in that: In step 200, an authentication signal observer is designed to estimate the state of the unmanned surface vessel in real time and calculate the estimation error, including: ； in, Indicates the authentication signal described from the outside by the unmanned ship system express estimates; represents the estimation error; represents the signal of the observer; represents the gain matrix of the observer; Represents auxiliary variables; represents the estimated value of the observer signal; represents a matrix of appropriate dimension.

6. A false data injection attack detection and system control method for unmanned surface ships according to claim 5, characterized in that: In step 200, detecting an FDI attack includes: Converge the estimation error to a bounded set and use the bounded set to design attack detection logic; In the proposed detection mechanism, the estimated error is calculated Use the actual speed of the USV ,in, ; If FDI does not occur, the estimation error The dynamic representation of is: ； Alarm indicates an attack alarm, 1 indicates an attack has occurred, and 0 indicates no attack has occurred; in, represents the error of the estimate; Indicates the average power of the received signal; Indicates the average power of the attack signal.

7. A false data injection attack detection and system control method for unmanned surface ships according to claim 1 or 6, characterized in that: In step 300, the attack observer includes a deviation attack observer, and the deviation attack observer is: ； In the above formula, express An estimated value of represents the auxiliary state variables of the observer; represents the bias gain selected by calculation; Indicates the actual speed of the USV; Indicates the quality of the ship; represents the external force input vector; is the nonlinear damping matrix; represents the estimated value of the ocean random disturbance vector; express The estimated value of .

8. The method for detecting and controlling false data injection attacks on unmanned surface vessels according to claim 7, characterized in that: In step 300, the attack observer includes a periodic attack observer, and the periodic attack observer is: ； In the above formula, represents a positive constant obtained from the gain matrix; represents the estimated value of the attack during the observation period; represents the auxiliary state variables of the observer; Indicates the cycle gain that can be selected; express; A known matrix of appropriate dimensions representing the system.

Images