Wind power unmanned ship path tracking control method based on active disturbance rejection linear quadratic regulation control

Abstract

The invention discloses a wind power unmanned ship path tracking control method based on active disturbance rejection linear quadratic regulation control. The method comprises the following steps: building kinematics and dynamics models based on a ship body physical structure and a system parameter identification method; according to the method, a linear quadratic regulator (LQR) is introduced into an active disturbance rejection control framework, and the strong disturbance compensation capability of the ADRC and the control accuracy of the LQR are combined; an unmanned ship guidance system is constructed based on the model, a dynamic curvature navigation control (DCNC) algorithm is adopted to dynamically adjust the navigational speed, and path tracking is ensured by controlling the course angle. The method for calculating the weight proportion of the LQR coefficient in real time overcomes the system instability caused by environment change in actual engineering application to a certain extent; the ADRC control strategy adopted by the invention effectively improves the robustness and optimality of the system.

Description

Technical Field

[0001] The invention relates to the field of automatic control, and in particular to a path tracking control method of a wind-powered unmanned ship based on ADLQRC. Background Art

[0002] As an automated navigation tool that relies on wind power for propulsion, wind-powered unmanned ships have shown great application potential in the fields of marine development, environmental protection, scientific research, and logistics and transportation in recent years. Since they do not consume fossil fuels, electric-driven wind-powered unmanned ships have the advantages of being green and environmentally friendly, and having a long endurance, and are suitable for long-distance navigation missions. However, considering the problem of lateral displacement of the hull caused by complex external disturbances on the water surface, the path tracking control of wind-powered unmanned ships faces many challenges. The main difficulty is how unmanned ships can adapt to the complexity and uncertainty of the marine environment. For example, factors such as wind speed changes, wave disturbances, and water currents will have a significant impact on the motion trajectory of the hull.

[0003] Traditional control methods, such as PID control and fuzzy control, often have certain limitations when dealing with these complex external disturbances. Although these methods can provide relatively ideal control effects in a static environment, when the external environment changes significantly, they are not robust enough and have difficulty coping with the uncertainty caused by changes in system parameters or external disturbances. This leads to a decrease in the path tracking accuracy of unmanned ships in complex marine environments, and even loss of control.

[0004] In recent years, Auto Disturbance Rejection Control (ADRC) has gradually become an effective means to solve control problems in complex dynamic environments. ADRC maintains the stability of the system by observing the lumped disturbances in the system in real time, including internal uncertainties and external disturbances, and compensating for them. The advantage of the ADRC method is that it does not require an accurate mathematical model and can effectively deal with uncertainties and strong external disturbances. Therefore, it is very suitable for the stable control of unmanned ships in harsh environments. However, although the control strategy of ADRC has good disturbance suppression capabilities, it lacks accuracy, especially in path tracking tasks, which require higher accuracy to ensure that the unmanned ship strictly follows the predetermined path.

[0005] On the other hand, LQR is a classic optimal control method that can optimize the performance indicators of the system through state feedback and is widely used in various automatic control systems. LQR has significant advantages in accuracy by designing a weight matrix so that the controller can minimize the control energy consumption or the sum of squared errors while satisfying the system stability. However, LQR has high requirements for the modeling accuracy of the system and lacks the ability to effectively respond to strong disturbances in the environment. Its applicability is limited when facing complex nonlinear environments.

[0006] Therefore, combining the robustness of ADRC and the control accuracy of linear quadratic regulator, a path tracking control method for wind-powered unmanned ships based on ADRC linear quadratic regulation control (ADLQRC) is proposed, which will effectively solve the problems of system instability and insufficient control accuracy caused by complex environmental changes in the existing technology. By introducing the LQR control strategy in the ADRC framework, this method can not only compensate for external disturbances in real time, but also maintain high control accuracy in path tracking tasks. In addition, the method of real-time adjustment of the LQR weight coefficient can dynamically adapt to different marine environments, further improving the robustness of the system.

[0007] The research background of this invention fully considers the challenges in the current field of unmanned ship control, and is committed to overcoming the limitations of traditional control methods in complex environments through the organic combination of self-disturbance rejection control and linear quadratic regulator, thereby providing a more efficient and stable path tracking control solution. This technology has broad application prospects, especially in the fields of unmanned ship navigation control, marine intelligent transportation, environmental monitoring, etc. Summary of the invention

[0008] The purpose of the present invention is to provide a path tracking control method for a wind-powered unmanned ship based on self-disturbance rejection linear quadratic regulation control, so as to solve the problems of system instability and insufficient control accuracy caused by environmental changes in the prior art.

[0009] The method of the present invention comprises the following steps:

[0010] S1. Kinematic and dynamic models were built based on the physical structure of the hull and the system parameter identification method;

[0011] S2. Construct LQR controller and introduce the anti-disturbance control framework to realize the yaw and speed control of unmanned ship, and combine the strong disturbance compensation capability of ADRC with the control accuracy of LQR;

[0012] S3. Based on the model established in step S1, an unmanned ship guidance system is constructed. The DCNC algorithm (dynamic curvature navigation control) adjusts the heading angle of the unmanned ship through heading control so that it gradually approaches the target path. Then, a variable speed control law is designed based on the desired path curvature to dynamically adjust the speed.

[0013] Furthermore, the specific process of step 1 is as follows: in order to build the kinematic and dynamic models, the geographic coordinate system E (X, Y, Z) and the hull coordinate system O (x, y, z) are introduced. The geographic coordinate system is also called the north-east coordinate system. The hull coordinate system is fixed to the unmanned ship, and the origin is the center of the unmanned ship. The motion of the unmanned ship includes parallel movement and rotation. In the geographic coordinate system, parallel movement is represented by linear displacement x, y, z. is the corresponding derivative of the current position (x, y) of the four-tilt-rotor wind-powered unmanned ship QTUSV, and the rotation is represented by the yaw angle ψ, is the corresponding derivative;

[0014] The kinematic equation for the transformation between the geographic coordinate system and the ship coordinate system is:

[0015]

[0016] Among them, u represents the longitudinal velocity of the unmanned ship relative to the hull coordinate system, v represents the lateral velocity, and r represents the bow angular velocity;

[0017] Formula (2-1) is expressed in matrix form as follows:

[0018]

[0019] The three-degree-of-freedom dynamic equation of the unmanned ship is established in the hull coordinate system:

[0020]

[0021] Where: M is the inertial parameter matrix including the additional mass, C(v) is the centripetal force and Coriolis force coefficient matrix, D is the damping coefficient matrix, τ is the control input of the system, including the longitudinal propulsion force f and the torque T of the unmanned ship, that is, τ = [f0T] T , τ′ is a lumped disturbance, and it is impossible to establish a definite mathematical model, so the ESO is used to observe the disturbance in the following text; the following is the coefficient definition matrix of M, C(v), D:

[0022]

[0023] in, represents the hydrodynamic parameters, which are obtained by towing test and parameter identification; m represents the mass of the unmanned ship; x g is the longitudinal distance of the center of gravity of the ship in the hull coordinate system O.

[0024] In order to make the inertia parameter matrix M and the damping coefficient matrix D positive definite diagonal matrices, assume that

[0025] make Then the coefficient definition matrix of M, C(v), and D can be expressed as:

[0026]

[0027]

[0028] Therefore, the dynamic equation of the unmanned ship is expressed as follows:

[0029]

[0030] The resultant force f = f1 + f2 - f3 - f4, the torque T = f l, l represents the distance from the motor to the center of the unmanned boat, and f1, f2, f3, and f4 represent the propulsion force generated by each rotor respectively;

[0031] Assuming that the lift generated by the rotation of the rotor is perpendicular to the corresponding rotor, the magnitude of the lift is linearly related to the square of the rotor's rotation speed, that is,

[0032] f i =C T ω i 2 (i=1,2,3,4) 2-11

[0033] where f i is the propulsion force of the corresponding rotor, C T is the rotor lift coefficient, ω i is the corresponding rotor speed;

[0034] In the process of motor control, the PWM signal input to the electronic speed regulator determines the driving voltage of the motor, that is, the motor speed. For the convenience of calculation, the relationship between the actual speed ω and the PWM input W can be set as:

[0035] ω=KW 2-12

[0036] Where K is the conversion coefficient between speed and PWM.

[0037] Further, the specific process of step 2 is:

[0038] According to the dynamic model of the unmanned ship, the state variables and control inputs are defined; the cost function of LQR is determined, that is, the state error and control input are weighted. The LQR cost function is:

[0039]

[0040] Among them, x is the state vector, u is the control input, Q and R are weight matrices, which reflect the importance of state error and control input respectively;

[0041] According to the Riccati equation, the state feedback gain matrix K (K = R -1 B TP); K links each component of the state variable with the corresponding control input and generates the optimal control input through linear combination. By optimizing K, LQR can ensure that during the entire control process, the system response can quickly approach the target state without overusing the control input: thrust or rudder angle; therefore, the gain matrix K can help the controller achieve optimal control performance, that is, to achieve the best dynamic response and minimum control cost while ensuring system stability;

[0042] Furthermore, the ADRC framework is introduced. Its core is to estimate and compensate the external disturbance of the system in real time through the extended state observer (ESO), and combine the ESO output (i.e., the estimated disturbance and dynamics) with the output of the LQR controller to generate a control law containing a disturbance compensation term. In this way, the accuracy of the LQR controller is combined with the strong disturbance compensation capability of the ADRC, which can better cope with the disturbance of the wind-powered unmanned ship in a complex environment. The final form of the control input is:

[0043] u(t)=-Ke(t)+u ADRC 2-13

[0044] Where, e(t) is the error, K is the LQR gain matrix, and u ADRC It is the compensation term of the active disturbance rejection control part.

[0045] Further, the specific process of step 3 is:

[0046] In the actual movement of the unmanned ship, the sideslip angle will affect the path tracking. The DCNC algorithm can estimate and compensate for the error caused by the sideslip angle. This method directly observes the relative angle between the unmanned ship and the target path, and combines the ship's speed information to adjust the ship's heading angle;

[0047] α is the desired heading, α=β-tan -1 ((y e +k i y int ) / Δ), β is the path direction, Δ is the forward distance, both of which are constants during the guidance process, y e is the path deviation of the unmanned ship at the current moment, is the integral of the deviation within the specified time period; k i is the parameter of the integral term;

[0048] By adding integral and differential terms, the guidance formula is as follows:

[0049]

[0050] where k d is the parameter of the differential term, y eold is the path deviation of the unmanned ship at the last moment, is the sideslip angle of the unmanned boat, It can be regarded as the scaling parameter of the DCNC algorithm.

[0051] In the process of path tracking, the unmanned ship not only needs to adjust the heading angle, but also needs to dynamically adjust the speed according to the curvature of the path. In this system, the speed is determined by the curvature. Furthermore, by establishing the relationship between the speed and the expected path curvature, a variable speed control law is designed so that the speed of the unmanned ship can be adaptively adjusted according to the curvature, that is,

[0052]

[0053] Among them, v_desire is the desired speed of the unmanned ship, v_last1 is the speed at the previous moment of v_desire, v_last2 is the speed at the previous moment of v_last1, and dire_flag is the identifier of the direction of travel of the bow, that is, it is judged whether the bow is moving forward or the stern is moving forward at this time.

[0054] The beneficial effects of the present invention are:

[0055] By introducing the extended state observer (ESO) through the active disturbance rejection controller (ADRC), the internal and external disturbances can be estimated and compensated in real time, improving the robustness and stability of the system in complex sea conditions. Combined with the optimization capability of the linear quadratic regulator (LQR), high-precision control of the position and heading of the unmanned ship can be achieved, effectively reducing the path tracking error. On the basis of maintaining the stability of the system, the ADLQRC framework significantly improves the dynamic response performance of path tracking by rationally designing the state feedback control law and the target tracker (TD). By utilizing the characteristics of ADRC that does not rely on accurate models, it can adapt to the influence of model uncertainty or changes in the external environment during the operation of wind-powered unmanned ships and maintain the stability of path tracking. The DCNC algorithm ensures that the unmanned ship can stably track and optimize the navigation efficiency under complex paths by accurately adjusting the heading angle and dynamically adjusting the speed. By adaptively adjusting the speed according to the path curvature, the algorithm improves the stability of navigation and energy efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0057] Figure 1 Schematic diagram of QTUSV planar three-degree-of-freedom motion;

[0058] Figure 2This is a schematic diagram of the QTUSV body coordinate system;

[0059] Figure 3 ADLQRC controller system framework for the present invention;

[0060] Figure 4 This is the schematic diagram of the DCNC algorithm guidance law path tracking principle. DETAILED DESCRIPTION

[0061] The process of establishing the kinematic and dynamic model of step S1 of the present invention based on the physical structure of the hull and the system parameter identification method of the quad tiltrotor wind-powered unmanned surface vehicle (QTWUSV) is as follows:

[0062] Before establishing the USV mathematical model, the following assumptions are made:

[0063] (1) The unmanned boat sails along the horizontal plane in an ideal water body.

[0064] (2) The origin of the hull coordinate system coincides with the center of the unmanned boat, and the center of gravity is at the center of the unmanned boat float.

[0065] (3) Ignore the high-frequency part of the external interference force and only consider the low-frequency effect.

[0066] In order to establish the mathematical model of USV, the geographic coordinate system E (X, Y, Z) and the hull coordinate system O (x, y, z) are introduced. The geographic coordinate system is also called the north-east coordinate system. The hull coordinate system is fixed to the unmanned ship, and the origin is the center of the unmanned ship. The motion of the unmanned ship includes parallel movement and rotation. In the geographic coordinate system, parallel movement is represented by linear displacement x, y, z, and rotation is represented by yaw angle ψ.

[0067] like Figure 1 and Figure 2 As shown in the figure, the kinematic equation for the transformation between the geographic coordinate system and the hull coordinate system is:

[0068]

[0069] Among them, u represents the longitudinal velocity of the unmanned ship relative to the hull coordinate system, v represents the lateral velocity, and r represents the bow angular velocity.

[0070] Formula (2-1) is expressed in matrix form as follows:

[0071]

[0072] The three-degree-of-freedom dynamic equation of the unmanned ship is established in the hull coordinate system:

[0073]

[0074] Where: M is the inertial parameter matrix including the additional mass, C(v) is the centripetal force and Coriolis force coefficient matrix, D is the damping coefficient matrix, τ is the control input of the system, including the longitudinal propulsion force f and the torque T of the unmanned ship, that is, τ = [f0T] T , τ′ is a lumped disturbance, and it is impossible to establish a definite mathematical model, so the ESO is used to observe the disturbance in the following text. The following is the coefficient definition matrix of M, C(v), D:

[0075]

[0076]

[0077] in, represents the hydrodynamic parameters, which are obtained by towing test and parameter identification; m represents the mass of the unmanned ship; x g is the longitudinal distance of the center of gravity of the ship in the hull coordinate system O.

[0078] In order to make the inertia parameter matrix M and the damping coefficient matrix D positive definite diagonal matrices, assume that

[0079] make Then the coefficient definition matrix of M, C(v), and D can be expressed as:

[0080]

[0081] Therefore, the dynamic equation of the unmanned ship is expressed as follows:

[0082]

[0083] The resultant force f = f1 + f2 - f3 - f4, the torque T = f l, l represents the distance from the motor to the center of the unmanned boat. f1, f2, f3, f4 represent the propulsion force generated by each rotor.

[0084] Assuming that the lift generated by the rotation of the rotor is perpendicular to the corresponding rotor, the magnitude of the lift is linearly related to the square of the rotor's rotation speed, that is,

[0085] f i =C T ω i 2 (i=1,2,3,4)2-11

[0086] Among them, f i is the propulsion force of the corresponding rotor, C T is the rotor lift coefficient, ω i is the corresponding rotor speed;

[0087] In the process of motor control, the PWM signal input to the electronic speed regulator determines the driving voltage of the motor, that is, the motor speed. For the convenience of calculation, the relationship between the actual speed ω and the PWM input W can be set as:

[0088] ω=KW 2-12, where K is the conversion coefficient between speed and PWM.

[0089] Further, the ADLQRC controller system described in step S2 includes a TD module, an LQR site tracking design module and an ESO observation module;

[0090] For QTWUSV, in order to achieve stable site tracking and heading control, it is crucial to design an easy-to-implement, anti-interference controller. Active Disturbance Rejection Control (ADRC) is an adaptive control algorithm that does not rely on the precise model of the system and has strong robustness. The improved ADLQRC combines the LQR control algorithm on the basis of traditional ADRC, which can effectively improve the system's anti-interference and response speed. It mainly consists of three parts: TD, LQR, and ESO error feedback law. TD is used to control the output of the system to achieve the desired control effect, LQR is used to achieve site tracking, and ESO is used to estimate the state and interference of the system. The ADLQRC controller system framework is as follows Figure 3 shown.

[0091] TD Design:

[0092]

[0093] Where: v0 is the target input value, v1 is the target expectation after TD smoothing, v2 is the differential of the target expectation after TD smoothing, and v3 is the secondary differential of the target expectation after TD smoothing; r is the speed factor; h is the filter factor; fhan(·) is the fastest comprehensive control function, and the fhan function is defined as follows:

[0094]

[0095] LQR site tracking design:

[0096] For QTWUSV, in order to design a stable heading control, the speed difference control method will be adopted. The specific design of the position tracking control method is as follows:

[0097] Get status feedback

[0098] u=-Kx 3-3

[0099] Let the gain matrix K = R -1 B TP, where x is the state vector and u is the control input.

[0100] Solve it using the Riccati equation,

[0101] A T P+PA+Q-PBR -1 B T P=0 3-4

[0102] By solving the algebraic Riccati equation given the system matrices A, B, and weight matrices Q and R, we get the matrix P. By comparing it with the lumped disturbance z observed by the ESO r The yaw control value f2 is obtained by subtracting, that is,

[0103] f2=uz r

[0104] ESO Design:

[0105] Furthermore, in order to track the internal state of the system and estimate the total disturbance inside and outside the system, the ESO is designed as follows:

[0106]

[0107] Where B(1) is the matrix coefficient of the control allocation matrix for the control quantity f2. r and z ψ is the expanded state quantity, that is, the total disturbance observed by ESO.

[0108] Furthermore, the DCNC algorithm is used to estimate and compensate the sideslip angle, and its path tracking principle diagram is as follows: Figure 4 shown.

[0109] α is the desired heading, α=β-tan -1 ((y e +k i y int ) / Δ). β is the path direction, Δ is the forward distance, both of which are constants during the guidance process. e is the path deviation of the unmanned ship at the current moment, is the integral of the deviation within the specified time period; k i is the parameter of the integral term.

[0110] By adding integral and differential terms, the guidance formula is as follows:

[0111]

[0112] where k d is the parameter of the differential term, y eold is the path deviation of the unmanned ship at the last moment, is the sideslip angle of the unmanned boat, It can be regarded as the scaling parameter of the DCNC algorithm.

[0113] Furthermore, in this system, the speed is determined by the curvature. Furthermore, by establishing the relationship between the speed and the desired path curvature, a variable speed control law is designed so that the speed of the unmanned ship can be adaptively adjusted according to the curvature, that is,

[0114]

[0115] Among them, v_desire is the desired speed of the unmanned ship, v_last1 is the speed at the previous moment of v_desire, v_last2 is the speed at the previous moment of v_last1, and dire_flag is the identifier of the direction of travel of the bow, that is, it is judged whether the bow is moving forward or the stern is moving forward at this time.

[0116] Furthermore, by using PID for speed control, the speed control value f1 can be obtained, that is,

[0117]

[0118] Among them, K p , K i , K d They are proportional coefficient, integral coefficient and differential coefficient respectively. The speed deviation is e(t) = v_desire-u.

[0119] Finally, according to the configuration of each motor and the mechanism constraints, the speed control value f1 output by the PID controller and the yaw control value f2 output by the LQR controller are controlled and distributed to meet the specific navigation requirements.

[0120] Compared with the prior art, the present invention has the following advantages:

[0121] 1) The method of real-time calculation of the weight ratio of the LQR coefficient adopted by the present invention overcomes the system instability caused by environmental changes in actual engineering applications to a great extent;

[0122] 2) The ADRC control strategy adopted in the present invention effectively improves the robustness and optimality of the system.

[0123] 3) The mode of combining LQR and ADRC adopted in the present invention not only brings into play the ability of ADRC adaptive disturbance suppression but also realizes the effect of LQR precise path tracking, making the system more adaptable to the environment and greatly improving the autonomy of QTWUSV in rivers, ponds, oceans and other fields.

[0124] 4) The DCNC algorithm adopted in the present invention effectively avoids the deviation caused by sideslip when turning by estimating and compensating the sideslip angle of the QTWUSV, thereby improving the heading accuracy.

[0125] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A path tracking control method for a wind-powered unmanned ship based on auto-disturbance rejection linear quadratic regulation control, characterized in that: The following steps are involved: S1. Kinematic and dynamic models were built based on the physical structure of the hull and the system parameter identification method; S2. Construct a linear quadratic regulator LQR controller and introduce the ADRC framework to achieve yaw and speed control of the unmanned ship, and combine the strong disturbance compensation capability of ADRC with the control accuracy of the quadratic regulator LQR. S3. Based on the model established in step S1, an unmanned ship guidance system is constructed. The heading angle of the unmanned ship is adjusted through the DCNC algorithm heading control so that it gradually approaches the target path. Then, a variable speed control law is designed based on the expected path curvature to dynamically adjust the speed.

2. The path tracking control method of a wind-powered unmanned ship based on auto-disturbance rejection linear quadratic regulation control according to claim 1 is characterized in that: The specific process of step 1 is as follows: In order to build the kinematic and dynamic models, the geographic coordinate system E (X, Y, Z) and the hull coordinate system O (x, y, z) are introduced. The geographic coordinate system is also called the north-east coordinate system. The hull coordinate system is fixed to the unmanned ship, and the origin is the center of the unmanned ship. The movement of the unmanned ship includes parallel movement and rotation. In the geographic coordinate system, parallel movement is represented by linear displacement x, y, z. is the corresponding derivative of the current position (x, y) of the four-tilt-rotor wind-powered unmanned ship QTUSV, and the rotation is represented by the yaw angle ψ, is the corresponding derivative; The kinematic equation for the transformation between the geographic coordinate system and the ship coordinate system is: Among them, u represents the longitudinal velocity of the unmanned ship relative to the hull coordinate system, v represents the lateral velocity, and r represents the bow angular velocity; Formula (2-1) is expressed in matrix form as follows: The three-degree-of-freedom dynamic equation of the unmanned ship is established in the hull coordinate system: Where: M is the inertial parameter matrix including the additional mass, C(v) is the centripetal force and Coriolis force coefficient matrix, D is the damping coefficient matrix, τ is the control input of the system, including the longitudinal propulsion force f and the torque T of the unmanned ship, that is, τ = [f 0T] T , τ′ is a lumped disturbance, and it is impossible to establish a definite mathematical model, so the ESO is used to observe the disturbance in the following text; the following is the coefficient definition matrix of M, C(v), D: Among them, X u , Y v , Y r , N v , N r , I z represents the hydrodynamic parameters, which are obtained by towing test and parameter identification; m represents the mass of the unmanned ship; x g is the longitudinal distance of the center of gravity of the ship in the hull coordinate system O. In order to make the inertia parameter matrix M and the damping coefficient matrix D positive definite diagonal matrices, assume that Y r =0, N v =0, X u <0, Y v <0、 N r <0, make d 11 =-X u , d 22 =-Y v , d 33 =-N r , then the coefficient definition matrix of M, C(v), and D can be expressed as: Therefore, the dynamic equation of the unmanned ship is expressed as follows: The resultant force f = f1 + f2 - f3 - f4, the torque T = f l, l represents the distance from the motor to the center of the unmanned boat, and f1, f2, f3, and f4 represent the propulsion force generated by each rotor respectively; Assuming that the lift generated by the rotation of the rotor is perpendicular to the corresponding rotor, the magnitude of the lift is linearly related to the square of the rotor's rotation speed, that is, f i =C T ω i 2 (i=1,2,3,4)2-11 where f i is the propulsion force of the corresponding rotor, C T is the rotor lift coefficient, ω i is the corresponding rotor speed; In the process of motor control, the PWM signal input to the electronic speed regulator determines the driving voltage of the motor, that is, the motor speed. For the convenience of calculation, the relationship between the actual speed ω and the PWM input W can be set as: ω=KW 2-12 Where K is the conversion coefficient between speed and PWM.

3. The path tracking control method of a wind-powered unmanned ship based on auto-disturbance rejection linear quadratic regulation control according to claim 1 is characterized in that: The specific process of step 2 is: According to the dynamic model of the unmanned ship, the state variables and control inputs are defined; the cost function of LQR is determined, that is, the state error and control input are weighted. The LQR cost function is: Among them, x is the state vector, u is the control input, Q and R are weight matrices, which reflect the importance of state error and control input respectively; According to the Riccati equation, the state feedback gain matrix K (K = R -1 B T P); K links each component of the state variable with the corresponding control input and generates the optimal control input through linear combination. By optimizing K, LQR can ensure that during the entire control process, the system response can quickly approach the target state without overusing the control input: thrust or rudder angle; therefore, the gain matrix K can help the controller achieve optimal control performance, that is, to achieve the best dynamic response and minimum control cost while ensuring system stability; Furthermore, the ADRC framework is introduced. Its core is to estimate and compensate the external disturbance of the system in real time through the extended state observer (ESO), and combine the ESO output (i.e., the estimated disturbance and dynamics) with the output of the LQR controller to generate a control law containing a disturbance compensation term. In this way, the accuracy of the LQR controller is combined with the strong disturbance compensation capability of the ADRC, which can better cope with the disturbance of the wind-powered unmanned ship in a complex environment. The final form of the control input is: u(t)=-I(t)+u ADRC 2-13 Where, e(t) is the error, K is the LQR gain matrix, and u ADRC It is the compensation term of the active disturbance rejection control part.

4. The path tracking control method of a wind-powered unmanned ship based on auto-disturbance rejection linear quadratic regulation control according to claim 1 is characterized in that: The specific process of step 3 is: In the actual movement of the unmanned ship, the sideslip angle will affect the path tracking. The DCNC algorithm can estimate and compensate for the error caused by the sideslip angle. This method directly observes the relative angle between the unmanned ship and the target path, and combines the ship's speed information to adjust the ship's heading angle; α is the desired heading, α=β-tan -1 ((y e +k i y int ) / Δ), β is the path direction, Δ is the forward distance, both are constants during the guidance process, y e is the path deviation of the unmanned ship at the current moment, is the integral of the deviation within the specified time period; k i is the parameter of the integral term; By adding integral and differential terms, the guidance formula is as follows: where k d is the parameter of the differential term, y eold is the path deviation of the unmanned ship at the last moment, is the sideslip angle of the unmanned boat, It can be regarded as the scaling parameter of the DCNC algorithm. In the process of path tracking, the unmanned ship not only needs to adjust the heading angle, but also needs to dynamically adjust the speed according to the curvature of the path. In this system, the speed is determined by the curvature. Furthermore, by establishing the relationship between the speed and the expected path curvature, a variable speed control law is designed so that the speed of the unmanned ship can be adaptively adjusted according to the curvature, that is, in, v_desire is the desired speed of the unmanned ship, v_last1 is the speed at the previous moment of v_desire, v_last2 is the speed at the previous moment of v_last1, and dire_flag is the identifier of the direction of travel of the bow, that is, it is judged whether the bow is moving forward or the stern is moving forward at this time.

Images