Preset performance model-free self-adaptive trajectory tracking method for double-propeller propulsion ship

Abstract

The invention discloses a preset performance model-free self-adaptive trajectory tracking method for a double-propeller propulsion ship, and relates to the technical field of ship trajectory tracking, and the method comprises the steps: generating an expected longitudinal speed and an expected turning head angular speed through a trajectory tracking guidance law module; defining a time-varying performance boundary function through a performance function design module; the original tracking error is converted into a bounded virtual error through an error conversion module; carrying out dynamic linearization processing on the nonlinear system by utilizing a pseudo Jacobian matrix through a tight format dynamic linearization module; estimating a pseudo Jacobian matrix in real time and setting a reset mechanism through a pseudo Jacobian matrix estimation module; and through the adaptive control law module, the control input is converted into the rotating speed of the propeller, and trajectory tracking control is realized. Through a data driving mode, the dependence of a traditional control method on an accurate system model is broken through, and accurate control on transient and steady-state performance of the system is realized in combination with a dynamic performance constraint mechanism.

Description

Technical Field

[0001] This invention relates to the field of ship trajectory tracking technology, and in particular to a model-free adaptive trajectory tracking method for a twin-propeller propulsion ship with preset performance. Background Technology

[0002] Against the backdrop of the continuous expansion and specialization of global maritime transport, twin-propeller propulsion vessels face significant technical challenges in trajectory tracking and control under complex sea conditions, mainly in the following four aspects:

[0003] 1) Twin-propeller propulsion ships are underactuated systems, and under environmental disturbances, twin-propeller propulsion ships are difficult to achieve accurate trajectory tracking control.

[0004] 2) Model Dependence: Many current control algorithms (such as linear quadratic regulators) typically rely on accurate hydrodynamic models. However, constructing hydrodynamic models is not only costly, but the hydrodynamic derivatives of ships often experience perturbations during actual navigation. This model dependency makes it difficult for algorithms to capture changes in the ship's dynamic characteristics in real time, leading to problems such as excessive overshoot and difficulty in quickly converging tracking errors. Therefore, designing robust and adaptable ship motion control algorithms under the condition of uncertainties in the hydrodynamic model has become a key challenge that urgently needs to be addressed.

[0005] 3) Insufficient Disturbance Suppression: Ships face complex environmental disturbances during actual navigation, including wind, waves, and currents. These disturbances are random and time-varying, making them difficult for traditional control algorithms to effectively suppress, resulting in a significant decrease in the accuracy and stability of ship motion control. Especially in high-precision navigation or complex mission scenarios, insufficient disturbance suppression capabilities significantly increase the risk of control failure. Therefore, developing control strategies with strong disturbance suppression capabilities is one of the key directions for improving ship motion control performance.

[0006] 4) Real-time limitations: With the development of artificial intelligence technology, learning-based control strategies (such as deep reinforcement learning [3]) have shown certain potential in ship motion control. However, such methods usually require a lot of computing resources and time for model training and decision generation, and are limited by network latency and hardware performance. In actual navigation, ships need to respond to environmental changes in a very short time, and the computational delay of learning-based control strategies may lead to decision lag, which in turn may cause safety hazards.

[0007] Therefore, improving the computational efficiency and real-time performance of algorithms while ensuring control accuracy is a crucial problem that learning-based control strategies urgently need to solve in practical applications. The aforementioned technical bottlenecks make it difficult for twin-propeller-driven ships to achieve the desired trajectory tracking performance under complex sea conditions, severely hindering the engineering application of autonomous navigation systems for unmanned vessels. Therefore, it is imperative to overcome these technical bottlenecks to achieve high efficiency, robustness, and real-time performance in ship motion control.

[0008] Ship trajectory tracking control is primarily based on the Proportional-Integral-Derivative (PID) control framework, achieved through its combination with a guidance law. However, traditional PID control has significant limitations in handling disturbances from wind, waves, and currents, and exhibits poor adaptability to the ship's nonlinear dynamic characteristics, making it difficult to achieve high-precision trajectory tracking. With the development of control theory, a series of advanced control algorithms have been proposed and successfully applied to ship motion control. These algorithms can effectively solve the problems faced by PID control, such as sliding mode control (SMC), fuzzy logic control (FLC), and backstepping control (BSC). Compared to PID control, these control algorithms can improve control performance to some extent, but they rely on the accuracy of the ship's dynamics model. However, during actual navigation, the ship's hydrodynamic derivatives are perturbed, so model inaccuracies can severely affect the performance of these control algorithms. In addition, these control algorithms also have other bottlenecks in engineering applications: the tuning of controller parameters depends on the experience of engineers, lacks a systematic parameter tuning framework, resulting in long deployment cycles and high costs, and fixed control parameters are difficult to adapt to complex and changing working conditions.

[0009] To overcome these limitations, researchers have begun exploring data-driven model-free control methods. Among them, Model-Free Adaptive Control (MFAC), as an effective control method for discrete-time nonlinear systems with unknown models, has attracted widespread attention from academia and industry due to its simple control law, high computational efficiency, convenient parameter tuning, and strong robustness. MFAC estimates the system's input-output characteristics in real time using pseudo-partial derivatives (PPDs), thus achieving effective control of nonlinear systems without the need for pre-calculation of hydrodynamic derivatives or system identification. Meanwhile, Prescribed Performance Control (PPC) ensures that the system tracking error remains within the preset performance boundaries through dynamic constraint mechanisms. This control method not only achieves precise constraints on controller performance but also significantly improves system stability. Although MFAC and PPC each exhibit unique advantages, existing methods still have significant limitations: MFAC lacks systematic design in terms of performance constraints, while PPC's dependence on the system model limits its application scope. Therefore, a new control architecture that integrates the advantages of both is needed to break through the dependence of traditional control methods on system models, and at the same time achieve the synergistic optimization of autonomous disturbance rejection capability and dynamic performance constraints, thereby providing a new solution for high-performance control in complex dynamic environments.

[0010] Therefore, those skilled in the art are dedicated to developing a model-free adaptive trajectory tracking method for pre-defined performance of twin-propeller-driven ships. This method is particularly suitable for trajectory tracking scenarios with strict dynamic performance constraints. It can effectively suppress the impact of environmental disturbances on ship motion and ensure high accuracy and stability of trajectory tracking under complex sea conditions, providing reliable technical support for the autonomous navigation of twin-propeller-driven ships. Summary of the Invention

[0011] In view of the above-mentioned deficiencies of the prior art, the technical problem to be solved by the present invention is how to achieve trajectory tracking control of a twin-propeller propulsion ship under environmental disturbances.

[0012] To achieve the above objectives, the present invention provides a model-free adaptive trajectory tracking method for a twin-propeller propulsion vessel with preset performance, the method comprising the following steps:

[0013] Step 1: Generate the desired longitudinal velocity and desired turning angular velocity using the trajectory tracking guidance law module;

[0014] Step 2: Using the performance function design module, define time-varying performance boundary functions to constrain the transient and steady-state behavior of the original tracking error;

[0015] Step 3: Convert the original tracking error into a bounded virtual error using the error conversion module;

[0016] Step 4: Dynamically linearize the motion control system of the twin-propeller propulsion ship using the pseudo-Jacobi matrix through the compact dynamic linearization module.

[0017] Step 5: Estimate the pseudo-Jacobi matrix in real time using the pseudo-Jacobi matrix estimation module and set a reset mechanism;

[0018] Step 6: Using the adaptive control law module, based on the outputs of the tight-form dynamic linearization module and the pseudo-Jacobi matrix estimation module, design an adaptive control law to convert the control input into the port propeller speed and the starboard propeller speed, thereby achieving trajectory tracking control.

[0019] Furthermore, step 1 also includes:

[0020] Step 1.1: Obtain ship motion state parameters;

[0021] Step 1.2: Obtain target tracking points and calculate longitudinal tracking error and lateral tracking error;

[0022] Step 1.3: Calculate the desired longitudinal velocity and desired turning angular velocity.

[0023] Furthermore, the ship's motion parameters include: ship position, heading angle ψ, longitudinal velocity u, lateral velocity v, and turning angular velocity r; and the drift angle β = arctan(v / u), heading angle χ = ψ + β, and the ship's resultant velocity are calculated.

[0024] Furthermore, step 1.2 also includes:

[0025] Dynamic reference point (x) on the desired trajectory after continuous smoothing r (t),y r (t) is defined as the current target tracking point, where t is the time variable; the tangential angle χ of the path. r (t)=atan2[y r ′(t),x r ′(t)], It is the partial derivative of the coordinates of the parameterized path with respect to the time variable t; the reference turning angular velocity is expressed as... The reference resultant velocity determined by the target path is expressed as: Δ represents the forward sight distance in the trajectory tracking guidance law;

[0026] Longitudinal tracking error x between the ship and the target path e and lateral tracking error y e Represented as:

[0027]

[0028] In the formula, k represents k·t s The transformation from a continuous system to a discrete system, t s Sampling time.

[0029] Furthermore, step 1.3 also includes:

[0030] The desired heading angle is calculated based on the path tangential angle, longitudinal tracking error, lateral tracking error, and foresight distance, as shown in the following formula:

[0031]

[0032] In the formula, the integral term y of the lateral tracking error int Defined as follows:

[0033]

[0034] The desired angular velocity is calculated based on the desired heading angle, as shown in the following formula:

[0035]

[0036] In the formula, T g A positive parameter used to control the convergence rate of the first-pitch tracking error;

[0037] The desired longitudinal velocity is calculated based on the reference combined velocity, longitudinal tracking error, and lateral tracking error, as shown in the following formula:

[0038]

[0039] Furthermore, the time-varying performance boundary function in step 2 is:

[0040] ρ(k)=(1-v)ρ(k-1)+vρ(∞)

[0041] The preset performance conditions for the original tracking error are:

[0042] -aρ(k)<e(k)<bρ(k)

[0043] In the formula, a and b are positive parameters of the two adjustable boundaries; This represents the initial tracking error for a twin-propeller propulsion vessel.

[0044] Furthermore, step 3 also includes: converting the original tracking error into a bounded virtual error using an invertible nonlinear mapping function.

[0045]

[0046] Where -1 < ξ(k) < 1.

[0047] Furthermore, step 4 also includes:

[0048] The motion control system of a twin-propeller propulsion ship can be represented as an equivalent compact-form dynamic linearized model:

[0049]

[0050] In the formula, Δξ(k+1)=ξ(k+1)-ξ(k) is the change in conversion error; Δu(k)=u(k)-u(k-1) is the change in control input; The pseudo-Jacobi matrix at time k is expressed as:

[0051]

[0052] Furthermore, step 5 also includes:

[0053] Using a compact-form dynamic linearized model of the motion control system of a twin-propeller-driven ship, the following penalty function is given:

[0054]

[0055] In the formula, μ > 0 is a weighting factor used to penalize excessively large variations in the pseudo-Jacobi matrix estimate; let And by simplifying the matrix inversion, we obtain:

[0056]

[0057] In the formula, η∈(0,2] is the set step size factor; It is a pseudo-Jacobi matrix The estimated value;

[0058] Configure the reset mechanism as follows:

[0059] if or or

[0060]

[0061] if or or

[0062]

[0063] In the formula, a, b1, and b2 are all positive parameters set to ensure the effectiveness of the controller.

[0064] Furthermore, step 6 also includes:

[0065] Set the following penalty function to control the input:

[0066] J2(u(k))=||ξ(k+1)|| 2 +λ||u(k)-u(k-1)|| 2

[0067] Where λ > 0 is a weighting factor used to prevent excessive changes in control input; substituting the compact form dynamic linearization model of the motion control system of a twin-propeller propulsion ship into the above equation, let By simplifying the matrix inversion, we can obtain:

[0068]

[0069] In the formula, ρ∈(0,1] is the set step size factor;

[0070] The obtained control inputs are converted into port propeller speed and starboard propeller speed:

[0071]

[0072] Compared with the prior art, the present invention has at least the following beneficial technical effects:

[0073] 1. This invention adaptively converts the error into the desired longitudinal velocity and desired bow angular velocity based on the state of a twin-propeller-driven vessel, providing a foundation for accurate trajectory tracking control of twin-propeller-driven vessels. Simultaneously, this trajectory tracking decoupling design significantly reduces control complexity and facilitates controller design and optimization.

[0074] 2. This invention employs data-driven control to eliminate the reliance on accurate dynamic models in controller design, enabling accurate trajectory tracking control even without a dynamic model for twin-propeller propulsion vessels. Simultaneously, the online update strategy of pseudo-partial derivatives endows the controller with strong anti-interference capabilities. The calculation speed is very fast during trajectory tracking control and pre-training is not required, thus exhibiting strong real-time performance.

[0075] 3. This invention ensures that the system tracking error is always kept within the preset performance boundary through a dynamic constraint mechanism, and that key indicators such as maximum overshoot and convergence rate meet the actual engineering requirements; by flexibly adjusting the performance function parameters, the error convergence speed can be specified according to different working conditions, further improving the system's adaptability and robustness.

[0076] The following will further explain the concept, specific structure, and technical effects of the present invention in conjunction with the accompanying drawings, so as to fully understand the purpose, features, and effects of the present invention. Attached Figure Description

[0077] Figure 1 This is a schematic diagram of a preferred embodiment of the present invention;

[0078] Figure 2 This is a schematic diagram of the trajectory tracking guidance law of a preferred embodiment of the present invention.

[0079] Figure 3 This is a schematic diagram of the coordinate system and parameters of a preferred embodiment of the present invention;

[0080] Figure 4 This is a schematic diagram of the target path setting in a preferred embodiment of the present invention. Detailed Implementation

[0081] The following description, with reference to the accompanying drawings, illustrates several preferred embodiments of the present invention to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms, and the scope of protection of the present invention is not limited to the embodiments mentioned herein.

[0082] In the accompanying drawings, components with the same structure are indicated by the same numerical designation, and components with similar structures or functions are indicated by similar numerical designations. The dimensions and thicknesses of each component shown in the drawings are arbitrary, and the present invention does not limit the dimensions and thicknesses of each component. To make the illustrations clearer, the thickness of some components has been appropriately exaggerated in the drawings.

[0083] This embodiment provides a model-free adaptive trajectory tracking method for a twin-propeller propulsion vessel with preset performance, such as... Figure 1 As shown, it includes the following steps:

[0084] Step 1: Through the trajectory tracking guidance law module, the desired longitudinal velocity and desired bow angle velocity are generated, thereby decoupling the trajectory tracking problem of a twin-propeller propulsion ship into tracking control of the desired longitudinal velocity and desired bow angle velocity, such as... Figure 2 As shown, it includes the following steps:

[0085] Step 1.1: Obtain ship motion state parameters.

[0086] To accurately characterize the three-degree-of-freedom maneuvering characteristics of a ship, such as Figure 3 As shown, two right-handed rectangular coordinate systems are established: the inertial coordinate system o0-x0y0z0 and the motion coordinate system o b -x b y b z bA ship's position and heading angle are described in an inertial coordinate system, while its velocity and turning angular velocity are described in a moving coordinate system. The inertial coordinate system, also known as the geodetic coordinate system, is fixed to the Earth's surface. b -x b y b The coordinate system coincides with the still water surface, with the origin o0 being a fixed point on the ground. The x0 axis points due north, the y0 axis points due east, and the z0 axis points vertically downwards as positive. This moving coordinate system, also known as the body-dependent coordinate system, is fixed within the ship, where o... b -x b y b The plane coincides with the still water surface, with the origin at o. b Located at the center of the ship, x b The axis points towards the bow, y b The axis points to the starboard side of the hull, z b The axis pointing vertically downwards is considered positive. Figure 3 In the coordinate system, (x, y) represents the position of the ship's center in the inertial coordinate system, ψ is the ship's heading angle, u, v, and r represent the longitudinal velocity, lateral velocity, and turning angular velocity in the hull coordinate system, respectively, β = arctan(v / u) is the drift angle, and χ = ψ + β is the heading angle. The values ​​represent the ship's resultant velocity, where n1 is the port propeller speed and n2 is the starboard propeller speed. In this embodiment, the ship's resultant velocity, heading angle, and position can be obtained through the Automatic Identification System (AIS), and the propeller speed can be obtained through a speed sensor.

[0087] Step 1.2: Obtain the target tracking point and calculate the longitudinal tracking error and lateral tracking error.

[0088] The trajectory tracking objective of a twin-propeller propulsion vessel is to ensure that the vessel accurately tracks a predetermined course at a specified time through control system output. This process requires real-time correction of the ship's spatial position and heading angle to match the desired trajectory, thereby achieving precise trajectory tracking control. For example... Figure 4 As shown, the dynamic reference point (x) on the desired trajectory after continuous smoothing. r (t),y r (t) is defined as the current target tracking point, where t is the time variable. The parameterized target trajectory can be obtained through a map and path tracking algorithms, and the tangent angle of the path... in It is the partial derivative of the coordinates of the parameterized path with respect to the time variable t; the reference turning angular velocity can be expressed as The reference resultant velocity determined by the target path can be expressed as: Δ represents the forward sight distance in the trajectory tracking guidance law.

[0089] Longitudinal tracking error x between the ship and the target path e and lateral tracking error y e It can be represented as:

[0090]

[0091] In the formula, k represents k·t s The transformation from a continuous system to a discrete system, t s Sampling time.

[0092] Step 1.3: Calculate the desired longitudinal velocity and desired turning angular velocity.

[0093] When a twin-propeller propulsion vessel is tracking its trajectory, external disturbances such as wind, waves, and currents can prevent it from accurately converging to the target path point. Borrowing the integral line-of-sight guidance method from path tracking, this approach is based on the path tangential angle χ. d Lateral tracking error y e The forward sight distance Δ is used to calculate the desired heading angle, transforming the tracking control problem into calculating the desired heading angle ψ. d The tracking problem is addressed by introducing the integral term y of the lateral tracking error. int To address the steady-state error in lateral tracking:

[0094]

[0095] In the formula, the integral term y of the lateral tracking error int Defined as follows:

[0096]

[0097] In the formula, γ is the integral gain of the integral term.

[0098] Meanwhile, trajectory tracking requires comprehensive consideration of controlling the ship's steering and speed, while path tracking tasks typically do not change the ship's speed and are relatively simple. Therefore, when applying the aforementioned line-of-sight guidance method to trajectory tracking control, certain improvements are needed:

[0099]

[0100] The improved formula described above incorporates the influence of longitudinal tracking error on the desired heading angle generated by the guidance law, ensuring that the ship can simultaneously achieve rapid convergence of both longitudinal and lateral tracking errors. Then, based on the desired heading angle obtained from equation (4), the desired turning angular velocity r is calculated as follows. d :

[0101]

[0102] In the formula, Tg A positive parameter used to control the convergence rate of the first-pitch tracking error.

[0103] The desired longitudinal velocity derived from the trajectory tracking guidance law is set as follows:

[0104]

[0105] Therefore, the trajectory tracking problem of a twin-propeller propulsion ship is transformed from a guidance law problem into a problem of tracking the desired longitudinal velocity and the desired turning angular velocity.

[0106] Step 2: Using the performance function design module, define time-varying performance boundary functions to constrain the transient and steady-state behavior of the tracking error.

[0107] To achieve preset performance control of a twin-propeller propulsion vessel, a performance function must first be established to define the allowable range of tracking error. The performance function ρ(·) needs to satisfy the following: the initial value ρ(0) must cover the initial error; it must decrease strictly monotonically with time, forcing the allowable error range to gradually shrink; as time approaches infinity, the ρ(∞) performance function must converge to a steady-state value, typically a small positive number or zero, ensuring the steady-state error of the control system is limited to a certain range; the performance function must be continuously differentiable to introduce derivative terms in the controller design, constructing conversion errors or analyzing stability; at any given time, the performance function must always be greater than 0 to avoid boundary failures. If ρ(∞), then the performance function must asymptotically approach zero but always remain non-negative. The preset performance function is designed according to the above rules as follows:

[0108] ρ(k)=(1-v)ρ(k-1)+vρ(∞) (8)

[0109] In classic preset performance control, the upper and lower bounds of the preset performance function are usually opposites. However, in actual control processes, the requirements of the control task on the original tracking error are asymmetrical. Therefore, the following preset performance conditions for the original tracking error are set:

[0110] -aρ(k)<e(k)<bρ(k) (9)

[0111] In the formula, a and b are the positive parameters of two adjustable boundaries; e(t) = [u d (k)-u(k),r d (k)-r(k)] Τ This represents the initial tracking error for a twin-propeller propulsion vessel.

[0112] Step 3: Using the error conversion module, the original tracking error is converted into a virtual error using a nonlinear mapping method;

[0113] The purpose of preset performance control is to ensure that the original tracking error is always kept within a predetermined boundary. In addition to constructing a preset performance function for the controlled system, error transformation is also required to limit the newly constructed error. In this embodiment, the original tracking error e(k) is transformed into a normalized bounded error signal ξ(k) through an invertible nonlinear mapping function, ensuring that the transformed error strictly satisfies -1 < ξ(k) < 1, thereby transforming the error constraint problem of the original system into a problem of controlling the transformed error.

[0114]

[0115] This technical solution effectively improves the transient response quality and steady-state tracking accuracy of the system, while reducing the complexity of controller design.

[0116] Step 4: The nonlinear twin-propeller propulsion ship motion control system is dynamically linearized using the compact form dynamic linearization module and the pseudo-Jacobi matrix.

[0117] Before performing tight-form dynamic linearization on the motion control system of a twin-propeller-driven ship, its control inputs are first processed to simplify the design of the subsequent controller and improve the control performance:

[0118]

[0119] The motion control system of a twin-propeller propulsion ship can be discretized into the following form:

[0120] ξ(k+1)=f(ξ(k),…,ξ(kn y ),u(k),…u(kn u (12)

[0121] In the formula, f(...): Represents an unknown function; n y and n u It is an unknown positive integer; the subscripts 'y' and 'u' represent the output and input, respectively.

[0122] For a twin-propeller propulsion ship motion control system, if f(...) has continuous partial derivatives for each variable component and satisfies the generalized Lipschitz condition, and ||Δu(k)||≠0 for all k, then the nonlinear system can be represented as an equivalent compact-form dynamic linearization model:

[0123]

[0124] In the formula, Δξ(k+1)=ξ(k+1)-ξ(k) is the change in conversion error; Δu(k)=u(k)-u(k-1) is the change in control input; The pseudo-Jacobi matrix at time k can be expressed as:

[0125]

[0126] Step 5: Estimate the pseudo-Jacobi matrix of the system in real time using the pseudo-Jacobi matrix estimation module and set a reset mechanism;

[0127] Using a compact-form dynamic linearized model of the motion control system of a twin-propeller-driven ship, the following penalty function is given:

[0128]

[0129] In the formula, μ > 0 is a weighting factor used to penalize excessively large variations in the pseudo-Jacobi matrix estimate. Let By simplifying the matrix inversion, we can obtain:

[0130]

[0131] In the formula, η∈(0,2] is the set step size factor; It is a pseudo-Jacobi matrix The estimated value.

[0132] Meanwhile, to improve the pseudo-Jacobi matrix estimation algorithm's ability to track time-varying parameters, a reset mechanism for the above algorithm is set up:

[0133]

[0134] In the formula, a, b1, and b2 are all positive parameters set to ensure the effectiveness of the controller.

[0135] Step 6: Using the adaptive control law module, design an adaptive control law based on the output of the above module to convert the control input into the rotational speed of the port and starboard propellers of the twin-propeller propulsion ship, thereby achieving trajectory tracking control.

[0136] Set the following penalty function to control the input:

[0137] J2(u(k))=||ξ(k+1)|| 2 +λ||u(k)-u(k-1)|| 2 (19)

[0138] Where λ>0 is a weighting factor used to prevent excessive changes in the control input. Substituting equation (13) into the above equation, let... By simplifying the matrix inversion, we can obtain:

[0139]

[0140] In the formula, ρ∈(0,1] is the step size factor set to make the control algorithm more general.

[0141] Finally, the obtained control inputs need to be converted into port propeller speed and starboard propeller speed:

[0142]

[0143] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A model-free adaptive trajectory tracking method for a twin-propeller propulsion vessel with preset performance, characterized in that, The method includes the following steps: Step 1: Generate the desired longitudinal velocity and desired turning angular velocity using the trajectory tracking guidance law module; Step 2: Using the performance function design module, define time-varying performance boundary functions to constrain the transient and steady-state behavior of the original tracking error; Step 3: Convert the original tracking error into a bounded virtual error using the error conversion module; Step 4: Dynamically linearize the motion control system of the twin-propeller propulsion ship using the pseudo-Jacobi matrix through the compact dynamic linearization module. Step 5: Estimate the pseudo-Jacobi matrix in real time using the pseudo-Jacobi matrix estimation module and set a reset mechanism; Step 6: Using the adaptive control law module, based on the outputs of the tight-form dynamic linearization module and the pseudo-Jacobi matrix estimation module, design an adaptive control law to convert the control input into the port propeller speed and the starboard propeller speed, thereby achieving trajectory tracking control.

2. The model-free adaptive trajectory tracking method for pre-defined performance of a twin-propeller propulsion vessel as described in claim 1, characterized in that, Step 1 further includes: Step 1.1: Obtain ship motion state parameters; Step 1.2: Obtain target tracking points and calculate longitudinal tracking error and lateral tracking error; Step 1.3: Calculate the desired longitudinal velocity and desired turning angular velocity.

3. The model-free adaptive trajectory tracking method for pre-defined performance of a twin-propeller propulsion vessel as described in claim 2, characterized in that, The ship's motion parameters include: ship position, heading angle ψ, longitudinal velocity u, lateral velocity v, and turning angular velocity r; and the drift angle β = arctan(v / u), heading angle χ = ψ + β, and the ship's resultant velocity are calculated.

4. The model-free adaptive trajectory tracking method for pre-defined performance of a twin-propeller propulsion vessel as described in claim 3, characterized in that, Step 1.2 further includes: Dynamic reference point (x) on the desired trajectory after continuous smoothing r (t),y r (t) is defined as the current target tracking point, where t is the time variable; the tangential angle χ of the path. r (t)=atan2[y r ′(t),x r ′(t)], It is the partial derivative of the coordinates of the parameterized path with respect to the time variable t; the reference turning angular velocity is expressed as... The reference resultant velocity determined by the target path is expressed as: Δ represents the forward sight distance in the trajectory tracking guidance law; Longitudinal tracking error x between the ship and the target path e and lateral tracking error y e Represented as: In the formula, k represents k·t s The transformation from a continuous system to a discrete system, t s Sampling time.

5. The model-free adaptive trajectory tracking method for pre-defined performance of a twin-propeller propulsion vessel as described in claim 4, characterized in that, Step 1.3 also includes: The desired heading angle is calculated based on the path tangential angle, longitudinal tracking error, lateral tracking error, and foresight distance, as shown in the following formula: In the formula, the integral term y of the lateral tracking error int Defined as follows: The desired angular velocity is calculated based on the desired heading angle, as shown in the following formula: In the formula, T g A positive parameter used to control the convergence rate of the first-pitch tracking error; The desired longitudinal velocity is calculated based on the reference combined velocity, longitudinal tracking error, and lateral tracking error, as shown in the following formula:

6. The model-free adaptive trajectory tracking method for pre-defined performance of a twin-propeller propulsion vessel as described in claim 5, characterized in that, The time-varying performance boundary function in step 2 is: ρ(k)=(1-v)ρ(k-1)+vρ(∞) The preset performance conditions for the original tracking error are: -aρ(k)<e(k)<bρ(k) In the formula, a and b are positive parameters of the two adjustable boundaries; This represents the initial tracking error for a twin-propeller propulsion vessel.

7. The model-free adaptive trajectory tracking method for pre-defined performance of a twin-propeller propulsion vessel as described in claim 6, characterized in that, Step 3 further includes: converting the original tracking error into a bounded virtual error using an invertible nonlinear mapping function. Where -1 < ξ(k) < 1.

8. The model-free adaptive trajectory tracking method for pre-defined performance of a twin-propeller propulsion vessel as described in claim 7, characterized in that, Step 4 also includes: The compact-form dynamic linearized model of the motion control system of a twin-propeller propulsion ship is as follows: In the formula, Δξ(k+1)=ξ(k+1)-ξ(k) is the change in conversion error; Δu(k)=u(k)-u(k-1) is the change in control input; The pseudo-Jacobi matrix at time k is expressed as:

9. The model-free adaptive trajectory tracking method for pre-defined performance of a twin-propeller propulsion vessel as described in claim 8, characterized in that, Step 5 further includes: Using a compact-form dynamic linearized model of the motion control system of a twin-propeller-driven ship, the following penalty function is given: In the formula, μ > 0 is a weighting factor used to penalize excessively large variations in the pseudo-Jacobi matrix estimate; let And by simplifying the matrix inversion, we obtain: In the formula, η∈(0,2] is the set step size factor; It is a pseudo-Jacobi matrix The estimated value; Configure the reset mechanism as follows: if or or if or or In the formula, a, b1, and b2 are all positive parameters set to ensure the effectiveness of the controller.

10. The model-free adaptive trajectory tracking method for a twin-propeller propulsion vessel with preset performance as described in claim 9, characterized in that, Step 6 also includes: Set the following penalty function to control the input: J2(u(k))=||ξ(k+1)|| 2 +λ||u(k)-u(k-1)|| 2 Where λ > 0 is a weighting factor used to prevent excessive changes in control input; substituting the compact form dynamic linearization model of the motion control system of a twin-propeller propulsion ship into the above equation, let By simplifying the matrix inversion, we can obtain: In the formula, ρ∈(0,1] is the set step size factor; The obtained control inputs are converted into port propeller speed and starboard propeller speed:

Images