Cooperative anti-swing control method and system for double ship cranes based on preset performance

Abstract

The invention provides a cooperative anti-swing control method and system for double marine cranes based on preset performance, and relates to the technical field of crane control. Comprising the following steps: establishing a marine crane dynamical model considering double-ship rolling disturbance by using a Lagrange method; determining a drivable variable, and establishing a tracking error of the drivable variable; introducing an exponential preset performance function, constraining the convergence rate, overshoot and steady-state value of the tracking error, and converting the constrained tracking error into an unconstrained variable by using a hyperbolic tangent function; a performance index function considering tracking errors and energy consumption is constructed, a Hamiltonian equation minimization performance index function is constructed, the Hamiltonian equation is solved through a reinforcement learning algorithm, and the optimal control rate is obtained. According to the invention, the convergence speed of the error of the system variable is limited in a preset interval, the steady-state and transient-state performance of error convergence is improved, and accurate control and anti-swing of the load are realized.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of crane control, and particularly relates to a double-ship crane cooperative sway control method and system based on preset performance. BACKGROUND

[0002] Offshore cranes are important offshore engineering equipment for performing offshore wind power installation, offshore platform construction, and underwater equipment launching and recovery operations. At present, most ship cranes still mainly perform single crane lifting operations. However, with the continuous increase in the size and weight of offshore equipment, the lifting capacity of a single crane has been difficult to meet the complex construction requirements, and it is necessary to use double-ship cranes to cooperatively perform lifting operations. In actual operations, the double-ship crane system is long-term affected by external disturbances such as ship motion and sea wind, which brings great challenges to the stability and safety of load lifting. Especially at the present stage, the ship crane is often operated by multiple workers, which not only reduces the load positioning accuracy, but also affects the overall operation efficiency. Therefore, it is necessary to design an efficient control system for the double-ship crane to accurately lift the load to the target position and effectively suppress the load swing, thereby further improving the safety and efficiency of offshore operations.

[0003] As a typical under-actuated system, the crane can effectively reduce cost, weight, and energy consumption due to its fewer control inputs than degrees of freedom. However, in the face of uncertainties and disturbances in complex marine environments, especially in double-ship cooperative operations, due to the increase in state variables, geometric constraints, and system coupling, traditional control methods are difficult to apply.

[0004] Many existing control methods based on nonlinearity, adaptivity, robustness, and learning are for control under known system parameters. When there are continuous unknown external disturbances and uncertain parameters in the system, many crane control algorithms cannot effectively achieve accurate positioning and swing suppression. SUMMARY

[0005] To overcome the deficiencies of the prior art, the present application provides a double-ship crane cooperative sway control method and system based on preset performance, which limits the convergence speed of the error of the system variable to a preset interval, improves the steady-state and transient-state performance of error convergence, and realizes accurate load control and anti-sway.

[0006] To achieve the above purpose, one or more embodiments of the present application provide the following technical solutions: The first aspect of the present application provides a double-ship crane cooperative sway control method based on preset performance.

[0007] The double-ship crane cooperative sway control method based on preset performance comprises the following steps: A ship crane dynamics model considering double ship roll disturbance is established using Lagrange method; The dynamics model is analyzed to determine the drivable variables therein, and tracking errors of the drivable variables are established; An exponential preset performance function is introduced to constrain the convergence speed, overshoot and steady-state value of the tracking errors, and the constrained tracking errors are converted into unconstrained variables using a hyperbolic tangent function; A performance index function considering the tracking errors and energy consumption is constructed, a Hamilton equation minimizing the performance index function is constructed, and the Hamilton equation is solved using a reinforcement learning algorithm to obtain an optimal control rate; Based on the optimal control rate, the double ship crane is subjected to collaborative anti-swing control.

[0008] The second aspect of the present application provides a double ship crane collaborative anti-swing control system based on preset performance.

[0009] The double ship crane collaborative anti-swing control system based on preset performance comprises: The dynamics modeling module is configured to establish a ship crane dynamics model considering double ship roll disturbance using Lagrange method; The tracking error establishing module is configured to analyze the dynamics model to determine the drivable variables therein, and establish tracking errors of the drivable variables; The tracking error constraint module is configured to introduce an exponential preset performance function to constrain the convergence speed, overshoot and steady-state value of the tracking errors, and convert the constrained tracking errors into unconstrained variables using a hyperbolic tangent function; The solving module is configured to construct a performance index function considering the tracking errors and energy consumption, construct a Hamilton equation minimizing the performance index function, and solve the Hamilton equation using a reinforcement learning algorithm to obtain an optimal control rate; The double ship crane optimal control module is configured to perform collaborative anti-swing control on the double ship crane based on the optimal control rate. The third aspect of the present application provides a computer readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps of the double ship crane collaborative anti-swing control method based on preset performance as described in the first aspect of the present application.

[0010] The fourth aspect of the present application provides an electronic device comprising a memory, a processor and a program stored on the memory and executable on the processor, wherein the processor executes the program to implement the steps of the double ship crane collaborative anti-swing control method based on preset performance as described in the first aspect of the present application.

[0011] The above one or more technical solutions have the following beneficial effects: (1) The adaptive anti-swing control algorithm based on preset performance is proposed, which limits the convergence speed of the error of system variables in a preset interval, improves the steady-state and transient performance of error convergence, realizes accurate load control and anti-swing, and verifies its effectiveness through strict mathematical analysis and simulation. First, the dynamic range of error convergence is set by means of preset performance function, and the limited tracking error is converted into an unconstrained variable by using hyperbolic tangent function; then, a comprehensive performance index function considering tracking error and energy consumption is constructed, and the optimal control law is solved by combining reinforcement learning strategy, and the adaptive anti-swing control algorithm is derived. Finally, the stability of the closed-loop system is proved by strictly theoretical derivation using Lyapunov function. The simulation test verifies the effectiveness of the method in the double-ship crane system, and the experimental results show that the control algorithm significantly improves the positioning accuracy of the load and the dynamic stability of the system.

[0012] (2) The invention uses Lagrange method to establish the dynamic model of the double-ship crane system, and through coordinate transformation, the external disturbance of the ship body is converted into internal variables of the system, which effectively simplifies the complexity of the dynamic model and provides convenience for the design of subsequent control algorithm.

[0013] (3) The double-ship crane control system of the invention introduces an exponential preset performance function to strictly constrain the convergence speed, overshoot and steady-state value of the tracking error, significantly improving the transient response and steady-state performance of the system. The limited tracking error is converted into an unconstrained variable by using hyperbolic tangent function.

[0014] (4) The invention designs a performance index function considering tracking error and energy consumption, and derives the optimal learning adaptive anti-swing control algorithm through reinforcement learning algorithm, effectively reducing the energy consumption in the control process, and realizing the balance between operation efficiency and energy consumption.

[0015] The advantages of the additional aspects of the invention will be partially given in the following description, partially will become obvious from the following description, or will be known by the practice of the invention. BRIEF DESCRIPTION OF DRAWINGS

[0016] The drawings accompanying the specification of the invention form a part of the specification and serve to further illustrate the illustrative embodiments of the invention and to explain the principles of the invention.

[0017] Figure 1 The method flowchart of example one.

[0018] Figure 2 The schematic diagram of the double-ship underactuated crane system of example one.

[0019] Figure 3 The control system block diagram of example one.

[0020] Figure 4 The results of the effectiveness verification experiment of the method of Example 1 and the conventional PD method are shown in the following figure.

[0021] Figure 5 The results of the method of Example 1 and the conventional PD method in response to external disturbances are shown in the following figure.

[0022] Figure 6 The results of the method of Example 1 and the conventional PD method in response to parameter uncertainty are shown in the following figure. DETAILED DESCRIPTION

[0023] It should be noted that the following detailed description is exemplary in nature and is intended to provide further description of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0024] It should be noted that the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application.

[0025] In the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0026] Overall inventive concept: For a dual-ship crane system, the present application designs a dual-ship crane adaptive anti-swing control algorithm based on preset performance. First, a dynamic model of the dual-ship crane is established; then, by introducing an exponential preset performance function, the tracking error convergence speed, overshoot and steady-state error of the system are strictly constrained; finally, a performance index function considering convergence error and energy consumption is proposed, and an optimal control law is derived based on reinforcement learning strategy, thereby ensuring high-precision positioning and fast and stable tracking performance of the dual-ship crane system.

[0027] The stability of the designed control algorithm is proved by using Lyapunov method. The experimental results show that the proposed control algorithm can quickly hoist the load to the target position, while reducing the swing and rotational motion of the load, and has strong robustness to parameter uncertainty and external disturbances.

[0028] Example 1 The present embodiment discloses a dual-ship crane cooperative anti-swing control method based on preset performance.

[0029] As shown in Figure 1 and Figure 3 The dual-ship crane cooperative anti-swing control method based on preset performance includes the following steps: A dynamic model of the ship crane considering dual-ship roll disturbance is established using Lagrange method; The driving variable in the dynamic model is determined by analyzing the dynamic model, and the tracking error of the driving variable is established; An exponential pre-specified performance function is introduced to constrain the convergence speed, overshoot and steady-state value of the tracking error, and the tracking error is converted into an unconstrained variable by using a hyperbolic tangent function. A performance index function considering the tracking error and energy consumption is constructed, a Hamilton equation is constructed to minimize the performance index function, and a reinforcement learning algorithm is used to solve the Hamilton equation to obtain the optimal control rate. Based on the optimal control rate, the collaborative anti-swing control of the double-ship crane is performed.

[0030] Overall, the technical points of the embodiment include: 1) The dynamic model of the double-ship crane cooperative operation considering the disturbance of the double-ship roll is established using the Lagrange method and analyzed, and the model includes two drivable state variables: the cantilever amplitude state of the double-ship; three non-driven state variables: the cable angle of the double-ship and the rotation angle of the load.

[0031] 2) For the double-ship crane system, the tracking error of the drivable variable (amplitude angle of the cantilever) is established, the convergence speed, overshoot and steady-state value of the tracking error are pre-constrained and limited by using an exponential pre-specified performance function, and then the constrained error variable is converted into an unconstrained state variable by using a hyperbolic function.

[0032] 3) Based on the converted unconstrained state variable, a filtering vector is constructed, and a performance index function of the double-ship crane system including the filtering vector and energy consumption is proposed. In order to minimize the performance index function, a Hamilton equation is constructed, a reinforcement learning algorithm is used to solve the Hamilton equation, and the optimal control rate is derived.

[0033] The method of the present application will be described in detail below.

[0034] (1) Dynamic model and analysis: Figure 2 is a schematic diagram of the double-ship crane structure in a two-dimensional environment, which is composed of cranes installed on two ships, and the two cranes cooperatively hoist the same load. and is a moving coordinate system with the ship body. is an inertial coordinate system, and its origin is fixed at A the center of rotation of the ship. It is assumed that the ship body only has roll motion, and the roll centers of the two ships are at the same height. The cable is regarded as a massless rod, and the flexibility and elasticity of the cable are not considered. The related physical parameters and definitions of the crane system are shown in Table 1.

[0035] Table 1 Parameters and variable definitions of the underactuated double-ship crane system

[0036] The dynamic model of the twin-ship crane system is established by the Lagrange method: (1) (2) (3) (4) (5) To simplify the study, the following simplified expressions are defined:

[0037] The twin-ship crane system has the following physical constraints: (6) (7) From the physical constraints, it can be seen that the state variables , , , , are not mutually opposed, and can be expressed by the other three variables: (8) where,

[0038] Based on the above analysis, the dynamic model of the twin-ship crane system is written in matrix-vector form as follows: (9) where, is the state variable of the system. , is the inertia matrix, the centripetal Coriolis matrix, the gravity vector, the control quantity and the external integrated disturbance. The specific expression of the system matrix is as follows: (10) where:

[0039]

[0040]

[0041]

[0042] Since the dual crane system is an under-actuated system, in order to facilitate the controller design, the system dynamics model is written in the form of the drivable and non-drivable subsystems by using the partial feedback linearization method: (11) (12) wherein, is the drivable state variable, is the non-drivable state variable. Further, the following is obtained: (13) wherein, (14) is the drivable variable, is the pitch angle of the two crane booms; represents the inertia matrix of the drivable dynamics system; represents the gravity vector of the drivable dynamics system; represents the Coriolis matrix of the drivable dynamics system; represents the disturbance integral term of the drivable dynamics system; represents the derivative of the drivable variable, i.e. the speed of the pitch of the two crane booms; represents the second derivative of the drivable variable, i.e. the acceleration of the pitch of the two crane booms.

[0043] (2) Control target The control target of the present application is as follows: 1) Precise positioning of the boom. Position the boom to the target position : (15) 2) Effectively suppress the swing of the load: (16) wherein, . represents the target position of the pitch of the two crane booms.

[0044] (3) Preset performance controller design Based on the preset performance function, a new coupling state variable vector is constructed to pre-constrain the convergence of the tracking error. Then, a nonlinear transformation function is introduced to map the constraint error into an unconstrained variable. Finally, based on the established unconstrained variable, a reinforcement learning algorithm is used to solve the optimal control term.

[0045] The dynamics equation is written in the following form: (17) where, , is the system output.

[0046] with defined as:

[0047] Next, the error signal of the system is designed : (18) In order to make the error signal bounded and guarantee its convergence speed, overshoot and other performance, the convergence of the error signal is limited in the following interval by using a pre-set performance function: (19) where, is an adjustable parameter. determines the performance of error convergence, is set as: (20) where, is a positive constant. is an exponential pre-set performance function; is the tracking error of the luffing angle of the two cranes' cantilevers; i = 1, 2.

[0048] Subsequently, in order to make the constrained tracking error into an unconstrained error, an auxiliary variable is proposed: (21) Therefore, based on equation (19), the error can be written as: (22) Therefore we can define as a strictly increasing smooth function: (23) where, Based on the above analysis we can get: (24) Taking the derivative of the above equation we get: (25) where,

[0049] Then, the following filtering vector is constructed : (26) where, , , and is a positive constant. Taking the derivative of the above equation gives: (27) where,

[0050] Since the disturbance is unobtainable in practice, in order to improve the control accuracy, the following radial basis function neural network is used for estimation and compensation: (28) where, is the input of the neural network, is the optimal network parameter, is the error under the optimal network parameter. is the Gaussian activation function: (29) where, is the center coefficient, is the width coefficient. The weight matrix of the neural network is generally unknown, so the actual output of the neural network is: (30) where, is the estimate of the optimal weight . The update rate of the neural network is designed as: (31) In the above equation, is the adjustable parameter, and the elements in the matrix are positive real numbers.

[0051] (4) Optimal controller design In order to solve the optimal control problem, a new performance index function is proposed based on the constructed new auxiliary variable (27): (32) where, , is the output of the actor neural network, is a diagonal matrix with positive real numbers on the diagonal. In order to solve the optimal performance function, the Hamilton equation is established: (33) where, is the partial derivative of the performance index function : . Therefore, the optimal index function satisfies the equation: (34) However, due to the partial differential property and inherent strong nonlinearity of equation (34), it is difficult to obtain an analytical solution of the HJB equation. Therefore, an actor-critic structure policy based on reinforcement learning is adopted to derive the optimal control term. A critic network is used to approximate the performance index function, i.e., (35) where , and denote the optimal weight matrix of the critic network, the activation function and the approximation error. The performance index function is approximated as The derivative of (36) where and are the gradient of the neural network activation function and the approximation error with respect to . Since the ideal weight vector is unknown, the actual output of the critic network can be expressed as: (37) where and are the estimated values of and . Substituting equation (37) into equation (34) gives the neural network approximation error: (38) where is the approximation error of the critic network parameters. To solve the HJB equation, the following objective function is established: (39) and the update rate of the critic network is proposed as: (40) where . An actor neural network is established to obtain the optimal control input, and the ideal actor network and the actual actor network are constructed as: (41) (42) where and are the ideal and actual optimal control inputs, and denote the ideal value and the estimate of the actor network weight matrix, is the activation function of the actor network, is the network approximation error. As for the critic network, in order to solve the optimal actor network parameters, the following objective function is constructed: (43) where, . Further, the update rate of the actor network is: (44) In summary, the adaptive controller is designed as: (45) The controller gain needs to satisfy the following condition: (46) At the same time, in order to facilitate subsequent stability analysis, the following assumptions are made: (47) (5) Stability analysis In order to prove the stability of the controller proposed in the present application, the Lyapunov function is designed as where, and are defined as follows: (48) Next, take the derivative of with respect to time to get: (49) Substitute the control law (45) and the update rate (31) to get: (50) Consider the following inequality: (51) Further, we get: (52) Subsequently, take the derivative of with respect to time to get: (53) According to formula (40) and (44), we can get: (54) (55) where, , , .

[0052] Bringing into formula (53) can be obtained: (56) Wherein, .

[0053] When the condition (47) is satisfied, the following results are obtained: (57) Wherein,

[0054] Then, by boundedness theorem, it is derived that the closed-loop signal is ultimately uniformly bounded, that is (58) We can get the error Converge to a small neighborhood around the equilibrium point. Based on Can be obtained .

[0055] Because there is: (59) We can get: (60) Therefore, we can get: (61) The maximum invariant set only contains the equilibrium point: (62) According to the Lasalle invariance principle, the system is asymptotically stable at the equilibrium point. Therefore, it can be obtained that the control algorithm proposed in the present application can guarantee that the double-ship crane transports the load to the target position and effectively suppresses the swing of the load, that is, the state variables of the closed-loop system will converge to the target value.

[0056] (6) Method verification and comparison In order to ensure the effectiveness of the method of the present application, the algorithm proposed in the present application is compared with the traditional PD control method, and the load is transported to the target position by using the two algorithms respectively. Figure 4 The experimental results of the comparison method between the proposed method and the optimal control are given, and the change process of the amplitude angle of the cantilever and the swing angle of the cable and the load is shown. It can be seen that the algorithm proposed in the present application can guarantee that the error of the drivable state variable converges to the target position quickly, and the convergence curve is limited in the preset range. It can be seen that the convergence speed of the algorithm proposed in the present application is faster than that of the traditional PD control algorithm, and the swing during transportation is smaller. Through the energy consumption curve, it can be seen that the algorithm proposed in the present application has lower energy consumption.

[0057] Figure 5 The comparison between the algorithm proposed in the present application and the traditional PD algorithm under external disturbance is shown. It can be seen that the deviation of the crane jib after the disturbance is smaller, and the swing of the cable and the load is also smaller. Therefore, the algorithm proposed in the present application has stronger robustness to external disturbance than the traditional algorithm.

[0058] Figure 6 The comparison between the algorithm proposed in the present application and the traditional PD algorithm under the condition of changing the system parameters and the initial position is shown. It can be seen that after changing the parameters, the algorithm proposed in the present application can still hoist and deliver the load to the target position faster and effectively suppress the swing of the cable and the load. Moreover, the algorithm proposed in the present application consumes less energy than the traditional PD algorithm, effectively saving the power resource.

[0059] The adaptive anti-swing control algorithm based on preset performance proposed in the present application can limit the convergence speed of the error of the system variable to a preset interval, improve the steady-state and transient-state performance of error convergence, realize accurate control of the load and anti-swing, and its effectiveness is verified through strict mathematical analysis and simulation. First, the dynamic range of error convergence is set by means of the preset performance function, and the limited tracking error is converted into an unconstrained variable by using the hyperbolic tangent function; then, a comprehensive performance index function considering the tracking error and energy consumption is constructed, and the optimal control law is solved by combining the reinforcement learning strategy, and the adaptive anti-swing control algorithm is derived. Finally, the stability of the closed-loop system is proved by strict theoretical derivation using Lyapunov function. The simulation test verifies the effectiveness of the method in the double-ship crane system, and the experimental results show that the control algorithm significantly improves the positioning accuracy of the load and the dynamic stability of the system.

[0060] Embodiment Two The embodiment discloses a double-ship crane cooperative anti-swing control system based on preset performance.

[0061] The double-ship crane cooperative anti-swing control system based on preset performance comprises: The dynamics modeling module is configured to establish a ship crane dynamics model considering double-ship roll disturbance by using the Lagrange method; The tracking error establishing module is configured to analyze the dynamics model, determine the drivable variables therein, and establish the tracking error of the drivable variables; The tracking error constraint module is configured to introduce an exponential preset performance function to constrain the convergence speed, overshoot and steady-state value of the tracking error, and convert the constrained tracking error into an unconstrained variable by using the hyperbolic tangent function; The solving module is configured to construct a performance index function considering tracking error and energy consumption, construct a Hamilton equation to minimize the performance index function, and solve the Hamilton equation by using a reinforcement learning algorithm to obtain an optimal control rate. The dual-ship crane optimal control module is configured to perform collaborative sway suppression control on the dual-ship crane based on the optimal control rate. Embodiment three An object of this embodiment is to provide a computer-readable storage medium.

[0062] A computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the steps in the dual-ship crane collaborative sway suppression control method based on preset performance according to Embodiment 1 of the present disclosure.

[0063] Embodiment four An object of this embodiment is to provide an electronic device.

[0064] An electronic device includes a memory, a processor, and a program stored on the memory and executable on the processor, and the processor implements the steps in the dual-ship crane collaborative sway suppression control method based on preset performance according to Embodiment 1 of the present disclosure when executing the program.

[0065] The steps and methods in the above embodiments two, three, and four correspond to Embodiment one, and the specific implementation can be seen from the relevant description of Embodiment one. The term “computer-readable storage medium” should be understood to include a single medium or multiple media of one or more instruction sets; it should also be understood to include any medium that can store, encode, or carry instruction sets for execution by a processor and cause the processor to perform any method in the present disclosure.

[0066] Those skilled in the art should understand that each module or step of the present disclosure described above can be implemented by a general computer device, and alternatively, they can be implemented by program code executable by a computing device, so that they can be stored in a storage device for execution by a computing device, or they can be respectively manufactured into individual integrated circuit modules, or a plurality of modules or steps among them can be manufactured into a single integrated circuit module. The present disclosure is not limited to any specific combination of hardware and software.

[0067] The above describes the specific embodiments of the present disclosure in conjunction with the accompanying drawings, but is not a limitation on the scope of protection of the present disclosure. Those skilled in the art should understand that various modifications or variations made by those skilled in the art on the basis of the technical solutions of the present disclosure without inventive labor are still within the scope of protection of the present disclosure.

Claims

1. A method for coordinated anti-sway control of two marine cranes based on preset performance, characterized in that, Includes the following steps: A dynamic model of a marine crane considering the rolling disturbance of two ships is established using the Lagrange method; Analyze the dynamic model, identify the driveable variables, and establish the tracking error of the driveable variables; An exponential preset performance function is introduced to constrain the convergence rate, overshoot, and steady-state value of the tracking error. The constrained tracking error is then converted into an unconstrained variable using the hyperbolic tangent function. Construct a performance index function that considers tracking error and energy consumption, build a performance index function that minimizes the Hamiltonian equation, and use a reinforcement learning algorithm to solve the Hamiltonian equation to obtain the optimal control law; Based on the optimal control law, coordinated anti-sway control is implemented for dual marine cranes.

2. The method for coordinated anti-sway control of dual marine cranes based on preset performance as described in claim 1, characterized in that, In the process of establishing a dynamic model for a marine crane that considers the rolling disturbance of two ships: Two cranes are installed on two separate ships, and the two cranes work together to lift the same load. Assume that the hulls only experience rolling motion, and that the rolling rotation centers of the two ships are at the same height; The cable is treated as a massless rod, without considering the flexibility and elasticity of the suspension rope; or, The dynamic model of the marine crane considering the rolling disturbance of two ships is as follows: As a driveable variable, The pitch angle of the two crane booms; The inertia matrix represents the inertia matrix of a driveable mechanical system; This represents the gravity vector of a driveable mechanical system. Represents the Coriolis matrix of a driveable mechanical system; This represents the integrated disturbance term of a driveable mechanical system; The derivative of the driving variable is the pitching speed of the two crane booms; This represents the second derivative of the driveable variable, i.e., the pitch acceleration of the two crane booms.

3. The method for coordinated anti-sway control of dual marine cranes based on preset performance as described in claim 1, characterized in that, An exponential pre-defined performance function is introduced to constrain the convergence rate, overshoot, and steady-state value of the tracking error, specifically including: ； ； in, These are adjustable parameters; A positive constant; Preset performance functions for exponential functions; Let i be the tracking error of the pitch angle of the two crane booms, i=1,2; t represents time t.

4. The method for coordinated anti-sway control of two marine cranes based on preset performance as described in claim 3, characterized in that, The constrained tracking error is transformed into an unconstrained variable using the hyperbolic tangent function, specifically including: Propose auxiliary variables : Write the error as: in, ; It is an increasing smooth function; This represents the auxiliary vector that has been established.

5. The method for coordinated anti-sway control of two marine cranes based on preset performance as described in claim 4, characterized in that, Construct a performance index function that considers tracking error and energy consumption, specifically as follows: in, Represents a performance metric function; , This is the output of the actor neural network. , It is a diagonal matrix, and all diagonal elements are positive real numbers; For the constructed filter vector, Denotes the constructed filter vector; where , , It is a positive number.

6. The method for coordinated anti-sway control of two marine cranes based on preset performance as described in claim 5, characterized in that, It also includes calculating the derivative of the constructed filter vector. : ； in, ; Represents the coefficient matrix; Represents the nonlinear function terms of the system; This represents the nonlinear gravity term of the system; Indicates the system input; Indicates the integrated term caused by error; The inertia matrix represents the dynamic system; express , The diagonal matrix formed; Representation matrix The derivative; Represents the system variable error vector The derivative; express , Form a diagonal matrix; Representation matrix The derivative; ; ; And on The unattainable perturbations are estimated and compensated using a radial basis function neural network. The output of the radial basis function neural network is: in, This is an estimate of the optimal weights; This represents the output of a radial basis function neural network; Represents the Gaussian activation function; This is the input to the radial basis function neural network.

7. The method for coordinated anti-sway control of dual marine cranes based on preset performance as described in claim 6, characterized in that, The optimal control rate is: ； in, This represents the estimation of the actor network weight matrix. For the activation function of the actor network, This represents a diagonal matrix composed of integer control gains; Indicates the synovial coefficient; Represents a symbolic function; Indicates control input; The controller gain must satisfy the following condition: ； in, Indicating neural network error The upper bound; Represents the coefficient matrix of the actor neural network The upper bound; express The upper bound; express The lower bound, For the actor neural network activation function; This represents the learning rate matrix of the critic network; This represents the learning rate matrix of the actor network; express The upper bound; express The upper bound of the gradient; express The lower bound, express The upper realm, , , ; This represents the coefficient matrix corresponding to the control input in the performance index function; The operator that retrieves the minimum eigenvector; This represents the operator for finding the largest eigenvector.

8. A dual-ship crane cooperative anti-sway control system based on preset performance, characterized in that, include: The dynamics modeling module is configured to: use the Lagrange method to establish a dynamics model of a marine crane that considers the rolling disturbance of two ships; The tracking error establishment module is configured to: analyze the dynamic model, determine the driveable variables, and establish the tracking error of the driveable variables; The tracking error constraint module is configured to: introduce an exponential preset performance function to constrain the convergence rate, overshoot and steady-state value of the tracking error, and use the hyperbolic tangent function to convert the constrained tracking error into an unconstrained variable; The solution module is configured to: construct a performance index function that considers tracking error and energy consumption, construct a performance index function that minimizes the Hamiltonian equation, solve the Hamiltonian equation using a reinforcement learning algorithm, and obtain the optimal control law; The optimal control module for dual marine cranes is configured to perform coordinated anti-sway control on the dual marine cranes based on the optimal control law.

9. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the method for coordinated anti-sway control of dual marine cranes based on preset performance as described in any one of claims 1-7.

10. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the dual marine crane cooperative anti-sway control method based on preset performance as described in any one of claims 1-7.

Images