Safety Control Method and System for Unmanned Surface Vessels Based on Adaptive Neural Network Observer

Abstract

This invention discloses a safety control method and system for unmanned surface vessels (USVs) based on an adaptive neural network observer, relating to the field of USV control technology. The method includes: constructing a second-order dynamic model of the USV using dynamic equations and obtaining its actual position and velocity state; designing an adaptive observer based on a neural network using the approximate nonlinear dynamics of a neural network to estimate the first-order system position state and the second-order system velocity state under DoS attacks and external disturbances; determining the position tracking error between the actual and estimated positions of the USV based on a preset desired trajectory and combining the actual and estimated positions, and constructing a virtual controller; designing a nonlinear filter based on the virtual control signal to obtain the trajectory tracking signal of the second-order velocity system, and then constructing an adaptive controller based on the estimated velocity state; and realizing the desired trajectory tracking of the USV under DoS attacks and disturbances based on the virtual controller and the adaptive controller.

Description

Technical Field

[0001] This invention relates to the field of unmanned surface vessel (USV) control technology, and in particular to a USV safety control method and system based on an adaptive neural network observer. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Unmanned surface vessels (USVs), also known as unmanned surface ships, are a type of nonlinear motion control system that operates on the water surface. They have wide applications in water transport, reconnaissance, exploration, and rescue. In practice, the navigation control of USVs is typically achieved by moving the vessel along a given reference trajectory; that is, the control objective is to ensure the vessel tracks a predetermined path. However, due to the complexity of the operating environment, USV systems are inevitably affected by external environmental disturbances, and are vulnerable to malicious attacks during the communication process from the USV's motion status to the control center, leading to safety issues. Therefore, ensuring that USVs maintain trajectory tracking under external disturbances and malicious communication attacks is of significant practical importance.

[0004] To address the security issues of the aforementioned unmanned surface vessel (USV) systems, three common attack behaviors exist: Denial-of-Service (DoS) attacks, data injection, and spoofing attacks. Among these, a DoS attack involves an attacker sending a large amount of useless data to block the channel's real-time transmission of system data, causing the system to malfunction. This attack is one of the most harmful to data transmission. Existing USV security control methods that consider DoS attacks primarily employ the design of observers to estimate the missing states caused by DoS attacks, thereby ensuring effective data transmission. This is currently one of the mainstream security control methods. However, accurately estimating the missing information of the system is quite difficult, especially under external disturbances, where existing observers are even less able to accurately estimate the missing information.

[0005] On the other hand, due to the cascaded nonlinearity of the unmanned surface vessel's position and velocity states, backstepping control is one of the effective control methods to ensure the safe operation of the unmanned surface vessel. However, the backstepping control process suffers from repeated state differentiation, leading to an explosion in computational complexity. Although designing auxiliary variables using filtering operators can solve the complexity explosion problem in the backstepping control process—for example, current proposals have suggested designing linear filters to handle the repeated differentiation problem in the backstepping safety control process under DoS attacks—existing linear filter designs lack general applicability and have certain filtering errors, resulting in inaccurate filtered signals and consequently affecting the accuracy of the tracking control strategy.

[0006] Moreover, in practical applications, various unexpected factors can affect the stable operation of unmanned surface vessels (USVs). For example, in complex environments, external disturbances such as strong winds, surges, and ocean currents can all affect the performance of USVs. Due to the suddenness and persistence of external disturbances, disturbance compensation is one of the major challenges for the smooth operation of the control system. However, existing USV control schemes rely on relatively strict assumptions regarding external disturbances, such as being constrained by the differentiability of the disturbance and the boundedness of its derivative. This limits them to a very small number of continuous and slowly varying external disturbances, resulting in poor applicability and accuracy of the control scheme. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a safety control method and system for unmanned surface vessels (USVs) based on an adaptive neural network observer. The method involves designing an observer based on an adaptive neural network to reconstruct the system's state, utilizing backstepping control principles to design stepwise virtual control signals, and designing a nonlinear filter with filtering error suppression to correlate these virtual control signals, thereby providing accurate auxiliary reference signals. A safety controller is then constructed based on the observer and filter to ensure the boundedness of all system signals. This adaptive control strategy suppresses the negative impacts of DoS attacks and external disturbances on USV operation during actual operation, solving the problems of poor applicability and accuracy in trajectory tracking control of USVs subjected to DoS attacks and interference.

[0008] In a first aspect, the present invention provides a safety control method for unmanned surface vessels based on an adaptive neural network observer.

[0009] A safety control method for unmanned surface vessels based on an adaptive neural network observer includes:

[0010] Using the dynamic equations, a second-order dynamic model of the unmanned surface vessel (USV) is constructed, and the actual position and velocity state of the USV are obtained.

[0011] By utilizing the online nonlinear dynamics of position and velocity states in a second-order dynamic model to approximate the second-order dynamic model using neural networks, an adaptive observer based on neural networks is designed to estimate the position state of the first-order system and the velocity state of the second-order system under DoS attacks and external disturbances.

[0012] Based on the preset expected trajectory and combined with the actual and estimated position states, the tracking error between the actual and estimated positions of the unmanned surface vessel is determined, thereby constructing a virtual controller for the first-order position system of the unmanned surface vessel.

[0013] Based on the virtual control signal output by the virtual controller, a nonlinear filter is designed to obtain the trajectory tracking signal of the second-order velocity system. Then, combined with the estimated velocity state, an adaptive controller for the second-order velocity system is constructed.

[0014] Based on virtual controllers and adaptive controllers, we can perform desired trajectory tracking of unmanned surface vessels under DoS attacks and disturbances.

[0015] Secondly, the present invention provides a safety control system for unmanned surface vessels based on an adaptive neural network observer.

[0016] A safety control system for unmanned surface vessels based on an adaptive neural network observer includes:

[0017] The data acquisition and model building module is used to construct a second-order dynamic model of the unmanned surface vessel using dynamic equations, and to obtain the actual position and velocity state of the unmanned surface vessel.

[0018] The state estimation module is used to approximate the online nonlinear dynamics of the position and velocity states in the second-order dynamic model using a neural network, and to design a neural network-based adaptive observer to estimate the first-order system position state and the second-order system velocity state under DoS attacks and external disturbances.

[0019] The virtual controller construction module is used to determine the actual and estimated position tracking error of the unmanned surface vessel (USV) based on the preset expected trajectory and the actual and estimated position states, thereby constructing the virtual controller of the USV's first-order position system.

[0020] The adaptive controller construction module is used to design a nonlinear filter based on the virtual control signal output by the virtual controller, obtain the trajectory tracking signal of the second-order velocity system, and then construct the adaptive controller of the second-order velocity system by combining the estimated velocity state.

[0021] The tracking control module is used to track the desired trajectory of the unmanned surface vessel under DoS attacks and disturbances, based on a virtual controller and an adaptive controller.

[0022] Thirdly, the present invention also provides an electronic device, comprising: a memory for storing executable instructions; and a processor for implementing the above-described unmanned surface vessel safety control method based on an adaptive neural network observer when executing the executable instructions stored in the memory.

[0023] Fourthly, the present invention also provides a computer-readable storage medium storing executable instructions for causing a processor to execute the executable instructions to implement the above-described unmanned surface vessel safety control method based on an adaptive neural network observer.

[0024] Fifthly, the present invention also provides a computer program product comprising executable instructions stored in a computer-readable storage medium; wherein, when the processor of an electronic device reads the executable instructions from the computer-readable storage medium and executes the executable instructions, the aforementioned unmanned surface vessel safety control method based on an adaptive neural network observer is implemented.

[0025] The above one or more technical solutions have the following beneficial effects:

[0026] 1. This invention provides a safety control method and system for unmanned surface vessels (USVs) based on an adaptive neural network observer. An observer based on an adaptive neural network is designed to reconstruct the system state. Utilizing the concept of backstepping control, stepwise virtual control signals are designed, and a nonlinear filter with filtering error suppression is designed to correlate the virtual control signals, thereby providing accurate auxiliary reference signals. A safety controller is then constructed based on the observer and the filter to ensure the boundedness of all system signals. Through the above adaptive control strategy, the negative impact of DoS attacks and external disturbances on the operation of USVs during actual operation is suppressed, solving the problems of poor applicability and accuracy of trajectory tracking control for USVs under DoS attacks and interference.

[0027] 2. In this invention, considering the nonlinear characteristics of unmanned surface vessels (USVs), neural network approximation technology is integrated to effectively handle nonlinear dynamics. A neural network-based observer is designed to enhance the ability to estimate missing information of USV systems under DoS attacks. Based on this, an adaptive observation function is introduced. The system state during the attack process is approximated by the neural network-based adaptive observer, and the influence of interference is suppressed by the adaptive observation function to ensure the accuracy of trajectory tracking. In addition, an adaptive safety controller based on the observer is designed to ensure bounded output tracking of the USV system under DoS attacks and interference, ensuring the effectiveness and applicability of trajectory tracking.

[0028] 3. The method designed in this invention has strong stability and robustness, can handle the negative impacts of DoS attacks and external disturbances, and enables the unmanned surface vessel system to navigate along a predetermined trajectory, which has practical significance.

[0029] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0030] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0031] Figure 1This is a flowchart of the unmanned surface vessel safety control method based on an adaptive neural network observer as described in an embodiment of the present invention;

[0032] Figure 2 This is a structural diagram of the unmanned surface vessel in an embodiment of the present invention;

[0033] Figure 3 This is a diagram showing the position tracking error of the unmanned surface vessel in an embodiment of the present invention.

[0034] Figure 4 This is a diagram showing the expected trajectory and actual position trajectory of the unmanned surface vessel in an embodiment of the present invention;

[0035] Figure 5 This is a diagram showing the actual speed trajectory of the unmanned surface vessel in an embodiment of the present invention;

[0036] Figure 6 This is a schematic diagram of the state response curve of the unmanned surface vessel filter in an embodiment of the present invention;

[0037] Figure 7 This is a position estimation diagram of the unmanned surface vessel observer in an embodiment of the present invention;

[0038] Figure 8 This is a velocity estimation diagram of the unmanned surface vessel observer in an embodiment of the present invention;

[0039] Figure 9 This is a schematic diagram of the adaptive estimation value of the unmanned surface vessel observer in an embodiment of the present invention;

[0040] Figure 10 This is a schematic diagram of the adaptive estimation value of the neural network weights of the unmanned surface vessel in an embodiment of the present invention;

[0041] Figure 11 This is a schematic diagram of the virtual control input response curve of the first-order system of the unmanned surface vessel in an embodiment of the present invention;

[0042] Figure 12 This is a schematic diagram of the actual control input response curve of the unmanned surface vessel in an embodiment of the present invention. Detailed Implementation

[0043] It should be noted that the following detailed descriptions are exemplary and are intended only to describe specific embodiments and to provide further explanation of the invention, and are not intended to limit the scope of exemplary embodiments of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0044] Terminology Explanation

[0045] (1) Unmanned surface vessels are unmanned surface ships.

[0046] (2) Adaptive control refers to the control system being able to adapt to changes in internal dynamics or external disturbances by estimating and adjusting the system control parameters online.

[0047] (3) Denial-of-service attack, also known as DoS attack, refers to an attacker sending a large amount of useless data to block the channel from transmitting system data in real time, causing the system to malfunction.

[0048] (4) A neural network is a computational model consisting of a large number of interconnected nodes (or neurons). Through multi-layer combination and stacking, neural networks can approximate complex nonlinear functions very well.

[0049] (5) External disturbances refer to disturbances in the external environment of the control system, such as strong winds, surges, ocean currents, etc., which can affect the performance of the unmanned surface vessel system.

[0050] Example 1

[0051] This embodiment provides a safety control method for unmanned surface vessels based on an adaptive neural network observer, such as... Figure 1 As shown, when the unmanned surface vessel (USV) system operates in an environment with DoS attacks and external disturbances, an adaptive observer and safety controller are established through system errors and neural networks to enable the USV to navigate and track a preset target trajectory. This method specifically includes the following steps:

[0052] Step S1: Using the dynamic equations, construct a second-order dynamic model of the unmanned surface vessel (USV) and obtain its actual position and velocity state.

[0053] First, we establish the dynamics model and the second-order system model of the unmanned surface vessel.

[0054] Step S101: Establish the dynamic model of the unmanned surface vessel (USV) using the Euler-Lagrange dynamics equations. Specifically, the USV structural diagram is shown below. Figure 2 As shown, the ground coordinate system X is set. e O e Y e and ship hull coordinate system X v O v Y v Define x s and y s The unmanned surface vessel X e Axis, Y e The position state along the axis, φ is the yaw angle of the unmanned surface vessel; correspondingly, u is defined. s ,v s ,r s For unmanned surface vessel X v Axis, Y vThe velocity state along the axial direction and the yaw rate. Using the Euler-Lagrange dynamics equations, the dynamic model of the unmanned surface vessel is established as follows:

[0055]

[0056] Where, x1:=(x s ,y s ,φ) T and v:=(u s ,v s ,r s ) T Let represent the position and velocity state vectors of the unmanned surface vessel, respectively. := is the definition symbol, and T indicates the transpose of the vectors. is the rotation matrix for transforming variables between the Earth's fixed inertial reference frame and the unmanned surface vessel's fixed reference frame; M is the positive definite symmetric inertial matrix; C(v) represents the Coriolis force and centrifugal force vectors; D(v) is the hydrodynamic damping parameter matrix; d is the unknown bounded external disturbance; τ is the generalized control force vector of the external input.

[0057] Step S102: Define the position and velocity state vectors of the unmanned surface vessel (USV), converting the USV's dynamic model into a second-order dynamic model. Specifically, define x2 = J(φ)v, and convert the USV's dynamic model into a second-order dynamic model, which can be expressed as a first-order position system dynamic model and a second-order velocity system dynamic model, as follows:

[0058]

[0059] Where y(t) is the position output of the unmanned surface vessel, and u(t) = J(φ)M -1 τ represents the control input, ω(t) = J(φ)M -1 d is a bounded external disturbance that satisfies It is an unknown constant. This is a nonlinear dynamic that can be represented by a continuous smooth function.

[0060] Step S2: Using a neural network to approximate the online nonlinear dynamics of the position and velocity states in the second-order dynamic model, design a neural network-based adaptive observer to estimate the first-order system position state and the second-order system velocity state under DoS attacks and external disturbances.

[0061] To accurately estimate the missing information in the system, this embodiment utilizes the online parameter adjustment capability of adaptive technology and the strong nonlinear dynamic approximation capability of neural network technology to design an observer based on an adaptive neural network. This achieves strong anti-interference capability and effectively estimates the missing information. Specifically, combining adaptive and neural network control techniques, the design of a neural network-based adaptive observer includes the following steps:

[0062] Step S201: Based on the constructed second-order dynamic model of the unmanned surface vessel (USV), the observed state of the USV's position and velocity is obtained. Combining this with the acquired actual position and velocity state of the USV, the observation error of the USV is constructed as follows:

[0063]

[0064] Where x = [x1, x2] T Let x1 and x2 be the actual state vector, representing the actual position state and the actual velocity state, respectively. It is the observed state vector. It is a location estimate. x s ,y s The estimated value of ψ It is a speed estimate. u s ,v s The estimated value of r; e1 is the position observation error, and e2 is the velocity observation error.

[0065] Step S202: The online nonlinear dynamic g(x(t)) of the position and velocity states in the approximation model using the radial basis function neural network in the neural network can be expressed as:

[0066] g(x|W)=W T Ψ(x(t));

[0067] Where W represents the ideal weights of the neural network, and Ψ(x(t)) is the Gaussian activation function, expressed as: Where c l It is the center of the activation function, b l is the width of the activation function, and exp represents the exponential function with the natural constant e as the base.

[0068] Furthermore, utilizing This represents the state vector x = [x1, x2]. T From the estimation, we can obtain:

[0069]

[0070] in, This represents the estimated value of the ideal weight W in the neural network.

[0071] Step S203: Introduce the observer perturbation compensation term and design an adaptive observer based on a neural network, which can be expressed as:

[0072]

[0073] in, It is an estimated output. Neural networks represent nonlinear dynamics to approximate dynamics. B = [0, I3] T Υ=[0,I3] T C = [I3,0] T I3 is a 3D identity matrix. K is the compensation term for the observer perturbation, designed as follows:

[0074]

[0075] in, S(0,∞) represents the normal operating time of the system, Ξ(0,∞) represents the time during which the system is attacked, η1 is a positive integer greater than 1, and η2 is a positive integer greater than 0. It is an unknown positive number. They are The estimated value is adaptively updated as follows:

[0076]

[0077] Where λ1, λ2, and λ3 are the positive constants of the adaptive weights. It is a normal number.

[0078] Step S3: Based on the preset expected trajectory and the actual and estimated position states, determine the actual and estimated position tracking error of the unmanned surface vessel (USV), thereby constructing a virtual controller for the first-order position system of the USV.

[0079] Step S301: Establish the first-order position system error model of the unmanned surface vessel. First, the desired, differentiable trajectory is defined as y. r (t):=[y r1 ,y r2 ,y r3 ] T Based on the set desired trajectory and the constructed dynamic model, the actual position tracking error of the unmanned surface vessel system is established as z1(t):=[z 1,1 ,z 1,2 ,z 1,3 ] T Defined as:

[0080] z1=yy r ;

[0081] Secondly, based on the observer and desired trajectory in step S203, the position tracking error of the unmanned surface vessel system under the presence / absence of DoS attack is established as follows:

[0082]

[0083] Step S302: Design the first-order virtual controller for the unmanned surface vessel (USV) system. Specifically, based on the estimated position tracking error of the USV obtained in the previous step, construct the first-order virtual controller for the USV's position system. Use the virtual controller to output a virtual control signal β1 to ensure the stability of the position error system.

[0084] The output virtual control signal is represented as follows:

[0085]

[0086] Where α > 0.5, It is the expected trajectory y r The derivative of .

[0087] Step S4: Based on the virtual control signal output by the virtual controller, design a nonlinear filter to obtain the trajectory tracking signal of the second-order velocity system. Then, combine the estimated velocity state to construct an adaptive controller for the second-order velocity system.

[0088] Step S401: Based on the virtual control signal output by the virtual controller, design a nonlinear filter to filter the virtual control signal and compensate for the filtering error, providing a reference trajectory for the second-order velocity system of the unmanned surface vessel. Here, let θ1 be the filtered virtual control vector, which can be expressed as:

[0089]

[0090] Where, α 1,2 It is a constant greater than 1, α 2,2 It is a constant greater than 0, satisfying (α) 1,2 -1)|q1| 2 ≥α 2,2 q1 = θ1 - β1 is the filtering error, with an initial value of θ1(0) = β1(0); The dynamics of θ1, that is, θ1 is based on This dynamic equation changes.

[0091] Furthermore, a second-order velocity system error model and safety control strategy for unmanned surface vessels are established.

[0092] Step S402: Based on the trajectory tracking signal of the second-order velocity system obtained in step S401 (i.e., the filtered virtual control vector θ1 mentioned above) and the estimated velocity state obtained in step S201, construct the estimated velocity tracking error of the unmanned surface vessel under a DoS attack, as follows:

[0093]

[0094] Step S403: Design a speed control strategy. Based on the speed tracking error estimated in step S402, and combining the adaptive observer (step S203) based on a neural network (step S202) and the nonlinear filter (step S401), an adaptive controller for the second-order speed system is designed according to the Lyapunov stability principle. The controller is as follows:

[0095]

[0096] Where α2 is a positive constant, The adaptive law is:

[0097]

[0098] Where l = 1, 2, ..., L, κ l It is a positive constant, r l It is an adaptive weight.

[0099] Step S6: Based on the virtual controller and adaptive controller, perform the desired trajectory tracking of the unmanned surface vessel under DoS attacks and disturbances.

[0100] As another implementation method, step S7, theoretical verification of the above scheme, namely: constructing a Lyapunov function, analyzing the unmanned surface vessel error system according to the Lyapunov stability theorem, and verifying the stability of the observation error and tracking error of the unmanned surface vessel system, as well as the adaptive controller and nonlinear filter, including the following steps:

[0101] Step S701: Theoretical verification of the observer.

[0102] The above observation error model e(t) can be rewritten as:

[0103]

[0104] Construct the Lyapunov function as follows:

[0105]

[0106] in,

[0107] Differentiating the above function yields:

[0108]

[0109] in, Satisfying ||ρ(t)||≤ρ 0 , ρ 0 >0 is a bounded constant.

[0110] exist In this case, substituting the parameter K from step S203, according to We can obtain:

[0111]

[0112] Where μ>0, By using the following inequality:

[0113]

[0114] The final result is:

[0115]

[0116] Step S702: Theoretical verification of the controller.

[0117] First, construct the Lyapunov function as follows:

[0118]

[0119] in, z1=yy r .

[0120] Taking the derivative of the above function, we get:

[0121]

[0122] The virtual control strategy in step S302 Substituting into the above equation, and considering both t∈S(0,∞) and t∈Ξ(0,∞), according to Young's inequality... available:

[0123]

[0124] Secondly, construct the Lyapunov function as follows:

[0125]

[0126] in,

[0127] Taking the derivative of the above function, we get:

[0128]

[0129] According to the definitions of q1 and θ1 in step S302, we can obtain:

[0130]

[0131] And when (α) 1,2 -1)||q1|| 2 ≥α2,2 When it is established, we can obtain:

[0132]

[0133] According to Young's inequality:

[0134]

[0135] Ultimately, we can obtain:

[0136]

[0137] in,

[0138] Select The final result of the above differentiation can be rewritten as:

[0139]

[0140] when hour, V(t)≤p.

[0141] According to Lyapunov's stability theorem, the controller can guarantee that the system output can track the reference signal and that all error signals can converge to a small region.

[0142] Preferably, to further verify the superiority of the proposed solution in this embodiment, the effectiveness of the method and system of the present invention is verified by actual operation using the computer simulation software MATLAB.

[0143] Specifically, this example uses the computer software MATLAB to verify the invented control method. The parameters selected for each step in the simulation are as follows, where the parameters for the unmanned surface vessel are:

[0144]

[0145] in,

[0146]

[0147] In the simulation, the desired trajectory is selected as y. r =[2×sin(2t)m,2cos(0.5t)m,0.5rad] T The parameters for the controller, observer, filter, and adaptive law involved in each step are selected as follows: μ = 0.4, α = 3, α² = 1, α… 1,2 =1.2, α 2,2 =0.01, eta1=1.2, eta2=0.01, λ k =0.2, κl =1, r l =0.2, k=1,2,3, l=1,2,3,4,5. The external disturbance is set as d(t)=-0.1×cos(0.2t)+0.2. The DoS attack occurs within the interval Ξ(τ,t):=[7,7.5]∪[13,13.7]∪[24,24.5]∪[27,27.7]∪[35,35.5].

[0148] like Figure 3 , Figure 4 As shown, the proposed safety control method enables the unmanned surface vessel to track a predetermined desired trajectory and operate stably under DoS attacks and external disturbances; as Figure 5 , Figure 6 As shown, the designed safety controller and filter can ensure stable operation of the unmanned surface vessel's speed status and filtered signal under DoS attacks and external disturbances; for example... Figure 7 , Figure 8 As shown, the designed neural network-based adaptive observer can estimate the position and velocity information of the unmanned surface vessel (USV) relatively well and has good robustness to external disturbances; for example... Figure 9 , Figure 10 As shown, the adaptive parameter estimation of the unmanned surface vessel (USV) enables the system to have strong adaptability to disturbances, and the estimated parameter values ​​can also change with the state without diverging; for example... Figure 11 , Figure 12 As shown, the virtual control input and actual control input of the unmanned surface vessel can adjust the control input to compensate for the impact of DoS attacks and external disturbances.

[0149] This embodiment addresses the safety control problem of unmanned surface vessel (USV) systems under DoS attacks and external interference. It proposes an adaptive observation and control method based on neural networks. This method approximates the state and designs auxiliary reference signals by constructing an observer based on an adaptive neural network and a nonlinear filter. Based on the signals from the observer and filter, a safety control strategy based on neural networks and backstepping control is proposed. Furthermore, using Lyapunov stability theory, the safety boundedness result for the USV system is obtained. Based on this, the effectiveness of the designed adaptive neural network-based observation and safety control strategy is verified using a class of USV systems as an example.

[0150] Example 2

[0151] This embodiment provides a safety control system for unmanned surface vessels based on an adaptive neural network observer, including:

[0152] The data acquisition and model building module is used to construct a second-order dynamic model of the unmanned surface vessel using dynamic equations, and to obtain the actual position and velocity state of the unmanned surface vessel.

[0153] The state estimation module is used to approximate the online nonlinear dynamics of the position and velocity states in the second-order dynamic model using a neural network, and to design a neural network-based adaptive observer to estimate the first-order system position state and the second-order system velocity state under DoS attacks and external disturbances.

[0154] The virtual controller construction module is used to determine the actual and estimated position tracking error of the unmanned surface vessel (USV) based on the preset expected trajectory and the actual and estimated position states, thereby constructing the virtual controller of the USV's first-order position system.

[0155] The adaptive controller construction module is used to design a nonlinear filter based on the virtual control signal output by the virtual controller, obtain the trajectory tracking signal of the second-order velocity system, and then construct the adaptive controller of the second-order velocity system by combining the estimated velocity state.

[0156] The tracking control module is used to track the desired trajectory of the unmanned surface vessel under DoS attacks and disturbances, based on a virtual controller and an adaptive controller.

[0157] Example 3

[0158] This embodiment provides an electronic device, including: a memory for storing executable instructions; and a processor for executing the executable instructions stored in the memory to implement the method provided in this embodiment.

[0159] Example 4

[0160] This embodiment also provides a computer-readable storage medium storing executable instructions, which, when executed by a processor, will cause the processor to execute the method described above in this embodiment.

[0161] Example 5

[0162] This embodiment provides a computer program product including executable instructions, which are computer instructions; the executable instructions are stored in a computer-readable storage medium. When the processor of an electronic device reads the executable instructions from the computer-readable storage medium and executes the executable instructions, the electronic device performs the method described in this embodiment.

[0163] The steps and methods involved in Embodiments 2 to 5 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0164] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0165] The above description is only a preferred embodiment of the present invention. Although the specific implementation of the present invention has been described in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that, based on the technical solution of the present invention, various modifications or variations that can be made by those skilled in the art without creative effort are still within the scope of protection of the present invention.

Claims

1. A safety control method for unmanned surface vessels based on an adaptive neural network observer, characterized in that, include: Using the dynamic equations, a second-order dynamic model of the unmanned surface vessel (USV) is constructed, and the actual position and velocity state of the USV are obtained. By utilizing the online nonlinear dynamics of position and velocity states in a second-order dynamic model to approximate the second-order dynamic model using neural networks, an adaptive observer based on neural networks is designed to estimate the position state of the first-order system and the velocity state of the second-order system under DoS attacks and external disturbances. Based on the preset expected trajectory and combined with the actual and estimated position states, the tracking error between the actual and estimated positions of the unmanned surface vessel is determined, thereby constructing a virtual controller for the first-order position system of the unmanned surface vessel. Based on the virtual control signal output by the virtual controller, a nonlinear filter is designed to obtain the trajectory tracking signal of the second-order velocity system. Then, combined with the estimated velocity state, an adaptive controller for the second-order velocity system is constructed. Based on virtual controllers and adaptive controllers, the desired trajectory of unmanned surface vessels is tracked under DoS attacks and disturbances; Based on the estimated position tracking error of the unmanned surface vessel (USV), a virtual controller for the USV's first-order position system is constructed. This virtual controller outputs a virtual control signal to ensure the stability of the position error system. The output virtual control signal is expressed as follows: ； in, , Represents the desired trajectory The derivative of This indicates the estimated position tracking error of the unmanned surface vessel under conditions of presence / absence of a DoS attack; Based on the virtual control signal output by the virtual controller, a nonlinear filter is designed to filter the virtual control signal and compensate for the filtering error, providing a reference trajectory for the second-order velocity system of the unmanned surface vessel. , represented as: ； in, It is a constant greater than 1. It is a constant greater than 0, satisfying , It is filtering error. It is the filtered virtual control vector. For virtual control signals, initial value .

2. The unmanned surface vessel safety control method based on an adaptive neural network observer as described in claim 1, characterized in that, The process of constructing the second-order dynamic model of the unmanned surface vessel is as follows: A dynamic model of the unmanned surface vessel is established using the Euler-Lagrange dynamics equations. Define the position and velocity state vectors of the unmanned surface vessel (USV) to transform the USV dynamic model into a second-order dynamic model; this model includes a first-order position system dynamic model and a second-order velocity system dynamic model.

3. The unmanned surface vessel safety control method based on an adaptive neural network observer as described in claim 1, characterized in that, An adaptive observer based on a neural network is designed by approximating the online nonlinear dynamics of position and velocity states in a radial basis function neural network model and introducing an observer perturbation compensation term.

4. The unmanned surface vessel safety control method based on an adaptive neural network observer as described in claim 1, characterized in that, Based on the trajectory tracking signal of the second-order velocity system and combined with the estimated velocity state, the estimated velocity tracking error of the unmanned surface vessel under a DoS attack is constructed as follows: ； in, It is a speed estimate. It is the filtered virtual control vector; Based on the estimated velocity tracking error, and combining a neural network-based adaptive observer and a nonlinear filter, an adaptive controller for the second-order velocity system is designed according to the Lyapunov stability principle: ； in, It is a positive number. It is a constant. The adaptive law is: ； in, , It is a positive constant. It is an adaptive weight. It is the observation state vector.

5. The unmanned surface vessel safety control method based on an adaptive neural network observer as described in claim 1, characterized in that, A Lyapunov function is constructed, and the stability of the observation error, tracking error, adaptive controller, and nonlinear filter of the unmanned surface vessel system is verified based on the Lyapunov stability theorem.

6. A safety control system for unmanned surface vessels based on an adaptive neural network observer, characterized in that, The unmanned surface vessel safety control method based on an adaptive neural network observer as described in any one of claims 1-5 includes: The data acquisition and model building module is used to construct a second-order dynamic model of the unmanned surface vessel using dynamic equations, and to obtain the actual position and velocity state of the unmanned surface vessel. The state estimation module is used to approximate the online nonlinear dynamics of the position and velocity states in the second-order dynamic model using a neural network, and to design a neural network-based adaptive observer to estimate the first-order system position state and the second-order system velocity state under DoS attacks and external disturbances. The virtual controller construction module is used to determine the actual and estimated position tracking error of the unmanned surface vessel (USV) based on the preset expected trajectory and the actual and estimated position states, thereby constructing the virtual controller of the USV's first-order position system. The adaptive controller construction module is used to design a nonlinear filter based on the virtual control signal output by the virtual controller, obtain the trajectory tracking signal of the second-order velocity system, and then construct the adaptive controller of the second-order velocity system by combining the estimated velocity state. The tracking control module is used to track the desired trajectory of the unmanned surface vessel under DoS attacks and disturbances, based on a virtual controller and an adaptive controller.

7. An electronic device, characterized in that, include: Memory, used to store executable instructions; The processor, when executing executable instructions stored in the memory, implements the unmanned surface vessel safety control method based on an adaptive neural network observer as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The device stores executable instructions that, when executed by a processor, implement the unmanned surface vessel safety control method based on an adaptive neural network observer as described in any one of claims 1-5.

9. A computer program product, characterized in that, The computer program product includes executable instructions stored in a computer-readable storage medium; When the processor of the electronic device reads the executable instructions from the computer-readable storage medium and executes the executable instructions, it implements the unmanned surface vessel safety control method based on an adaptive neural network observer as described in any one of claims 1-5.

Images