An AUV ditching control method based on event triggering and preset performance

Abstract

The application discloses an AUV landing control method based on event triggering and preset performance, and belongs to the technical field of ocean exploration, and comprises the following steps: constructing a sea current observer, estimating a sea current speed through the sea current observer; constructing a flow speed inversion model by taking the sea current speed as an influencing factor; constructing a radial basis function neural network; determining a Gaussian basis function and an optimal weight matrix corresponding to the Gaussian basis function through the radial basis function neural network under the condition of considering the total uncertainty of a system set composed of sea current disturbance, modeling uncertainty and propeller failure; constructing a finite time performance function; constructing a finite time preset performance controller according to the finite time performance function; introducing an incremental signal double-threshold event triggering control strategy into the finite time preset performance controller; and controlling the AUV landing process through the finite time preset performance controller. The control precision can be greatly improved, the convergence speed is fast, and the trajectory tracking control of the finite time preset performance is realized.

Description

Technical Field

[0001] This invention belongs to the field of marine exploration technology, specifically relating to an AUV docking control method based on event triggering and preset performance. Background Technology

[0002] Recently, with increasing emphasis on marine exploration and resource utilization, autonomous underwater vehicles (AUVs) have played an increasingly important role in this field. High-precision operations in near-shore environments are a significant research trend, such as comprehensive inspections and safety patrols of offshore engineering structures like wind power facilities. However, when approaching the seabed, AUVs face challenges such as poor maneuverability, weak close-range detection capabilities, and limited detection angles. These problems are difficult to solve using traditional AUV control methods. Traditional AUV control methods often cannot explicitly set the convergence time, resulting in slow convergence speeds, low control accuracy, and compromised dynamic performance. Summary of the Invention

[0003] The purpose of this invention is to provide an AUV docking control method based on event triggering and preset performance, which can solve the technical problems in the prior art, such as the inability to explicitly set the convergence time, slow convergence speed, low control accuracy, and inability to guarantee dynamic performance.

[0004] To solve the above-mentioned technical problems, the present invention is implemented as follows:

[0005] This invention provides an AUV docking control method based on event triggering and preset performance, comprising:

[0006] S101: Construct an ocean current observer and estimate ocean current velocity using the ocean current observer;

[0007] S102: Construct a current velocity inversion model using ocean current velocity as an influencing factor;

[0008] S103: Construct a radial basis function neural network to determine the Gaussian function and the optimal weight matrix corresponding to the Gaussian function, considering the total uncertainty of the system set consisting of ocean current disturbance, modeling uncertainty and thruster failure.

[0009] S104: Construct a finite-time performance function;

[0010] S105: Construct a finite-time preset performance controller based on the finite-time performance function;

[0011] S106: Introduce an incremental signal dual-threshold event-triggered control strategy in a finite-time preset performance controller;

[0012] S107: The AUV docking process is controlled by a finite-time preset performance controller.

[0013] The present invention has at least the following beneficial technical effects:

[0014] In this embodiment of the invention, a current velocity inversion model is constructed by using ocean current velocity as an influencing factor, and a radial basis function neural network is built. Considering the total uncertainty of the system set consisting of ocean current disturbances, modeling uncertainties, and thruster failures, the radial basis function neural network determines the Gaussian function and the optimal weight matrix corresponding to the Gaussian function, significantly improving control accuracy. Introducing an incremental signal dual-threshold event-triggered control strategy into the finite-time preset performance controller allows for explicit setting of the convergence time, resulting in fast convergence. This enables trajectory tracking control of finite-time preset performance, ensuring the dynamic performance of the AUV and facilitating AUV control practices under different operating conditions, as well as precise operations and complex multi-degree-of-freedom trajectory operations, especially the realization of vertical docking. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating an AUV docking control method based on event triggering and preset performance provided in an embodiment of the present invention;

[0016] Figure 2 This is a comparison chart of the new performance function provided in this embodiment of the invention and the traditional performance function (k is 0.1);

[0017] Figure 3 This is a comparison chart of the new performance function provided in this embodiment of the invention and the traditional performance function (k is 0.3);

[0018] Figure 4 This is a comparison chart of the new performance function provided by the embodiments of the present invention and the traditional performance function (k is 0.175);

[0019] Figure 5 This is a comparison graph of the curvature of the new performance function provided in this embodiment of the invention and the traditional performance function (k is 0.175).

[0020] The realization of the objective, functional characteristics and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0022] The AUV docking control method provided by the present invention will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.

[0023] Reference Figure 1 The diagram shows a flowchart of an AUV docking control method based on event triggering and preset performance provided by an embodiment of the present invention.

[0024] This invention provides an AUV (Autonomous Underwater Vehicle) docking control method based on event triggering and preset performance, primarily applied to disc-shaped AUVs. Disc-shaped AUVs, as a type of maneuverable AUV, differ from traditional torpedo-shaped underactuated AUVs. They can perform flexible movements in six degrees of freedom, such as low-drag swaying and full-circle turning in the horizontal plane, or hovering in a fixed position in the vertical plane. Their dynamic model is closer to the fully actuated model of ROVs than the underactuated model of traditional AUVs, but lacks the mooring force term found in the ROV dynamic model; therefore, it is suitable to be represented by the Fossen six-degree-of-freedom model.

[0025] AUV docking control methods include:

[0026] S101: Construct an ocean current observer and estimate ocean current speed using the ocean current observer.

[0027] In one possible implementation, S101 specifically includes sub-steps S1011 and S1012:

[0028] S1011: Constructing an ocean current observer:

[0029]

[0030] in, For v c The estimated values ​​are: ω is an intermediate variable, λ is a positive design parameter, η is the position and attitude vector of the AUV, and J(η) is the transformation matrix.

[0031] S1022: Define the observation error of ocean current velocity The observation error converges over time, and the ocean current velocity v is measured using an ocean current observer. c Make an estimate.

[0032] S102: Construct a current velocity inversion model using ocean current velocity as an influencing factor.

[0033] It should be noted that when a disc-shaped AUV performs high-precision, complex trajectory operations, the ocean current velocity and its impact on the system may exhibit complex, time-varying characteristics, leading to a decrease in the accuracy of the acquired AUV velocity status. Therefore, to more accurately describe the impact of ocean currents on the AUV trajectory tracking system, this patent considers the velocity status of the disc-shaped AUV as a combination of its relative ocean current velocity and the ocean current velocity, while retaining the nonlinear term of ocean current disturbance in the dynamic equation.

[0034] In one possible implementation, S102 specifically includes sub-step S1021:

[0035] S1021: Constructing a current velocity inversion model using ocean current velocity as an influencing factor:

[0036]

[0037] Among them, v c M represents the ocean current speed. η =MJ -1 M is the mass inertia matrix, and J is the transformation matrix between the geodetic coordinate system and the attached coordinate system. η = [η Dx ,η Dy ,η Dz ,η Dφ ,η Dθ ,η Dψ ] T This represents the position and attitude vector of the AUV in the geodetic coordinate system. This represents the velocity and angular velocity of the AUV relative to the ocean current in the attached coordinate system. C RB Let C represent the Coriolis force and centripetal force matrix of the AUV rigid body. Aη =C A (v re )J -1 C A D represents the Coriolis force and centripetal force matrix representing the added mass of the AUV. η =D F (v re )J -1 D F This is the hydrodynamic damping matrix. η The force and moment vector provided for gravity and buoyancy. τ th This refers to the control force and torque generated by the propulsion system.

[0038] The transformation matrix J can be expressed as:

[0039]

[0040]

[0041] Since the acceleration of an AUV during vertical motion is limited, the change in ocean current velocity caused by changes in depth also has an upper limit; furthermore, at the same depth, the change in ocean current velocity near the seabed is relatively small. Based on these characteristics, the following reasonable assumptions are proposed:

[0042] Assumption 1: Ocean current velocity v c first derivative Bounded, that is, satisfied

[0043] The following observer is introduced here to compensate for the ocean current velocity v in the model. c :

[0044]

[0045] in For v c The estimated value is ω, where ω is an intermediate variable and λ is a positive design parameter.

[0046] Definition of observation error of ocean current velocity Using the above formula to estimate ocean current velocity, the observation error will converge over time.

[0047] The specific proof is given below:

[0048] The Lyapunov function is defined as follows:

[0049]

[0050] Find its derivative with respect to time.

[0051]

[0052] Where a0 is a positive constant.

[0053] Based on hypothesis 1, we can further obtain

[0054] If a suitable gain λ is chosen, it satisfies The error in flow velocity observation It will exhibit a convergent form.

[0055] Therefore, the error in ocean current observations is bounded. Let... It is a positive constant.

[0056] Furthermore, the thruster is a crucial component of the AUV and a major source of malfunctions. Therefore, the impact of thruster failures must be considered when controlling the AUV.

[0057] In one possible implementation, S102 further includes sub-steps S1022 to S1024:

[0058] S1022: The impact of thruster failure is represented by the thrust distribution matrix ΔB, and the actual control force and torque are adjusted to τ. th +Δτ:

[0059] τ th +Δτ=(B0-K th B)u=(B0+ΔB)u

[0060] Where B0 represents the nominal value of the thrust distribution matrix, u represents the control signal of the thruster, and K th It is a diagonal matrix whose element k thii ∈[0,1], representing the corresponding thruster fault process.

[0061] S1023: Taking the impact of thruster failure as an influencing factor, a flow velocity inversion model is constructed:

[0062]

[0063] Among them, C RBη0 C Aη0 D η0 and g η0 Representing C respectively RBη C Aη D η and g η The standard value,

[0064] F Δ This indicates the impact of environmental disturbances, thruster failures, and model uncertainties.

[0065] Among them, F Δ The expression is:

[0066]

[0067] Where, η r C represents the AUV's displacement relative to water. Aη η r +D η η r Representing the influence of ocean currents, Represents the state variables of the control system.

[0068] S1024: The dynamic equations of the AUV are represented using state vectors:

[0069]

[0070] This expression can be integrated as follows:

[0071]

[0072] in,

[0073] S103: Construct a radial basis function neural network (RBFNN) to determine the Gaussian function and the optimal weight matrix corresponding to the Gaussian function by considering the total uncertainty of the system set consisting of ocean current disturbance, modeling uncertainty and thruster failure.

[0074] In one possible implementation, S103 specifically includes sub-steps S1031 to S1034:

[0075] S1031: Determine the RBFNN estimate of the nonlinear term F, which consists of model uncertainties and thruster failures, over the interval Ω:

[0076]

[0077] in, The weight matrix h(x) represents the estimation of a radial basis function neural network. N )=[h1(x N ),h2(x N ),...,h j (x N ),...,h m (x N )] T ∈R m h j (x N ) represents the Gaussian function, and h is the Gaussian function. j (x N The form can be represented as:

[0078]

[0079] Among them, c j Let b represent the center vector of the j-th node. j >0 indicates the base width of the j-th node. In this study, the hidden layer has 9 nodes, and the input to the neural network is designed as follows: Therefore, the dimensions of each element are x. N ∈R 18 c j ∈R 18 c = [c1, c2, ..., c m ]∈R 18×9 ,h(x N )∈R 9 ,

[0080] S1032: Define estimation error Weighting error but

[0081]

[0082] S1033: The adaptive law for calculating the radial basis function neural network is:

[0083]

[0084] Where, τ i Let Γ = diag[τ1, τ2, ..., τ n Elements in the gain matrix, γ W Positive design parameters express The value of the i-th degree of freedom.

[0085] S1034: Through the adaptive law, the estimated weight matrix is ​​made... Gradually approaching the optimal weight matrix W * .

[0086] The specific proof is given below:

[0087] The Lyapunov function is defined as follows:

[0088]

[0089] V W Taking the first derivative with respect to time, we get:

[0090]

[0091] According to Lemma 1, for any and ε * >0, for The following inequalities hold true;

[0092]

[0093] It can be deduced that:

[0094]

[0095] It should also be noted that:

[0096]

[0097] Therefore, it can be rewritten as:

[0098]

[0099] Design a suitable gain to meet the requirements. At this point, the approximation error of the neural network's weight matrix will converge to a compact set.

[0100]

[0101] The estimated weight matrix of the neural network will approximate the optimal weight matrix, thus satisfying any approximation accuracy μ. * .set up It is a positive constant.

[0102] S104: Construct a finite-time performance function.

[0103] It should be noted that the traditional performance function is quite broad, and its form is as follows:

[0104] ρ(t)=(ρ0-ρ ∞ )exp(-kt)+ρ ∞

[0105] Where, ρ0, ρ ∞ k is a pre-given positive constant.

[0106] The six degrees of freedom of the AUV's motion trajectory are constrained by performance functions, which satisfy the following:

[0107] -ρ i (t)<e i (t)<ρ i (t)

[0108] Among them, e i (t)=x i -x di =η i -η di x di =η di The desired trajectory is defined under the i-th degree of freedom, e. i (t) defines the deviation between the actual trajectory and the expected value in the i-th degree of freedom. If the error variable e i The initial value of (t) satisfies -ρ i (0) < e i (0) < ρ i (0), then the six-degree-of-freedom trajectory vector η of AUH is strictly restricted to the performance boundary ±ρ. i (t) inside.

[0109] Furthermore, the parameter k in the performance function limits the minimum convergence rate of the tracking error, while ρ i∞ An upper bound is given on the allowable steady-state tracking error. Therefore, an appropriate performance function ρ needs to be designed. i The desired system error response can be obtained by (t).

[0110] The error transformation takes the form of:

[0111]

[0112]

[0113] The physical meaning of the relationship between the parameter k and the convergence rate in traditional performance functions is rather vague, and the convergence time cannot be controlled.

[0114] Therefore, to improve the dynamic response of AUV trajectory tracking, this paper proposes a new performance function. This novel performance function has high overall smoothness and can improve the dynamic response of AUV trajectory tracking by adjusting the parameter t. f The terminal settling time, i.e. the convergence terminal time, can be manually set, and the form of the performance function can be made more varied by changing the cutoff time and linear term parameters, making it easier to adjust the convergence rate.

[0115] In one possible implementation, S104 specifically includes sub-step S1041:

[0116] S1041: Constructing a novel finite-time performance function:

[0117]

[0118]

[0119] Where ρ0 and ρ ∞ Let κ be a pre-given positive constant, t be a constant, and t be time.

[0120] Here, we will compare it with the classic performance function with different convergence rate parameters, referring to... Figures 2 to 5 .

[0121] When the parameter k of the classical performance function is small (e.g.) Figure 2 As shown), the performance function's constraint on the error lacks speed; when k is a large value (such as...), Figure 3 As shown), this can easily lead to system errors exceeding the constraint boundaries in the initial stage due to the short action time of the nonlinear compensation system and control signals; when k is selected as a moderate value to make it similar to the designed new performance function (such as... Figure 4 As shown in the figure, it can be observed that the new performance function is smoother in the initial stage, more efficient in the convergence stage, and has a higher overall smoothness than the traditional form. Figure 5 (This demonstrates a comparison of the curvature of the two.)

[0122] S105: Construct a finite-time preset performance controller based on the finite-time performance function.

[0123] In one possible implementation, S105 specifically includes sub-steps S1051 to S1055:

[0124] S1051: Define the tracking error and related variables of a system state variable x1 of the AUV:

[0125] e1 = x1 - x 1d

[0126]

[0127] Where x1 represents the system state variable, and x2 is the first derivative of x1;

[0128] S1052: Constructing a virtual control law:

[0129]

[0130] Where k1 is the design parameter, Let x1 be the first derivative of the expected value of state x1.

[0131] S1053: Define the tracking error for another system state variable x2:

[0132] e² = x² - x 2d

[0133] S1054: Define the sliding surface based on the terminal sliding mode and finite time:

[0134] ξ i =e 2i +sig(ε 1i ) a i = 1, 2, ..., 6

[0135] Where, ξ i Let the matrix be [ξ1,ξ2,...,ξ] i ,...,ξ6] T The element in sig(ε) 1i ) a =sign(ε 1i )|ε 1i | a e 2i Let ε represent the error e2 under the i-th degree of freedom. 1i ε1 represents the transformation error under the i-th degree of freedom.

[0136] S1055: Trajectory tracking control input for a given finite-time preset performance controller:

[0137]

[0138] in, a∈(0,1), b∈(0,1), k2 is the design parameter, sig(ξ) i ) b =sign(ξ i )|ξ i | b .

[0139] For AUV unknown control direction trajectory tracking systems, in complex environments with drastic changes and simultaneous propeller failures, the controller provided in this invention can quickly stabilize the tracking error within a finite time, ensuring that the system meets the preset performance.

[0140] The following is a detailed proof that the system's sliding surface converges rapidly in a finite time:

[0141] The Lyapunov function is defined as follows:

[0142]

[0143] Find its derivative with respect to time:

[0144]

[0145] Note but

[0146]

[0147] Right now

[0148]

[0149] in

[0150] make Then equation (70) satisfies Lemma 3, if there exists a positive definite continuous function V(x):U→R satisfying Where λ1>0, λ2>0, m∈(0,1), The system is then said to be practically extended finite-time stable. The system state converges to the following set in finite time:

[0151]

[0152] Where 0 < θ0 < 1.

[0153] At this point, the system will converge to the set:

[0154]

[0155] Convergence time satisfies:

[0156]

[0157] in λ2 = 1.

[0158] Next, we will prove that the error signals are bounded when the sliding surface of the system tends to converge rapidly in a finite time.

[0159] The Lyapunov function is defined as follows:

[0160]

[0161] Find its derivative with respect to time:

[0162]

[0163] Where sig(ε1) a =[sig(ε 11 ) a ,...,sig(ε 1i ) a ,...,sig(ε 16 ) a ] T i = 1, 2, ..., 6. Therefore, the above formula can be rewritten as...

[0164]

[0165] make Then the transformation error ε1 and the ocean current observation error can be obtained. and neural network weight estimation error Convergence status:

[0166]

[0167] When the sliding surface S approaches 0, according to equation (63), the boundedness of errors e2 and ε2 depends on ε1. Combining equation (75), since ε1 is bounded and convergent, ε2 and e2 are also bounded.

[0168] S106: Introduce an incremental signal dual-threshold event triggering control strategy in a finite-time preset performance controller.

[0169] In one possible implementation, S106 specifically includes sub-steps S1061 and S1062:

[0170] S1061: Constructing the output update law for a finite-time preset performance controller:

[0171]

[0172]

[0173] zev (t)=u(t)-u * (t)

[0174] Where inf represents taking the minimum value, z ev (t) represents the difference between the control input u(t) at time t and the control output u*(t) at time tq, u * Refers to the AUV propulsion system in t q The output at time q = 0, 1, 2, ..., c k Indicates the event triggering thresholds, c1, c2, T * The design parameters are positive, and c1 > c2.

[0175] S1062: Transmit control signals, which can be represented as:

[0176]

[0177] Where q = 1, 2, ...

[0178] This event-triggered control strategy only requires inputting a binary increment. This allows for the updating of the actuator's control signals, saving data transmission costs.

[0179] Design parameters c1 and c2 represent event triggering thresholds, T * This represents the time threshold with design considerations. To ensure the dynamic performance of the system, when t∈(0,T) * In the initial control phase, the event trigger threshold should be selected as a relatively large threshold (i.e., c1) to ensure that the control signal effectively handles the initial tracking error. To guarantee the steady-state performance of the system, when the tracking error gradually stabilizes over time (i.e., t≥T), the tracking error should be controlled to stabilize. * The control signal should gradually converge to zero, so a smaller event trigger threshold (i.e., c2) should be designed.

[0180] The following is a detailed proof demonstrating that there is no Zeno phenomenon in data transmission over the controller-to-actuator network:

[0181] In t q ≤t<t q+1 For z ev Differentiate (t):

[0182]

[0183] According to (64), we can obtain Established, and make t q ≤t<t q+1 In the formula, ψ is a positive constant.

[0184] Note and Then there is

[0185]

[0186] This expression can be rewritten as follows:

[0187]

[0188] It can be seen that the control signal uses the minimum transmission interval The data is sent intermittently to the actuator, and there is no Zeno phenomenon during the data transmission.

[0189] Thus, this patent demonstrates that, under conditions of unknown ocean current velocity, environmental disturbances, modeling uncertainties, and the combined effects of thruster failure, the AUV trajectory tracking system, utilizing the ocean current observer, neural network weight adaptive law, finite-time performance function and control law, and event triggering mechanism designed in this patent, can achieve rapid and stable convergence within a finite time, and the system's steady-state and dynamic performance meet the pre-set performance requirements.

[0190] S107: The AUV docking process is controlled by a finite-time preset performance controller.

[0191] The present invention has at least the following beneficial technical effects:

[0192] In this embodiment of the invention, a current velocity inversion model is constructed by using ocean current velocity as an influencing factor, and a radial basis function neural network is built. Considering the total uncertainty of the system set consisting of ocean current disturbances, modeling uncertainties, and thruster failures, the radial basis function neural network determines the Gaussian function and the optimal weight matrix corresponding to the Gaussian function, significantly improving control accuracy. Introducing an incremental signal dual-threshold event-triggered control strategy into the finite-time preset performance controller allows for explicit setting of the convergence time, resulting in fast convergence. This enables trajectory tracking control of finite-time preset performance, ensuring the dynamic performance of the AUV and facilitating AUV control practices under different operating conditions, as well as precise operations and complex multi-degree-of-freedom trajectory operations, especially the realization of vertical docking.

[0193] Currently, the control algorithms used for AUV trajectory tracking mainly include traditional PID control, fuzzy control, sliding mode control, and adaptive control. This patent, while utilizing preset performance control to ensure high system precision and good dynamic performance, combines an event-triggered mechanism and a finite-time algorithm, balancing the system's convergence speed, accuracy, and practicality, thereby achieving high-precision operation under complex AUV trajectories.

[0194] This invention proposes a finite-time preset performance method that combines an event-triggered mechanism. It fully considers the impact of ocean current velocity and its resulting disturbances, designs an observer to compensate for ocean current velocity, establishes an AUV current velocity correlation model, and uses a neural network to approximate environmental disturbances, modeling uncertainties, and thruster failures. Finally, it utilizes finite-time preset performance control to achieve high-precision and rapid trajectory tracking of the AUV. Furthermore, this patent proposes a novel finite-time performance function with good smoothness and flexibility, and allows for the manual setting of the system's terminal stabilization time. In addition, the dual-threshold incremental event-triggered mechanism designed in this patent can change the control input solely based on binary incremental signals, and can employ different event trigger thresholds according to different operating conditions and trajectory tracking stages, which helps reduce signal transmission costs and actuator losses.

[0195] The above are merely embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A docking control method for an autonomous underwater vehicle (AUV) based on event triggering and preset performance, characterized in that, include: S101: Construct an ocean current observer and estimate the ocean current velocity using the ocean current observer; S102: Construct a velocity inversion model using ocean current velocity as an influencing factor; S103: Construct a radial basis function neural network to determine the Gaussian function and the optimal weight matrix corresponding to the Gaussian function by considering the total uncertainty of the system set consisting of ocean current disturbance, modeling uncertainty and thruster failure. S104: Construct a finite-time performance function; S105: Construct a finite-time preset performance controller based on the finite-time performance function; S106: In the finite-time preset performance controller, an incremental signal dual-threshold event triggering control strategy is introduced; S107: The AUV docking process is controlled by a finite-time preset performance controller; Specifically, S101 includes: S1011: Construct the aforementioned ocean current observer: in, for The estimated value, As an intermediate variable, Positive design parameters Here are the position and attitude vectors of the AUV. This is the transformation matrix; S1012: Define the observation error of the described ocean current velocity. The observation error converges over time, and the ocean current velocity is measured by the ocean current observer. Make an estimate; Specifically, S102 includes: S1021: Construct the velocity inversion model by taking ocean current velocity as an influencing factor: in, For ocean current velocity, , M Here is the mass inertia matrix. J This is the transformation matrix between the geodetic coordinate system and the body coordinate system; For AUV in the geodetic coordinate system Position and attitude vectors This represents the velocity and angular velocity of the AUV relative to the ocean current in the attached coordinate system; Represents the Coriolis force and centripetal force matrix of the AUV rigid body; , The matrix representing the Coriolis force and centripetal force of the added mass of the AUV; , This is the fluid dynamics damping matrix; The force and moment vectors provided for gravity and buoyancy; These are the control forces and torques generated by the thrusters; Wherein, the transformation matrix J It can be represented as: ; ; S1022: Through thrust distribution matrix To represent the impact of a thruster malfunction, the actual control force and torque are adjusted to... ; in, The nominal value representing the thrust distribution matrix. The control signal representing the thruster, It is a diagonal matrix whose element k thii ∈[0,1], representing the corresponding thruster fault process; S1023: Taking the impact of thruster failure as an influencing factor, construct the flow velocity inversion model: in, and Represent and The standard value, This indicates the impact of environmental disturbances, thruster failures, and model uncertainties; in, The expression is: in, This represents the AUV's response to water displacement. Representing the influence of ocean currents, Represents the state variables of the control system; S1024: The dynamic equations of the AUV are represented using state vectors: This expression can be integrated as follows: in, , ; Specifically, S103 includes: S1031: In the interval Above, the nonlinear terms consisting of model uncertainties and thruster failures are determined. F RBFNN estimate: in, The weight matrix represents the estimated value of the radial basis function neural network. , Gaussian function, Gaussian function The form can be expressed as: in, Indicates the first The center vector of each node Indicates the first The base width of each node is set; the hidden layer has 9 nodes, and the input to the neural network is designed as follows: Therefore, the dimensions of each element are as follows: ; S1032: Define estimation error Weighting error ,but ; S1033: The adaptive law of the radial basis function neural network is calculated as follows: in, for Elements in the gain matrix Positive design parameters express No. The value of each degree of freedom; S1034: Through the aforementioned adaptive law, the estimated weight matrix is... Gradually approaching the optimal weight matrix ; Specifically, S104 includes: S1041: Constructing a novel finite-time performance function: in, and For a pre-given positive constant, It is a constant. For time; Specifically, S105 includes: S1051: Define a system state variable for the AUV Tracking error and related variables: in, Represents system state variables. for The first derivative; S1052: Constructing a virtual control law: in, For design parameters, For state The first derivative of the expected value; S1053: Define another system state variable Tracking error: ； S1054: Define the sliding surface based on the terminal sliding mode and finite time: in, For matrix The elements in , Representing the Error under each degree of freedom , Representing the Transformation error under each degree of freedom ; S1055: Given the trajectory tracking control input of the finite-time preset performance controller: in, , , For design parameters, .

2. The AUV docking control method according to claim 1, characterized in that, S106 specifically includes: S1061: Construct the output update law of the finite-time preset performance controller: in, This indicates taking the minimum value. Indicates the current time control input and Control output at any time difference, Refers to the AUV propulsion system in Output at any moment , Indicates the event trigger threshold. The design parameters are positive, and ; S1062: Transmit control signal, the control signal can be expressed as: in, .

Images