Anti-deception attack adaptive sliding mode control method and system for unmanned ship and medium

Abstract

The application relates to the technical field of unmanned ship safety control, and discloses an anti-deception attack adaptive sliding mode control method and system for an unmanned ship and a medium, the method comprising the following steps: S1, constructing a T-S fuzzy model of the unmanned ship under deception attack and external disturbance; S2, designing a double adaptive sliding mode observer based on the T-S fuzzy model of the unmanned ship constructed in S1; S3, designing a sliding mode controller based on S1 and S2; S4, forming a closed-loop control system for sliding mode control based on S1-S3: integrating the T-S fuzzy model of the unmanned ship constructed in S1, the double adaptive sliding mode observer designed in S2 and the sliding mode controller designed in S3 to form a closed-loop control system and realize sliding mode control. The application can significantly improve the dynamic inhibition capability of composite threats, realize accurate compensation for deception attack signals and marine external disturbances, realize high-precision zero-vibration control under the coexistence of deception attack and external disturbance, and provide an autonomous control solution with safety, robustness and real-time performance for the unmanned ship.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of unmanned ship safety control, and particularly relates to an anti-deception attack adaptive sliding mode control method and system for an unmanned ship and a medium. BACKGROUND

[0002] Unmanned ships play a key role in ocean exploration, environmental monitoring and offshore engineering due to their environmental adaptability and high efficiency. However, their operation faces two serious challenges. First, physical layer disturbance: dynamic disturbances such as wind and waves in the ocean environment cause the system to exhibit strong nonlinearity, making it difficult for traditional linear controllers to achieve stable trajectory tracking. Second, network layer attack: remote control relies on wireless communication networks, which are vulnerable to deception attacks. Attackers can tamper with control instructions by injecting false signals, causing system performance degradation or even instability.

[0003] Traditional sliding mode control (SMC) is robust to disturbances, but requires accurate knowledge of the disturbance boundary. However, the upper bound of actual ocean disturbances is unknown and time-varying. The fixed switching gain of SMC can induce high-frequency chattering, exacerbating actuator wear. Existing network security control research assumes fixed attack patterns and does not consider real-time adjustment of attack strategies. Existing observers cannot simultaneously estimate attack signals and disturbance boundaries, resulting in insufficient control accuracy under combined threats. T-S fuzzy modeling can linearize nonlinear systems, but lacks a coordinated control framework for network attacks, making it difficult to handle the coupling effects of deception attacks and disturbances.

[0004] Existing methods cannot achieve high-precision stable control of unmanned ships in the presence of unknown deception attacks and time-varying disturbances. The root cause is that the dynamic coupling of attack signals and disturbances has not been decoupled, there is a lack of online joint estimation mechanism for attack weights and disturbance boundaries, and chattering suppression and stability guarantee are not unified in a strict mathematical framework (input-state stability theory). SUMMARY

[0005] The technical problem to be solved by the present application is to overcome the shortcomings of the prior art. The present application provides an anti-deception attack adaptive sliding mode control method, system and medium for an unmanned ship. By online joint estimation of deception attack signals ε∇(x(t)) and disturbance boundaries α, β, real-time compensation of coupled disturbances, dynamic decoupling of combined threats, design of adaptive switching law for dynamic adjustment of sliding mode gain, elimination of actuator wear caused by fixed gain, suppression of chattering, and proof of input-state stability (ISS) of the closed-loop control system based on Lyapunov theory and linear matrix inequality (LMI), strict stability is guaranteed.

[0006] To achieve the above purpose, the first aspect of the present application provides an anti-deception attack adaptive sliding mode control method for an unmanned ship, comprising the following steps:

[0007] S1, constructing a T-S fuzzy model of an unmanned ship under deception attacks and external disturbances:

[0008] The deception attacks and external disturbances are coupled in the T-S fuzzy model of the unmanned ship ;

[0009] S2, designing a double adaptive sliding mode observer based on the T-S fuzzy model of the unmanned ship constructed in S1:

[0010] The double adaptive sliding mode observer reconstructs state estimation values using measurable outputs and estimates the upper bounds of deception attacks and external disturbances online;

[0011] At least including designing a double adaptive law, which is used to complete the joint estimation of the upper bounds of deception attacks and external disturbances , and dynamically adjusting adaptive rates and adaptive rates through the double adaptive law, wherein the adaptive rates are used to estimate the upper bounds of external disturbances, and the adaptive rates are used to estimate the upper bounds of deception attacks; and,

[0012] At least including designing a robust term , which uses the adaptive rates and the adaptive rates to generate a compensation signal through a sign function , which is used to actively offset the effects of deception attacks and external disturbances ;

[0013] S3, designing a sliding mode controller based on S1 and S2:

[0014] Based on the T-S fuzzy model of the unmanned ship constructed in S1, the state estimation values , adaptive rates and adaptive rates provided by the double adaptive sliding mode observer in S2 are used to design a sliding mode controller;

[0015] S4, forming a closed-loop control system for sliding mode control based on S1-S3:

[0016] Integrating the T-S fuzzy model of the unmanned ship constructed in S1, the double adaptive sliding mode observer designed in S2 and the sliding mode controller designed in S3, a closed-loop control system is formed to realize the sliding mode control of the T-S fuzzy model of the unmanned ship in S1.​​

[0017] Optionally, the unmanned ship T-S fuzzy model constructed in S1 is represented as:

[0018] ,

[0019] ,

[0020] (1);

[0021] wherein, represents the system full state of the unmanned ship model, the constant i = 1, 2, 3, 4, represents the performance output, represents the actual output, t is a time variable, is a linear velocity or angular velocity or position vector matrix, is a control rate, is a deception attack weighting matrix, is a deception attack mathematical model, is an external disturbance, is a premise variable, is a fuzzy set, is a damping ratio inertia matrix, is a mooring ratio inertia matrix, is a measurement output constant matrix, is a control output constant matrix.

[0022] Optionally, the double adaptive sliding mode observer in S2 is designed, comprising: the double adaptive sliding mode observer adopts a T-S fuzzy model and injects an output estimation error term to reconstruct the unmeasurable system state , the dynamic equation of the double adaptive sliding mode observer is represented as:

[0023] ,

[0024] ,

[0025] (2);

[0026] wherein, represents the system full state of the sliding mode observer, is an estimated value of , is an estimated value of , is an estimated value of , is an estimated value of , is an observer gain, is a robust term.

[0027] Optionally, S2 at least includes designing a double adaptive law, denoted as:

[0028] (4) ;

[0029] (5) ;

[0030] wherein, denotes a disturbance boundary estimation update law, denotes an attack boundary estimation update law, is an anti-network attack adaptive rate constant, is an anti-disturbance adaptive rate constant, is an upper limit of network attack, denotes the transpose of an observer error, denotes the transpose of a measurement output constant matrix, denotes the actual output norm.

[0031] Optionally, S2 at least further includes designing a robust term , the robust term utilizes the adaptive rate and the adaptive rate to produce a compensation signal through a sign function for actively counteracting the effects of a deception attack and an external disturbance ;

[0032] the robust term is denoted as:

[0033] (3) ;

[0034] wherein, are estimated values of respectively, is an anti-disturbance adaptive rate, is an anti-network attack adaptive rate, is a robust term constant, is an observer error weight matrix, is an observer error.

[0035] Optionally, S2 at least further includes designing an error dynamic analysis, specifically:

[0036] is defined as:

[0037] In combination with formula (1) and formula (2), the error dynamic equation is denoted as: ​

[0038] (6);

[0039] (7);

[0040] wherein, represents an observer error update rate, is an observer error, is an observer error residual term.

[0041] Optionally, the sliding mode controller in S3 includes designing a sliding mode surface, denoted as:

[0042] (8);

[0043] wherein, is a sliding mode surface function, is a sliding mode matrix, is a controller gain, is a controller integral variable;

[0044] Combining formula (2) and formula (8), the sliding mode surface function rate of change is obtained , denoted as:

[0045] (9);

[0046] The time-varying weighting matrix is defined as , and the equivalent control rate is obtained, denoted as:

[0047] (10);

[0048] Substituting formula (10) into formula (2), the sliding mode dynamics equation is obtained, denoted as:

[0049] (11).

[0050] Optionally, the sliding mode controller in S3 further includes:

[0051] The Lyapunov function is constructed, denoted as:

[0052] (38);

[0053] wherein, represents the transpose of ;

[0054] The control law for realizing the sliding mode region reachability construction under the spoofing attack is constructed, denoted as:

[0055] (39);

[0056] wherein, is the controller combined gain, is represented as a time-varying weighting matrix inverse;

[0057] Definition ;

[0058] wherein, is the anti-disturbance adaptive rate, is the anti-network attack adaptive rate, is a robust term constant, is a reaching condition constant;

[0059] Derivation of Lyapunov function , obtains:

[0060] (40);

[0061] wherein, represents the derivative of the Lyapunov function, represents the rate of change of the sliding surface function, which indicates that under the condition of deception attack, the closed-loop system is driven to the sliding surface by the sliding mode control rate formula (39) .

[0062] To achieve the above purpose, the second aspect of the present application provides an anti-deception attack adaptive sliding mode control system of an unmanned ship, the control system comprising:

[0063] A first design unit constructs a T-S fuzzy model of an unmanned ship under deception attack and external disturbance: the T-S fuzzy model of the unmanned ship is coupled with deception attack and external disturbance ;

[0064] A second design unit designs a double adaptive sliding mode observer based on the constructed T-S fuzzy model of the unmanned ship: the double adaptive sliding mode observer reconstructs the state estimation value using the measurable output , and estimates the upper bound of the deception attack and external disturbance online;

[0065] At least comprising a double adaptive law, the double adaptive law is used to complete the joint estimation of the upper bound of the deception attack and external disturbance , through the double adaptive law dynamically adjusting the adaptive rate and the adaptive rate , wherein the adaptive rate is used to estimate the upper bound of the external disturbance, and the adaptive rate is used to estimate the upper bound of the deception attack; and,

[0066] at least including a design robust term , the robust term using the adaptive rate and the adaptive rate , a compensation signal is produced by a sign function for actively canceling the influence of spoofing attacks and external disturbances ;

[0067] A third design unit designs a sliding mode controller based on the first design unit and the second design unit: based on the constructed T-S fuzzy model of the unmanned ship, the state estimation value provided by the double adaptive sliding mode observer in S2 , the adaptive rate and the adaptive rate , a sliding mode controller is designed.

[0068] A closed-loop control unit forms a closed-loop control system for sliding mode control based on the first design unit, the second design unit and the third design unit: the T-S fuzzy model of the unmanned ship constructed by the first design unit, the double adaptive sliding mode observer designed by the second design unit and the sliding mode controller designed by the third design unit are integrated to form a closed-loop control system, and the sliding mode control of the T-S fuzzy model of the unmanned ship is realized.

[0069] To achieve the above-mentioned purposes, the third aspect of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the anti-spoofing attack adaptive sliding mode control method of the unmanned ship as described above.

[0070] Compared with the prior art, the above technical scheme has the following beneficial effects:

[0071] In the present application, by designing two independent online adaptive laws, the upper bound of the amplitude of the spoofing attack signal and the upper bound of the energy of the unknown external disturbance in the ocean are estimated in real time respectively, the system output error is continuously monitored, and the compensation gain parameter in the controller is dynamically adjusted according to the above, so as to accurately match the threat intensity currently faced, overcome the disadvantages caused by the large gain of the traditional sliding mode control for coping with the worst case, and effectively realize high-precision control.

[0072] In the present application, for the unmanned ship facing complex threats (coexistence of deception attack signal injection and unknown external disturbance in the ocean) in complex ocean environment, through the T-S fuzzy modeling framework, the complex nonlinear unmanned ship system is converted into a weighted sum of a series of linear subsystems, and the anti-deception attack adaptive sliding mode controller is designed. The controller has the approximation ability of fuzzy logic, the strong robustness of sliding mode control and the online learning ability of adaptive algorithm, so that the closed-loop control system can dynamically identify, separate and suppress different types of threats, significantly improving the overall suppression ability and survivability of the system to complex threats.

[0073] In the present application, high-precision trajectory tracking or point control can be achieved under the condition that both deception attacks and disturbances are unknown and coexist, and high-precision zero-vibration control can be achieved. Unknown ocean disturbances and model uncertainties can be overcome, and control performance can be maintained without degradation. At the same time, through the online updating ability of the adaptive algorithm, offline precalculation is not required, and time-varying threat environment can be dealt with, ensuring that the unmanned ship can safely, reliably and efficiently operate autonomously in real and harsh ocean missions.

[0074] The specific embodiments of the present application will be described in further detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0075] The accompanying drawings, which are part of the present application, serve to provide a further understanding of the present application, and the illustrative embodiments of the present application and their descriptions serve to explain the present application, but do not constitute an improper limitation on the present application. Obviously, the drawings described below are only some embodiments, and other drawings can be obtained from these drawings by those of ordinary skill in the art without creating creative labor.

[0076] In the drawings:

[0077] Figure 1 Flowchart of the anti-deception attack adaptive sliding mode control method of the unmanned ship in the present specific embodiment;

[0078] Figure 2 Logic diagram of the anti-deception attack adaptive sliding mode control method of the unmanned ship in the present specific embodiment;

[0079] Figure 3 Position response curve diagram and speed response curve diagram of the unmanned ship without any control input in the present specific embodiment, wherein a is the position response curve diagram of the unmanned ship without any control input, and b is the speed response curve diagram of the unmanned ship without any control input;

[0080] Figure 4Fig. 1 is a schematic diagram of the position response curve and the speed response curve of the unmanned ship after the anti-deception attack adaptive sliding mode control method of the unmanned ship is adopted in the embodiment of the present application, wherein a is the position response curve of the unmanned ship after the anti-deception attack adaptive sliding mode control method of the unmanned ship is adopted, and b is the speed response curve of the unmanned ship after the anti-deception attack adaptive sliding mode control method of the unmanned ship is adopted;

[0081] Figure 5 Fig. 2 is a schematic diagram of the position observer error curve and the speed observer error curve of the double adaptive sliding mode observer in the embodiment of the present application, wherein a is the position observer error curve of the double adaptive sliding mode observer, and b is the speed observer error curve of the double adaptive sliding mode observer;

[0082] Figure 6 Fig. 3 is a schematic diagram of the time response curve of the sliding mode surface function and the control rate in the embodiment of the present application, wherein a is the time response curve of the sliding mode surface function , and b is the time response curve of the control rate .

[0083] Figure 7 Fig. 4 is a schematic diagram of the structure of the anti-deception attack adaptive sliding mode control system of the unmanned ship in the embodiment of the present application. DETAILED DESCRIPTION

[0084] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. The following embodiments are used to illustrate the present application, but not to limit the scope of the present application.

[0085] As described in the background, the unmanned ship faces the composite threat of physical layer time-varying disturbance and network layer deception attack in the marine environment. The existing technology has the following problems: the traditional sliding mode control (SMC) has robustness to disturbance, but relies on the known upper bound of disturbance, and the fixed switching gain causes chattering; the existing observer cannot simultaneously decouple and jointly estimate the boundary of time-varying disturbance and dynamic attack; there is a lack of unified control framework, which makes it difficult to guarantee stability while suppressing chattering and resisting composite threats.

[0086] In the face of this problem, those skilled in the art may try to improve the traditional SMC, for example, adopt a single adaptive law to estimate the upper bound of a "total disturbance" instead of a fixed switching gain, but cannot distinguish the strength of the disturbance and the attack, resulting in blind compensation of the controller and inability to accurately regulate different types of threats; Or respectively design a disturbance observer and an attack detector, and then simply superimpose the outputs of the two to control compensation, but the two independently designed modules lack coordination, the dynamic response may not match, or even conflict with each other.

[0087] Based on this, see Figure 1 and Figure 2 , the application provides an anti-deception attack adaptive sliding mode control method for unmanned ships, comprising the following steps:

[0088] S1, constructing an unmanned ship T-S fuzzy model under deception attack and external disturbance:

[0089] The deception attack and the external disturbance are coupled in the unmanned ship T-S fuzzy model.

[0090] S2, based on the unmanned ship T-S fuzzy model constructed in S1, designing a double adaptive sliding mode observer:

[0091] The double adaptive sliding mode observer reconstructs the state estimation value using the measurable output , and estimates the upper bound of the deception attack and the external disturbance online;

[0092] At least including designing a double adaptive law, the double adaptive law is used to complete the joint estimation of the upper bound of the deception attack and the external disturbance , and dynamically adjusts the adaptive rate and the adaptive rate through the double adaptive law, wherein the adaptive rate is used to estimate the upper bound of the external disturbance, and the adaptive rate is used to estimate the upper bound of the deception attack; and,

[0093] At least including designing a robust term , the robust term uses the adaptive rate and the adaptive rate to produce a compensation signal through a sign function , which is used to actively offset the influence of the deception attack and the external disturbance ;

[0094] S3, design a sliding mode controller based on S1 and S2:

[0095] Based on the T-S fuzzy model of the unmanned ship constructed in S1, the state estimation value provided by the double adaptive sliding mode observer in S2 , the adaptive rate and the adaptive rate , design a sliding mode controller;

[0096] S4, form a closed-loop control system based on S1-S3 for sliding mode control:

[0097] Integrate the T-S fuzzy model of the unmanned ship constructed in S1, the double adaptive sliding mode observer designed in S2 and the sliding mode controller designed in S3 to form a closed-loop control system, and realize sliding mode control of the T-S fuzzy model of the unmanned ship.

[0098] It can be understood that the present application realizes accurate distinction and independent estimation of disturbances and attacks by designing a double adaptive law, solves the problem of blind compensation, realizes more delicate and smooth control, realizes deep integration of perception and control in a unified sliding mode theory framework, and guarantees the global stability of the closed-loop system.

[0099] It should be noted that the execution subject of the anti-deception attack adaptive sliding mode control method of the unmanned ship in the embodiment is an anti-deception attack adaptive sliding mode control device of the unmanned ship, which can be an electronic device, a component in an electronic device, an integrated circuit, or a chip. The electronic device can be a mobile electronic device or a non-mobile electronic device. Illustratively, the mobile electronic device can be a mobile phone, a tablet computer, a notebook computer, a palm computer, a vehicle-mounted electronic device, a wearable device, etc., and the non-mobile electronic device can be a server and a personal computer, etc., which are not limited in the present application. The anti-deception attack adaptive sliding mode control method of the unmanned ship in the embodiment will be described below with the execution subject being a server as an example.

[0100] In addition, the terms "first" and "second" are only for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined as "first" and "second" can explicitly or implicitly include one or more of the features. In the description of the embodiments of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0101] As a specific implementation, the T-S fuzzy model of the unmanned ship constructed in S1 is represented as:

[0102] ,

[0103] ,

[0104] (1);

[0105] in, Let i represent the total state of the unmanned surface vessel model, with constants i = 1, 2, 3, 4. Indicates performance output. This represents the actual output, where t is the time variable. It is a linear velocity, angular velocity, or position vector matrix. For control rate, For deception attack weighted matrix, To deceive and attack mathematical models, Due to external disturbances, As a prerequisite variable, For fuzzy sets, The damping ratio inertia matrix, Here is the mooring ratio inertia matrix. To measure the output constant matrix, To control the output constant matrix.

[0106] As a specific implementation, S2 designs a dual adaptive sliding mode observer, including: the dual adaptive sliding mode observer utilizes measurable system output The error term is estimated using the TS fuzzy model and the injected output. To reconstruct the unpredictable system state The dynamic equations of the dual adaptive sliding mode observer are expressed as:

[0107] ,

[0108] ,

[0109] (2);

[0110] in, This represents the output system state of the unmanned surface vessel (USV) TS fuzzy model with dual adaptive sliding mode observers. for The estimated value, for The estimated value, for The estimated value, for The estimated value, For observer gain, It is a robust term.

[0111] It should be noted that the estimation of the system state by the dual adaptive sliding mode observer is completed through formula (2), and the robustness term... To ensure that the estimation error still converges in the face of attacks and external disturbances.

[0112] In the presence of deception attacks and complex ocean disturbances, the total upper bound of the system's uncertainty is unknown and time-varying. Traditional SMC has to choose a very large and extremely conservative fixed switching gain to ensure robustness, which inevitably leads to severe control input chattering, harmful to actuators and difficult to apply in practice. The dual adaptive law of the present application aims to fundamentally solve this problem, and the beneficial effects brought about are: the system does not need to know the worst-case scenario in advance, the adaptive law can learn by itself, completely eliminating the dependence on fixed upper bound; the gain varies as needed, rather than always being the maximum value, generating a smoother control signal that significantly suppresses chattering; even if the attack pattern or disturbance intensity changes over time, the sliding mode controller can continuously maintain excellent performance through adaptive adjustment, enhancing the system's resilience.

[0113] The uniqueness of dual adaptation lies in: dual estimation targets, it is not to estimate a single "total disturbance", but to estimate two different physical meanings of unknown boundaries in parallel and independently, namely: the gain upper bound of external disturbance relative to output And the weighted matrix norm upper bound of attack signal Dual design channels: although the structures of the two adaptive laws are similar, the driving signals and normalization factors are designed for their respective targets, that is: the external disturbance adaptive rate is coupled with the output energy The attack adaptive law is coupled with the system information known by the attacker. .

[0114] As a specific embodiment, S2 at least includes designing a dual adaptive law, the dual adaptive law is represented as:

[0115] (4);

[0116] (5);

[0117] Wherein, represents the disturbance boundary estimation update law, represents the attack boundary estimation update law, is the anti-network attack adaptive rate constant, is the anti-disturbance adaptive rate constant, is the upper limit of network attack, represents the transpose of the observer error, represents the transpose of the measured output constant matrix, represents the actual output norm.

[0118] As a specific embodiment, S2 at least also includes designing a robust term , the robust term with adaptive rates and adaptive rates , through sign functions to produce compensation signals for actively canceling the effects of spoofing attacks and external disturbances .

[0119] robust terms , denoted as:

[0120] (3)

[0121] where are the estimates of , respectively, is the anti-disturbance adaptive rate, is the anti-network attack adaptive rate, is the robust term constant, is the observer error weight matrix, is the observer error.

[0122] It can be understood that the robust terms in formula (3) and the switching control terms in the following formula (39) work together to achieve real-time compensation for spoofing attacks and external disturbances, and the robust terms use adaptive rates and , through sign functions to produce compensation signals, which are injected into the double adaptive sliding mode observer and the sliding mode controller, for actively canceling the effects of spoofing attacks and external disturbances . The switching terms in formula (39) further ensure that the system trajectory can approach and remain on the sliding surface, enhancing the robustness of compensation.

[0123] It should be noted that the joint estimation of the spoofing attack and external disturbance boundaries is completed through formula (4) and formula (5), which dynamically adjust the adaptive rates and adaptive rates online through information such as double adaptive sliding mode observer error and actual output , where is for the anti-disturbance adaptive rate , is for the anti-network attack adaptive rate .

[0124] As a specific embodiment, S2 further includes at least a design error dynamic analysis, specifically:

[0125] definition ;

[0126] Combining formulas (1) and (2), the error dynamic equation is expressed as:

[0127] (6);

[0128] (7);

[0129] in, This represents the error update rate of the dual adaptive sliding mode observer. For the dual adaptive sliding mode observer error, This represents the error residual term of the dual adaptive sliding mode observer.

[0130] It is worth noting the error of the dual adaptive sliding mode observer. Quantify the true state of the system With system estimated state The difference between them, the goal of the sliding mode controller is to make the error of the dual adaptive sliding mode observer... Approaching 0, thus ensuring The accuracy, ultimately ensuring based on The control rate formula (39) Effective operation. The update rates of formulas (4) and (5) directly depend on It is obvious that the dual adaptive sliding mode observer has errors. It is the driving force behind the adaptive process. When the attack or disturbance increases, it leads to errors in the dual adaptive sliding mode observer. As the value increases, the adaptive law will accelerate the adjustment of the adaptive rate. and adaptive rate As the attack or disturbance decreases, the state estimation becomes more accurate, leading to errors in the dual adaptive sliding mode observer. As the value decreases, the adjustment of the adaptive law also slows down, while avoiding the adaptive rate. and adaptive rate Excessive drift.

[0131] Ideally, the dual adaptive sliding mode observer uses fuzzy model weights. Weights of the real system Ideally, they should be completely consistent; however, due to factors such as modeling errors, such consistency is virtually impossible. (Dual adaptive sliding mode observer error residuals) This can precisely capture and encapsulate the modeling uncertainty caused by fuzzy weight mismatch, using the error residual term of the dual adaptive sliding mode observer. Due to weighting error and system dynamics Two terms, indicating the error residual term of the dual adaptive sliding mode observer is not an arbitrary disturbance, but a term related to the system state with clear physical meaning, assuming the error residual term of the dual adaptive sliding mode observer , the complexity of stability analysis proof can be simplified.

[0132] The adaptive law and compensation mechanism form a closed loop of perception, learning, compensation and convergence, where perception: the error of the dual adaptive sliding mode observer increases, indicating that there is a mismatched disturbance / attack; learning: the adaptive law formula (4) and formula (5) increase the estimates of the error of the dual adaptive sliding mode observer and the system state, and the adaptive rate and the adaptive rate ; compensation: the increase of the adaptive rate and the adaptive rate increases the amplitude of the robust term , thereby generating stronger control action to suppress disturbances and attacks; convergence: after the disturbance and attack are suppressed, the error of the dual adaptive sliding mode observer decreases, the adjustment effect of the adaptive law weakens, and the system reaches a new equilibrium.

[0133] As a specific embodiment, a sliding mode controller is designed in S3, including designing a sliding surface, denoted as:

[0134] (8);

[0135] wherein, is a sliding surface function, is a sliding matrix, is a sliding mode controller gain, is a sliding mode controller integral variable;

[0136] Combining formula (2) and formula (8), the rate of change of the sliding surface function is obtained, denoted as:

[0137] (9);

[0138] Define the time-varying weighting matrix , let , the equivalent control rate is obtained, denoted as:

[0139] (10);

[0140] Substitute formula (10) into formula (2) to obtain the sliding mode dynamics equation, denoted as:

[0141] (11).

[0142] As a specific embodiment, the sliding mode controller in S3 is designed, further comprising:

[0143] A Lyapunov function is constructed, denoted as:

[0144] (38);

[0145] wherein, represents the transpose of;

[0146] A control law is implemented for the sliding mode region reachability construction under deception attacks, denoted as:

[0147] (39);

[0148] wherein, is the combined gain of the controller, is denoted as a time-varying weighted matrix inverse;

[0149] Define , wherein, is the anti-disturbance adaptive rate, is the anti-network attack adaptive rate, is a robust term constant, is a reaching condition constant;

[0150] Derive the Lyapunov function , to obtain:

[0151] (40);

[0152] wherein, represents the derivative of the Lyapunov function, represents the rate of change of the sliding surface function, which indicates that under the condition of deception attacks, the closed-loop control system is driven to the sliding surface by the sliding mode control law formula (39) .

[0153] In an implementable embodiment, further comprising state stability analysis, which is specifically analyzed as follows:

[0154] A Lyapunov function function is constructed:

[0155] (12);

[0156] Take , wherein, is the anti-disturbance adaptive rate error, is the anti-network attack adaptive rate error, denotes the upper bound coefficient of the disturbance, denotes the upper bound coefficient of the attack weight, is the state estimation weight matrix, is the estimation error weight matrix;

[0157] Taking the derivative of the constructed Lyapunov function, we have

[0158] (14);

[0159] where, denotes the inverse of the anti-network attack adaptive rate constant, denotes the inverse of the anti-disturbance adaptive rate constant;

[0160] Substituting formula (6) and formula (11) into formula (14), we have

[0161] (15);

[0162] Taking Substituting formula (15), we have

[0163] (16);

[0164] Taking , we have Substituting formula (3), we have

[0165] (17);

[0166] where, denotes the fuzzy weighted observation error norm;

[0167] Substituting formula (4) and formula (5) into and calculating, we have

[0168] (18);

[0169] Adding formula (17) and formula (18), we have

[0170] (19);

[0171] Using Young's inequality, we have

[0172] (20);

[0173] where, is the disturbance weighted matrix, denotes the inverse of the disturbance weighted matrixthe maximum eigenvalue of represents an upper bound of the fuzzy modeling residual error energy.

[0174] Substitute formula (19) and formula (20) into formula (16), and obtain:

[0175] (21);

[0176] where , is an augmented state vector;

[0177] (22);

[0178] where is a conditional matrix. Construct a state intermediate matrix where represents a sliding mode dynamic relaxation matrix, represents an observation error dynamic relaxation matrix, then:

[0179] (23);

[0180] Construct a conversion matrix , satisfying where represents a unit matrix;

[0181] Multiply formula (23) by respectively on the left and right, and obtain:

[0182] (24);

[0183] where the controller gain intermediate variable ; since:

[0184] (25);

[0185] (26);

[0186] In order to facilitate the MATLAB LMI toolbox to solve, transform formula (35) and formula (36) into the following inequalities:

[0187] (27);

[0188] (28);

[0189] where the state intermediate variable , the disturbance intermediate variable ;

[0190] From equation (21) and equation (24), we have:

[0191] where, is the intermediate state matrix, is the intermediate state matrix is the minimum eigenvalue of, is the augmented state vector;

[0192] Define the state transition parameter and the disturbance transition parameter, equation (29) can be obtained by:

[0193] (30);

[0194] Solve the differential inequality of equation (12) and equation (29), we have:

[0195] (31);

[0196] Continue to solve, we have:

[0197] (32);

[0198] where, denotes the stable analysis integral variable, and the comprehensive transition parameter ;

[0199] Thus, we have, the function and the function :

[0200] (33);

[0201] (34);

[0202] When is fixed, for , it is linearly increasing, which is a function of ; when is fixed, for , it is exponentially decaying to 0, satisfying the property of function; continuous and strictly increasing, and , satisfying the property of function; Therefore, equation (30) is a sufficient condition for the input-state stability of the closed-loop system.

[0203] Therefore, if there exists

[0204] ,​ satisfies:

[0205] (35);

[0206] (36);

[0207] (37);

[0208] For , the sliding mode dynamic equation formula (11) is input to the state stability.

[0209] In an implementable embodiment, further comprising the step of simulating by giving the parameters of the closed-loop control system composed of the controller and the observer, specifically comprising:

[0210] , ,

[0211] , ,

[0212] , ,

[0213] , .

[0214] wherein A is the damping characteristic matrix, B is the restoring force characteristic matrix, D is the control and disturbance input matrix, is the control output constant matrix 1, is the control output constant matrix 2, is the control output constant matrix 3, is the control output constant matrix 4.

[0215] , , , , , ;

[0216] wherein, are rotation matrix 1, rotation matrix 2, rotation matrix 3 and rotation matrix 4 respectively. The parameters in the sliding mode control rate are set to respectively. The parameters in the deception attack model are set to respectively. The ocean disturbance is set to ocean disturbance 1, ocean disturbance 2, ocean disturbance 3, . The sliding mode parameters are set to wherein denotes the generalized inverse. Obviously, is nonsingular.

[0217] The system is simulated according to the simulation conditions given above to verify the input-state stability of the system. The application provides comparative simulation graphs to highlight the necessity of the control strategy of the application.

[0218] Please refer to Figure 3 , Figure 3 , which shows the open-loop response of the unmanned ship system when it is subjected to only deception attacks and external disturbances without applying any control input. The position response curve represents the position and and the yaw angle of the unmanned ship over time. The position response curve shows divergence or sustained large oscillation, indicating that the unmanned ship cannot maintain the predetermined heading or position and is out of control. The speed response curve represents the longitudinal speed , lateral speed and yaw angular speed of the unmanned ship over time, and the speed response curve also shows the characteristics of divergence or severe jitter, indicating that the kinetic energy is not effectively controlled.

[0219] Please refer to Figure 3 , Figure 3 , which is a contrastive graph showing that the unmanned ship system cannot maintain stability if an effective control strategy is not adopted under the security threat of deception attacks and external disturbances, thereby demonstrating the necessity and importance of the control strategy proposed in the application from the negative side.

[0220] Please refer to Figure 4 , Figure 4 , which is the position and speed response curve of the unmanned ship after the adaptive sliding mode control law proposed in the application is adopted, for demonstrating the control effect of the application. The position response curve represents the position and and the yaw angle of the unmanned ship, and the curve can quickly converge to the expected value or remain bounded within a small fluctuation in its neighborhood, which indicates that the controller of the application can effectively overcome the influence of attacks and disturbances and achieve accurate trajectory tracking or point keeping. The speed response curve represents the longitudinal speed , lateral speed and yaw angular speed of the unmanned ship over time, which also shows good convergence and smoothness, tending to a stable value or zero, indicating that the dynamic process of the system is smooth.

[0221] Please refer to Figure 3 and Figure 4 ,​Figure 4 In contrast, the effectiveness and robustness of the control method proposed in this application are directly verified. It is proved that this application can ensure the stability of the closed-loop system and satisfactory performance indicators when the unmanned ship is subjected to network deception attacks and complex ocean environment disturbances at the same time. Figure 3 Please see

[0222] , Figure 5 , Figure 5 is the state estimation error curve of the double adaptive sliding mode observer, which is used to show the estimation accuracy of the double adaptive sliding mode observer. Figure 5 The error curve of the double adaptive sliding mode observer is shown with time, which contains 6 lines corresponding to the estimation error of different state components of the system. All estimation error curves show the characteristics of rapid convergence from the initial error to a very small neighborhood around zero, and then always remain within the bounded range.

[0223] Please continue to see Figure 5 , Figure 5 It is proved that the double adaptive sliding mode observer designed in this application has good performance, which can accurately and quickly reconstruct all the states of the system by relying only on the actual output , providing reliable input for the state feedback-based control law. The bounded convergence of the double adaptive sliding mode observer error also provides key support for the stability proof of the entire closed-loop system.

[0224] Please see Figure 6 , Figure 6 is the time response curve of the sliding surface function and the control input , which is used to show the achievement of sliding mode motion and the physical realizability of control input. The sliding surface function curve shows the value of the sliding surface function defined according to formula (8) changes with time. The curve quickly converges to a boundary layer around zero in a limited time, and maintains in this area thereafter. This strictly proves that the system trajectory is driven and maintained near the sliding surface, meeting the basic requirements of sliding mode control, thereby ensuring the robustness of the system. The control input curve shows the control signal of the actual output of the controller. Although the signal is continuous, it contains high-frequency switching components. It is worth noting that its amplitude is bounded, and the strength of switching chattering is effectively suppressed within an acceptable range.

[0225] Please continue to see Figure 6 , Figure 6The convergence through the sliding mode surface proves the realization of the control target; and the adaptive mechanism effectively weakens the inherent "chattering" problem of the traditional sliding mode control, generates a more smooth and more easily implemented physical control signal by the actuator, and proves the engineering practicability of the application.

[0226] The simulation results show that the adaptive sliding mode control method based on T-S fuzzy modeling of the application can effectively combine the strong robustness of the sliding mode control, the approximation ability of the fuzzy logic to the nonlinear system and the online learning advantage of the adaptive technology, and realize the dynamic compensation of the deception attack and the external disturbance.

[0227] Please refer to Figure 3 , Figure 4 and Figure 6 , as shown in comparison with Figure 3 and Figure 4 , the application enables the system to quickly recover and remain stable when suffering from a composite threat, with fast response and strong anti-interference capability. Figure 4 It is proved that the double adaptive sliding mode observer designed in the application can provide accurate state estimation, with high tracking accuracy and small estimation error, providing a reliable basis for the control law. Figure 6 Further, the application effectively suppresses the chattering of the control input through the adaptive mechanism, ensuring robustness while having good control flexibility and engineering realizability. Figure 6 In the application, is a longitudinal sliding mode surface, is a lateral sliding mode surface, is a steering sliding mode surface; is a longitudinal control input, is a lateral control input, is a steering control input.

[0228] Based on the same inventive concept, please refer to Figure 7 , the application also provides an anti-deception attack adaptive sliding mode control system for an unmanned ship, which comprises:

[0229] A first design unit constructs a T-S fuzzy model of the unmanned ship under deception attack and external disturbance: the T-S fuzzy model of the unmanned ship is coupled with deception attack and external disturbance ;

[0230] A second design unit designs a double adaptive sliding mode observer based on the constructed T-S fuzzy model of the unmanned ship: the double adaptive sliding mode observer reconstructs the state estimation value using the measurable output , and online estimates the upper bound of the deception attack and the external disturbance .

[0231] at least comprising a double adaptive law designed to accomplish the estimation of the upper bound of the spoofing attack and the external disturbance the joint estimation of the boundary, the adaptive rate is dynamically adjusted by the double adaptive law and the adaptive rate , wherein the adaptive rate is used to estimate the upper bound of the external disturbance, the adaptive rate is used to estimate the upper bound of the spoofing attack; and,

[0232] at least comprising a robust term designed , the robust term utilizes the adaptive rate and the adaptive rate , a compensation signal is produced by a sign function to actively offset the influence of the spoofing attack and the external disturbance ;

[0233] a third design unit, based on the first design unit and the second design unit, designs a sliding mode controller: based on the constructed T-S fuzzy model of the unmanned ship, the state estimation value provided by the double adaptive sliding mode observer in S2 , the adaptive rate and the adaptive rate , the sliding mode controller is designed;

[0234] a closed-loop control unit, based on the first design unit, the second design unit and the third design unit, forms a closed-loop control system for sliding mode control: the T-S fuzzy model of the unmanned ship constructed by the first design unit, the double adaptive sliding mode observer designed by the second design unit and the sliding mode controller designed by the third design unit are integrated to form a closed-loop control system, and the sliding mode control is realized.

[0235] Based on the same inventive concept, the application also provides a computer readable storage medium storing a computer program, and the computer program is executed by a processor to realize the anti-spoofing attack adaptive sliding mode method of the unmanned ship as described above.

[0236] The program product of the application for realizing the above-mentioned method can adopt a portable compact disc read-only memory and includes a program code, and can be run on a terminal device, such as a personal computer. However, the program product of the application is not limited to this, and in the application, the readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, device or apparatus.

[0237] It is understood in light of the worldwide nature of commerce, that the present application can be practiced in a number of environments, some of which can be different from those described herein. It is also understood that, regardless of the environment used to practice the application, the principles of the present application can be useful. It is therefore understood that there is a further aspect of the application that relates to the use of the systems, devices and methods disclosed herein and / or any combinations of the systems, devices and methods disclosed herein in the environments of commerce. As such, at least one further aspect of the application relates to a use of the systems, devices and methods disclosed herein and / or any combinations of the systems, devices and methods disclosed herein in the environments of commerce.

[0238] The above merely preferred embodiments of the present application and are not intended to limit the present application. Since the application has been disclosed with the preferred embodiments, modifications and variations of the preferred embodiments could become apparent to those skilled in the art in light of the above teachings. It is therefore contemplated that the use of such modifications and variations are intended to be covered by the above description. It is intended, therefore, that the present application be considered in all its aspects as only a preferred embodiment of the application, and that all changes and modifications of the preferred embodiment come within the scope of the application. Accordingly, what is desired to be secured by Letters Patent is the application as defined by the appended claims, the description and the drawings.

Claims

1. An anti-deception attack adaptive sliding mode control method for an unmanned ship, characterized in that, The method comprises the following steps: S1, constructing a T-S fuzzy model of the unmanned ship under a deception attack and external disturbance: The unmanned ship T-S fuzzy model is coupled with deception attack and external disturbance ; S2, designing a double adaptive sliding mode observer based on the T-S fuzzy model of the unmanned ship constructed in S1: The dual adaptive sliding mode observer utilizes measurable outputs to reconstruct state estimates and online estimate upper bounds of spoofing attacks and external disturbances ​ at least comprising a double adaptive law for accomplishing a joint estimation of the upper bound of the spoofing attack and the external disturbance and the adaptive rate and the adaptive rate , wherein the adaptive rate is used for estimating the upper bound of the external disturbance is used for estimating the upper bound of the spoofing attack; and, At least include design robustness items The robustness term Using the adaptive rate and the adaptive rate Through symbolic functions Production compensation signals are used to proactively counteract deceptive attacks. and external disturbances The impact; S3, designing a sliding mode controller based on S1 and S2: Based on the S1 constructed unmanned ship T-S fuzzy model, using S2 in the state estimation value provided by double adaptive sliding mode observer , adaptive rate and adaptive rate , design sliding mode controller; S4, forming a closed-loop control system based on S1-S3 for sliding mode control: The T-S fuzzy model of the unmanned ship constructed in S1, the double adaptive sliding mode observer designed in S2, and the sliding mode controller designed in S3 are integrated to form a closed-loop control system, thereby realizing the sliding mode control of the T-S fuzzy model of the unmanned ship in S1.

2. The method of claim 1, wherein, The T-S fuzzy model of the unmanned ship constructed in S1 is expressed as: ， ， (1)； wherein, x represents the system full state of the unmanned ship model, constant i = 1, 2, 3, 4, y represents the performance output, y represents the actual output, t is a time variable, is a linear velocity or angular velocity or position vector matrix, is a control rate, is a spoofing attack weighting matrix, is a spoofing attack mathematical model, is an external disturbance, is a premise variable, is a fuzzy set, is a damping ratio inertia matrix, is a mooring ratio inertia matrix, is a measurement output constant matrix, is a control output constant matrix.

3. The method of claim 2, wherein, In S2, a double adaptive sliding mode observer is designed, including: the double adaptive sliding mode observer utilizes measurable system outputs , adopts a T-S fuzzy model and injects an output estimation error term to reconstruct unmeasurable system states , and a dynamic equation of the double adaptive sliding mode observer is expressed as: ， ， (2)； wherein, represents the system full state of the dual adaptive sliding mode observer, is an estimate of is an estimate of is an estimate of is an estimate of is an estimate of is an estimate of is an estimate of is an estimate of is the dual adaptive sliding mode observer gain, is the robust term.

4. The method of claim 3, wherein, In S2, at least a double adaptive law is designed, which is expressed as: (4)； (5)； wherein, represents a disturbance boundary estimation update law, represents an attack boundary estimation update law, represents a double adaptive sliding mode observer error weight matrix, is an anti-network attack adaptive rate constant, is an anti-disturbance adaptive rate constant, is a network attack upper bound, represents a transpose of a double adaptive sliding mode observer error, represents a transpose of a measurement output constant matrix, represents an actual output norm.

5. The method of claim 4, wherein, S2 comprises at least a design robustness term , said robustness term utilizes said adaptive rate and said adaptive rate , a compensation signal is produced by a sign function for actively counteracting the influence of spoofing attacks and external disturbances ; The robust term is represented as: (3)； wherein are respectively an estimate of is an anti-disturbance adaptive rate, is an anti-network attack adaptive rate, is a robust term constant, is a dual adaptive sliding mode observer error weight matrix, is a dual adaptive sliding mode observer error.

6. The method of claim 4, wherein, In S2, at least an error dynamic analysis is also designed, which is specifically: Definitions ; In combination with formula (1) and formula (2), the error dynamic equation is expressed as: (6)； (7)； wherein, represents a double adaptive sliding mode observer error update rate, is a double adaptive sliding mode observer error, is a double adaptive sliding mode observer error residual term.

7. The method of claim 4, wherein, In S3, the sliding mode controller is designed, including a sliding mode surface, which is expressed as: (8)； wherein, is a sliding mode surface function, is a sliding mode matrix, is a sliding mode controller gain, is a sliding mode controller integral variable; Combining equation (2) and equation (8), the rate of change of the sliding mode surface function is obtained , which is expressed as: (9)； Defining the time-varying weighting matrix , let , the equivalent control rate is expressed as: (10)； Formula (10) is substituted into formula (2) to obtain a sliding mode dynamics equation, which is expressed as: (11)； wherein, is denoted as a time-varying weighting matrix inverse.

8. The method of claim 7, wherein, In S3, the sliding mode controller also includes: A Lyapunov function is constructed, which is expressed as: (38)； wherein denotes the transpose of A control law for realizing the reachability of the sliding mode region under a deception attack is constructed, which is expressed as: (39)； wherein, Gc is the controller combination gain; Definitions ; wherein, is an anti-jamming adaptive rate, is an anti-network attack adaptive rate, is a robust term constant, is a reach condition constant; lyapunov function taking the derivative, we obtain: (40)； wherein, denotes the derivative of the Lyapunov function, denotes the rate of change of the sliding surface function, which indicates that under the condition of a deception attack, the closed-loop system is driven onto the sliding surface by the sliding mode control law (39) .

9. An anti-deception attack adaptive sliding mode control system for an unmanned ship, characterized by, The sliding mode control system comprises: A first design unit, a T-S fuzzy model of the unmanned ship under the deception attack and external disturbance is constructed: the T-S fuzzy model of the unmanned ship is coupled with the deception attack and the external disturbance and the external disturbance ​ The second design unit, based on the constructed unmanned vessel TS fuzzy model, designs a dual adaptive sliding mode observer: the dual adaptive sliding mode observer utilizes measurable output Reconstructed state estimate And estimate deception attacks online. and external disturbances The upper bound; at least comprising a double adaptive law for accomplishing a joint estimation of the upper bound of the spoofing attack and the external disturbance and the adaptive rate and the adaptive rate , wherein the adaptive rate is used for estimating the upper bound of the external disturbance is used for estimating the upper bound of the spoofing attack; and, at least comprising a design robust term , the robust term utilizing the adaptive rate and the adaptive rate , a compensation signal is produced by a sign function for actively counteracting the influence of spoofing attacks and external disturbances ; A third design unit designs the sliding mode controller based on the first design unit and the second design unit: based on the constructed T-S fuzzy model of the unmanned ship, the state estimation value provided by the double adaptive sliding mode observer in S2 is used to design the sliding mode controller , the adaptive rate and the adaptive rate ​ A closed-loop control unit forms a closed-loop control system based on the first design unit, the second design unit, and the third design unit for sliding mode control: the T-S fuzzy model of the unmanned ship constructed by the first design unit, the double adaptive sliding mode observer designed by the second design unit, and the sliding mode controller designed by the third design unit are integrated to form a closed-loop control system, thereby realizing the sliding mode control of the T-S fuzzy model of the unmanned ship.

10. A computer readable storage medium storing a computer program, characterized in that, The computer program, when executed by a processor, is used to realize the adaptive sliding mode control method of the unmanned ship against deception attacks according to any one of claims 1-8.

Images