Water surface unmanned ship control method and system under path tracking

Abstract

The invention discloses a water surface unmanned ship control method and system under path tracking, and relates to the technical field of attitude control, and the method comprises the steps: building a kinetic model of a water surface unmanned ship; establishing a path tracking error equation for describing a system error of the unmanned surface ship relative to an expected path under a Serret-Frenet coordinate system; constructing a path tracking controller based on a linear active disturbance rejection controller; inputting the system error and a water surface unmanned ship control signal into a fuzzy RBF neural network, and outputting an optimal control parameter of a path tracking controller by taking minimization of the system error as a target; updating a path tracking controller according to the optimized control parameters; and re-estimating external disturbance by using the updated path tracking controller, outputting a control signal for optimizing the water surface unmanned ship, and controlling the water surface unmanned ship to advance along an expected path. According to the invention, the attitude control precision under strong ocean current and changeable weather is improved, and the path tracking precision of the water surface unmanned ship is greatly improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of attitude control, and relates to a water unmanned ship control method and system, in particular to a water unmanned ship control method and system under path tracking. BACKGROUND

[0002] An unmanned surface vehicle (USV) is a type of ship that does not require human operation for navigation. It is usually equipped with a pre-programmed control system that allows it to navigate autonomously. These vessels are widely used in fields such as oceanographic research, environmental monitoring, and military reconnaissance, and have the ability to perform tasks such as autonomous navigation, data collection, and remote operation. They rely on the propulsion force of the ship itself or natural forces such as waves and wind to drive them, allowing them to perform tasks on the water surface for a long time.

[0003] The autonomous control capability of the unmanned surface vehicle is crucial because it often performs tasks in harsh marine environments. An effective control system can ensure that the unmanned ship maintains stable navigation under complex sea currents, wind waves, and other natural disturbances, achieving precise path tracking and navigation. In addition, the robustness and adaptability of the control system are crucial for ensuring that the unmanned surface vehicle can respond to unexpected situations, avoid collisions, or deviate from the path when performing tasks remotely.

[0004] However, the current control of unmanned surface vehicles in complex marine environments has low control accuracy, especially in strong sea currents or variable weather conditions. The control system cannot effectively respond to nonlinear disturbances in dynamic environments and internal system noise, resulting in large errors in heading control and path tracking. SUMMARY

[0005] In view of the above shortcomings of the prior art, the present application provides a water unmanned ship control method and system under path tracking, which can solve the technical problems of the prior art, i.e., the current water unmanned ship has low control accuracy in complex marine environments, especially in strong sea currents or variable weather conditions, and the control system cannot effectively respond to nonlinear disturbances in dynamic environments and internal system noise, resulting in large errors in heading control and path tracking.

[0006] The technical solution adopted by the present application is as follows:

[0007] A water unmanned ship control method under path tracking, comprising the following steps:

[0008] Constructing a dynamic model of the water unmanned ship;

[0009] Combining the dynamic model, a path tracking error equation is established in the Serret-Frenet coordinate system, which is used to quantify the system error between the actual navigation state of the water unmanned ship and the desired path.

[0010] constructing a path tracking controller based on a linear active disturbance rejection controller, estimating and compensating external disturbances faced by the water unmanned ship through the path tracking controller, and outputting an initial water unmanned ship control signal;

[0011] inputting the system error and the initial water unmanned ship control signal into a fuzzy RBF neural network to minimize the system error, and outputting optimized control parameters of the path tracking controller;

[0012] updating the path tracking controller according to the optimized control parameters, re-estimating external disturbances by using the updated path tracking controller, and outputting an optimized water unmanned ship control signal; and controlling the water unmanned ship to travel along the expected path based on the optimized water unmanned ship control signal.

[0013] Further, the specific steps of constructing the dynamic model of the water unmanned ship include:

[0014] determining the position vector and the velocity vector of the water unmanned ship in a preset coordinate system according to the Newton-Euler equation, and constructing the dynamic model in combination with the position vector and the velocity vector.

[0015] Further, the preset coordinate system includes a system coordinate system and a geostationary coordinate system;

[0016] the origin of the system coordinate system is the center of the water unmanned ship, the x-axis points to the forward direction of the water unmanned ship, the y-axis is perpendicular to the x-axis and points to the starboard of the water unmanned ship, and the z-axis is determined based on the x-axis and the y-axis through the right-hand rule;

[0017] the origin of the geostationary coordinate system is the center of the earth, the x-axis points to the geographic north pole, the y-axis points to the geographic east pole, and the z-axis is vertically downward.

[0018] Further, the system error includes a lateral error and a heading error; the lateral error is the vertical deviation of the water unmanned ship from the expected path; and the heading error is the angular deviation of the actual heading of the water unmanned ship from the corresponding heading of the expected path.

[0019] Further, the path tracking controller introduces a plurality of feedback control gains and a plurality of observer gains, estimates the heading, the heading rate of change and the total external disturbance of the water unmanned ship at the current time through the observer gains in real time, adjusts the control parameters in combination with the feedback control gains, so as to realize the compensation of the external disturbance.

[0020] Further, the fuzzy RBF neural network includes an input layer, a fuzzification layer, a fuzzy inference layer and an output layer connected in sequence, and the fuzzification layer and the output layer are directly connected.

[0021] Further, the input layer receives system error and initial water unmanned ship control signal, and transmits input data by using a linear activation function;

[0022] The fuzzification layer performs fuzzification processing on the input data, and calculates fuzzy quantities of the system error and the initial water unmanned ship control signal belonging to different fuzzy sets respectively;

[0023] The fuzzy inference layer performs a two-by-two product operation on the fuzzy quantities of different fuzzy sets corresponding to each input data, and outputs an inference result;

[0024] The output layer receives the inference result output by the fuzzy inference layer, and calculates an initial optimization control parameter based on a preset connection weight matrix.

[0025] Further, the fuzzy RBF neural network optimizes the connection weight matrix by introducing a learning momentum factor and a learning rate, calculates a performance index function according to the deviation between ideal output and actual output of the output layer under the current iteration number, and outputs the initial optimization control parameter as a final optimization control parameter when the value of the performance index function is less than a preset threshold. If not, the learning momentum factor and the learning rate are updated, the initial optimization control parameter is recalculated, and the optimization process is repeated.

[0026] Further, the optimization control parameter includes an observer bandwidth assumption value, an external disturbance influence estimation value and a feedback control gain.

[0027] A water unmanned ship control system under path tracking, comprising:

[0028] One or more processors;

[0029] A memory for storing one or more programs;

[0030] When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned water unmanned ship control method under path tracking.

[0031] Compared with the prior art, the beneficial effects of the present application are:

[0032] The application can accurately describe the motion characteristics of the water unmanned ship by constructing a kinetic model, and establish a path tracking error equation in combination with the Serret-Frenet coordinate system, thereby improving the accuracy of error estimation. Secondly, the linear active disturbance rejection controller is used for estimation and compensation of external disturbance, so that the water unmanned ship can effectively cope with nonlinear disturbance and internal noise in complex marine environment, and the accuracy of the heading control is significantly improved. The fuzzy RBF neural network is used for dynamic optimization of the controller, which can not only adjust the control parameters in real time, but also reduce the error caused by disturbance, so as to ensure that the water unmanned ship stably sails along the expected path. Further improve the control precision and enhance the robustness, especially in strong sea current and changeable weather, the performance is excellent, which greatly improves the path tracking accuracy of the water unmanned ship. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 is a flowchart of a water unmanned ship control method under path tracking provided by an embodiment of the application.

[0034] Figure 2 is a coordinate system diagram in the embodiment of the application.

[0035] Figure 3 is a schematic diagram of a water unmanned ship control system under path tracking in the embodiment of the application. DETAILED DESCRIPTION

[0036] The technical solutions of the application will be further clearly and specifically described in combination with the drawings and specific examples.

[0037] As shown in Figure 1 is a flowchart of a water unmanned ship control method under path tracking provided by an embodiment of the application.

[0038] The embodiment of the application provides a water unmanned ship control method under path tracking, which can include the following steps:

[0039] S1: establishing a kinetic model of the water unmanned ship.

[0040] The kinetic model refers to a mathematical description of the motion behavior of the water unmanned ship, which is usually expressed by equations to express the mechanical properties, speed, acceleration, direction and other factors of the ship. This model is usually based on physical laws, such as Newton-Euler equation, considering the mass, inertia, external force (such as water flow, wind force) and internal force (such as propulsion force) of the water unmanned ship, so as to predict the motion state of the unmanned ship in different environments.

[0041] It should be noted that establishing a dynamic model of the unmanned surface vessel provides an accurate mathematical basis for subsequent control strategies, ensuring that the control system can make reasonable adjustments based on the vessel's actual motion characteristics. This model provides a precise dynamic description for path tracking, which helps improve control accuracy and system response robustness in complex environments.

[0042] In one possible implementation, S1 specifically includes:

[0043] S11: Determine the position and velocity vectors of the unmanned surface vessel in the Earth-fixed coordinate system based on the Newton-Euler equations:

[0044]

[0045]

[0046]

[0047]

[0048]

[0049]

[0050] in, Indicates the course of the unmanned surface vessel. Represents the arctangent function. and Let represent the lateral and longitudinal velocities of the unmanned surface vessel in the Earth-fixed coordinate system, respectively. This represents the yaw angle of the unmanned surface vessel's floating body in the system coordinate system. This represents the yaw angle of the unmanned surface vessel in the system coordinate system. express and The deviation between Represents the velocity vector. and These represent the lateral and longitudinal velocities of the unmanned surface vessel relative to the ocean current in the system coordinate system, respectively, with the subscript T indicating transpose. This represents the roll angle of the unmanned surface vessel in the system coordinate system. This represents the pitch angle of the underwater drone in the system coordinate system. ) represents the position vector. Let represent the heading angle of the unmanned surface vessel, p represent the angular depth of the unmanned surface vessel around the x-axis, and q represent the angular velocity of the unmanned surface vessel around the y-axis. This represents the yaw angle and angular velocity of the unmanned surface vessel. This represents the yaw angle and angular velocity of the unmanned surface vessel's traction mechanism. represents the velocity of the surface unmanned ship in the x-axis direction, represents the velocity of the surface unmanned ship in the y-axis direction, represents the longitudinal velocity of the ocean current, represents the transverse velocity of the ocean current, represents the longitudinal velocity of the wind, represents the first derivative of the transverse position of the surface unmanned ship, represents the cosine value of represents the cosine value of represents the sine value of represents the sine value of represents the cosine value of represents the first derivative of represents the tangent value of represents the sine value of represents the angular velocity of the roll angle of the surface unmanned ship, represents the first derivative of the longitudinal position of the surface unmanned ship, represents the first derivative of represents the first derivative of represents the first derivative of represents the angular velocity of the heading angle of the surface unmanned ship's float, represents the angular velocity of the heading angle of the surface unmanned ship's tractor, w represents the vertical velocity of the surface unmanned ship, represents the component of the surface unmanned ship along the z-axis direction related to , , .

[0051] Specifically, is the component of the surface unmanned ship velocity along the z-axis direction is converted to the inertial or geographic coordinate system by a part of the projection of the Euler angle (especially , , .

[0052] wherein the position vector and the velocity vector describe the position and the velocity of the surface unmanned ship in space.

[0053] In one possible implementation, the origin of the system coordinate system is the center of the unmanned surface vessel, the x-axis points in the direction of the unmanned surface vessel's movement, the y-axis is perpendicular to the x-axis and points to the starboard side of the unmanned surface vessel, and the z-axis points in a vertical direction determined by the right-hand rule based on the x-axis and y-axis.

[0054] The origin of the Earth-fixed coordinate system is the Earth's center of mass, the x-axis points towards the geographic North Pole, the y-axis points towards the geographic East Pole, and the z-axis points vertically downwards.

[0055] Specifically, such as Figure 2 As shown, the system coordinate system (superscript: ) S ): Fix it at the system's center of gravity, i.e., on the armored cable of the unmanned surface vessel. Define the coordinate system. Perpendicular to the armored cable, and the arrow points in the direction of the system's forward velocity. Definition From the origin O of the system coordinate system S Pointing towards the glider along the armored cable, and in the system coordinate system The right-hand rule is satisfied.

[0056] Earth-fixed coordinate system (superscript in the upper right corner) N The location of the unmanned surface vessel can be well represented using a northeast-oriented coordinate system, where... Pointing due north, Pointing due east, and Vertically downwards.

[0057] S12: Establish a dynamic model by combining position and velocity vectors:

[0058]

[0059] in, Represents the rigid mass matrix. Indicates and The relevant inertial hydrodynamic matrix, Indicates and The relevant Coriolis centripetal force coefficient matrix, Indicates and The relevant hydrodynamic Coriolis centripetal force matrix, Indicates and The relevant damping force matrix, Indicates and The relevant restoring force matrix, Let R represent the active control force matrix, and let R represent the real number field. This represents a real matrix with six rows and six columns. This represents a real matrix with six rows and one column. This represents the disturbance vector.

[0060] Specifically, the process of establishing the dynamic model of the water unmanned ship determines its position and velocity vector through Newton-Euler equations, and describes its motion behavior by combining its inertia, external forces (such as sea currents, wind forces) and internal forces (such as propulsion forces). By introducing mechanical parameters such as rigid mass matrix, Coriolis force matrix, damping force matrix and restoring force matrix, the model can comprehensively consider the dynamic response of the water unmanned ship in complex environments. This model provides an accurate theoretical basis for the control strategy, enabling the control system to better adapt to external disturbances, improve path tracking accuracy, and enhance the robustness and stability of the system in different sea conditions.

[0061] S2: In combination with the dynamic model, a path tracking error equation is established in the Serret-Frenet coordinate system to describe the system error of the water unmanned ship relative to the desired path.

[0062] The Serret-Frenet coordinate system is a coordinate system used to describe points on a curve and tangent to the curve, which defines a local coordinate system where the x-axis is along the tangent direction of the curve; the y-axis is perpendicular to the tangent and points to the normal direction of the curve; the z-axis is perpendicular to the plane and is usually used to represent the normal of the curve. The desired path refers to the ideal navigation route that the water unmanned ship needs to follow. This path is pre-set and is usually planned according to task requirements or target positions. The path tracking error equation is used to measure the deviation between the actual path of the water unmanned ship and the desired path, usually including lateral error (indicating the vertical deviation of the ship from the path) and heading error (indicating the deviation between the actual heading of the ship and the desired heading).

[0063] It should be noted that by establishing the path tracking error equation in the Serret-Frenet coordinate system, the actual motion of the water unmanned ship and the deviation from the desired path can be effectively quantified. This method can accurately describe the change of error, help design more accurate control strategies, and thus achieve higher accuracy of path tracking, especially in complex environments, which can effectively reduce error accumulation and improve system stability.

[0064] In one possible implementation, the system error includes lateral error and heading error.

[0065] The lateral error is the deviation along the path, and the heading error is the deviation of the heading angle from the desired direction. By dividing the system error into lateral error and heading error, the deviation of the water unmanned ship during path tracking can be accurately quantified. The lateral error directly reflects the vertical deviation of the ship from the predetermined path, while the heading error reveals the deviation between the ship's heading and the target direction. This subdivision helps to design more accurate control strategies, which can adjust the lateral position and heading angle of the ship respectively, ensuring more accurate path tracking and enhancing the stability and robustness of the system.

[0066] In one possible implementation, the path tracking error equation is specifically:

[0067]

[0068]

[0069]

[0070] wherein, represents the sailing speed of the surface unmanned ship, represents the yaw angle of the surface unmanned ship, represents the lateral error, represents the first derivative of represents the first derivative of the heading error , represents the first derivative of the yaw angle error of the surface unmanned ship, represents the lateral speed of the surface unmanned ship in the Serret-Frenet coordinate system, represents the yaw angle velocity at the reference point p, represents the longitudinal error at the reference point p, represents the cosine value of the yaw angle error , represents the sine value of the yaw angle error , represents the yaw angle at the reference point p.

[0071] It should be noted that in this path tracking error equation, the lateral error measures the degree of deviation between the surface unmanned ship and the expected path. The heading error represents the difference between the actual heading of the surface unmanned ship and the expected heading. By clearly distinguishing between the lateral error and the heading error, the control system can more accurately adjust the motion of the surface unmanned ship, respectively optimizing the position and heading control of the ship. This method makes the path tracking process more detailed and flexible, which can improve the control accuracy in complex environments, reduce error accumulation, and thus enhance the stability and robustness of the system.

[0072] S3: Construct a path tracking controller based on a linear active disturbance rejection controller, estimate the external disturbance of the surface unmanned ship and compensate for the external disturbance, and output the control signal of the surface unmanned ship.

[0073] The linear active disturbance rejection controller (LADRC) is a method based on an extended state observer (ESO) for real-time estimation and compensation of external disturbances and internal disturbances of the system. It improves the stability and robustness of the system by observing and compensating for disturbances in the system. Compared with traditional controllers, LADRC does not rely on accurate mathematical models and has strong anti-interference ability. The path tracking controller is a controller used to guide the water unmanned ship to travel along the predetermined path. It adjusts the heading and speed of the unmanned ship based on the path tracking error equation to ensure that the ship can accurately follow the desired path. External disturbances refer to external factors that affect the motion of the water unmanned ship, such as ocean currents, wind, waves, etc. These disturbances can cause the heading and position of the ship to deviate from the expected trajectory, affecting the accuracy of path tracking. The water unmanned ship control signal is a signal generated by the controller, and the control signal is executed by actuators such as rudders and propulsion systems to adjust the heading and speed of the water unmanned ship, making it move towards the desired path.

[0074] It should be noted that by combining the linear active disturbance rejection controller, the influence of external disturbances such as ocean currents, wind and waves on the water unmanned ship can be estimated and compensated in real time, significantly improving control accuracy. It not only effectively deals with uncertainties in dynamic environments, but also ensures system stability, thereby achieving more accurate path tracking.

[0075] In one possible implementation, the control law of the path tracking controller is specifically:

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082] wherein, represents the water unmanned ship control compensation signal for external disturbances output by the path tracking controller at time k, represents the water unmanned ship control signal output by the path tracking controller at time k, represents the observer bandwidth assumption value, represents the undamped natural frequency, and represent the first feedback control gain and the second feedback control gain, respectively, , and respectively represent the first, second, and third observer gains, represents the estimation of the heading at time k, represents the estimation of the heading derivative at time k, represents the estimation of the total external disturbance at time k, represents the error signal of the path tracking controller at time k, represents the first order error of the path tracking controller at time k, represents the second order error of the path tracking controller at time k, represents the first order virtual control quantity of the path tracking controller at time k, represents the second order virtual control quantity of the path tracking controller at time k, represents the estimation of the external disturbance influence.

[0083] It is worth noting that this path tracking controller's control law can effectively deal with the influence of external disturbances by introducing multiple feedback gains and dynamic adjustment mechanisms. Through the optimization of control compensation signals and control signals, the control system can estimate the heading error, heading derivative, and external disturbance in real time, thus more accurately adjusting the controller parameters. The use of observer gains and feedback gains to adjust the controller effectively enhances the robustness of the control system. Especially in the face of environmental changes and complex disturbances, the system can maintain high-precision path tracking, improving control stability and adaptability.

[0084] S4: input the system error and water surface unmanned ship control signal into the fuzzy RBF neural network to minimize the system error, and output the optimized control parameters of the path tracking controller.

[0085] wherein the fuzzy RBF neural network is a hybrid network combining fuzzy logic and radial basis function (RBF) neural network. It first uses the fuzzification layer to convert input data (such as system error and control signal) into fuzzy sets, and then processes these fuzzy inputs through the RBF neural network to output control parameters. The RBF network maps through radial basis functions, which can effectively extract features from the input space and optimize them. In path tracking control, the fuzzy RBF neural network can dynamically optimize the controller parameters according to the goal of minimizing system error, adapting to environmental changes and disturbances.

[0086] ​It is worth noting that by inputting the system error and control signal into the fuzzy RBF neural network, the parameters of the path tracking controller can be adaptively optimized. This process can adjust the controller in real time to minimize errors and improve the accuracy and robustness of the system. The introduction of the fuzzy RBF neural network enables the controller to better respond in complex and dynamic environments and optimize control performance, especially when facing nonlinear disturbances.

[0087] In one possible implementation, the optimization control parameters include an observer bandwidth assumption value, an external disturbance influence estimation value, a first feedback control gain, and a second feedback control gain.

[0088] It is worth noting that by optimizing the control parameters, including the observer bandwidth, external disturbance estimation value, and feedback control gain, the control system can be dynamically adjusted to more accurately respond to external disturbances and optimize path tracking accuracy. This enhances the adaptability and robustness of the system, ensuring stable operation in complex environments.

[0089] In one possible implementation, the fuzzy RBF neural network includes an input layer, a fuzzification layer, a fuzzy inference layer, and an output layer connected in sequence, and the fuzzification layer is connected to the output layer.

[0090] S4 specifically includes:

[0091] S41: input the system error and the water surface unmanned ship control signal as the input data of the input layer, and the activation function of the input layer is:

[0092]

[0093] where x represents the input data, represents the output data corresponding to the input data x of the input layer.

[0094] S42: perform fuzzification processing on the input data through the fuzzification layer, and calculate the fuzzy quantities of the system error and the water surface unmanned ship control signal belonging to different fuzzy sets:

[0095]

[0096] where variable i=1,2 represents the input data label, i=1 represents that the input data is the system error, i=2 represents that the input data is the water surface unmanned ship control signal, variable j=1,2,3,4,5,6 represents the fuzzy set label, represents the natural exponential function, represents the output data of the input layer when the input data is i, represents the mean difference of the membership function of the jth fuzzy set of the ith input variable of the Gaussian function, a standard deviation of a membership function of the jth fuzzy set of the ith input variable representing a Gaussian function, a fuzzy quantity representing that the ith input data of the fuzzification layer output belongs to the jth fuzzy set.

[0097] S43: multiply the fuzzy quantities of each input data corresponding to different fuzzy sets and each other, to obtain the fuzzy inference layer output data:

[0098]

[0099] wherein, the fuzzy inference layer output data, represents N represents the total number of input variables, represents the total number of fuzzy sets, represents the membership degree product of each input variable in the ith rule, represents the product of the number of fuzzy subsets of all input variables.

[0100] S44: input the fuzzy inference layer output data to the output layer to obtain the initial optimization control parameter:

[0101]

[0102]

[0103]

[0104]

[0105] wherein, , , and respectively represent the observer bandwidth assumption value, the external disturbance influence estimation value, the first feedback control gain and the second feedback control gain, , , and comprise the initial optimization control parameter, represents a connection weight matrix, represents the weight connection from the 1st output layer to the jth node, represents the weight connection from the 2nd output layer to the jth node, represents the weight connection from the 3rd output layer to the jth node, represents the weight connection from the 4th output layer to the jth node.

[0106] Optionally, the connection weight matrix is a connection weight matrix of 36 rows and 4 columns.

[0107] S45: optimizing the connection weight matrix;

[0108]

[0109]

[0110]

[0111] wherein, denotes a learning momentum factor, denotes the number of iterations, denotes a learning rate, denotes an ideal output of the output layer in the kth iteration, denotes an actual output of the output layer in the kth iteration, denotes a partial derivative, denotes a performance index function, denotes a weight value of the jth connection in the kth iteration, denotes the jth weight in the k-1th iteration, denotes an increment of the weight in the kth iteration, denotes a weight value of the jth connection in the k-2th iteration, denotes a part of the partial derivative of the output layer with respect to the input, denotes an actual output, denotes an input variable, denotes a center of the jth node, denotes an output of the jth neuron of the 4th layer, denotes a connection weight of the jth node.

[0112] S46: in a case where the performance index function value is less than a preset performance function index value, outputting the initial optimization control parameter as the optimization control parameter, otherwise, updating the hyperparameters of the fuzzy RBF neural network, and returning to step S44, wherein the hyperparameters include the learning momentum factor and the learning rate.

[0113] Specifically, this process utilizes a fuzzy RBF neural network to optimize the parameters of the path tracking controller. First, the system error and control signal are input into the input layer, then they are converted into fuzzy sets through the fuzzification layer, and the membership degree of each input data is calculated. Next, the fuzzy inference layer outputs data based on the interaction of fuzzy sets and passes it to the output layer to calculate the preliminary optimized control parameters, such as observer bandwidth, external disturbance estimation value, feedback gain, etc. Then, by optimizing the connection weight matrix, the neural network continuously adjusts these control parameters to enable the system to adaptively improve path tracking accuracy and stability in different environments. The advantage is that it can cope with dynamic changes and complex disturbances of the system, enhance the flexibility and robustness of the control system, and through optimization iteration and adjustment of hyperparameters, the system can always maintain high efficiency and low error in actual operation.

[0114] It should be noted that the size of the preset energy function index value can be set by the skilled person according to actual needs, and the present application does not limit it.

[0115] S5: updating the path tracking controller according to the optimized control parameters.

[0116] It can be understood that the path tracking controller is updated according to the control parameters optimized by the fuzzy RBF neural network. This enables the controller to adjust its own parameters in real time to adapt to environmental changes or new dynamic conditions, ensuring that the water surface unmanned ship always tracks the path in the optimal state.

[0117] S6: re-estimating the external disturbance using the updated path tracking controller and outputting the optimized water surface unmanned ship control signal.

[0118] It should be noted that by re-estimating the external disturbance using the updated path tracking controller, the control system can adjust itself in new environmental conditions. Re-estimating the disturbance helps to improve control accuracy and ensures that the system can accurately respond to dynamic changes such as changes in ocean currents or wind, thereby outputting optimized control signals. This adjustment process enhances the robustness of the system, ensuring that the water surface unmanned ship always maintains stable path tracking capability in complex environments.

[0119] S7: controlling the water surface unmanned ship to travel along the desired path using the optimized water surface unmanned ship control signal.

[0120] In practical applications, first, the dynamic model of the unmanned surface ship is established to provide a theoretical basis for subsequent control. Then, the path tracking error equation is established in the Serret-Frenet coordinate system to accurately quantify the deviation between the ship and the desired path. Next, the linear active disturbance rejection controller (LADRC) is used to estimate and compensate external disturbances in real time, ensuring that the ship can respond to dynamic changes such as ocean currents, waves, and wind. The fuzzy RBF neural network is further used to optimize the control parameters, adjusting the controller in real time to reduce errors and improve control accuracy and stability. The method has the advantages of effectively dealing with nonlinear disturbances in complex environments, optimizing control performance, and ensuring that the unmanned surface ship can stably and accurately follow the desired path under different conditions.

[0121] As Figure 3 shown, it is a structural schematic diagram of the path tracking control system of the unmanned surface ship according to the embodiment of the present application.

[0122] The embodiment of the present application provides a path tracking control system 20 of an unmanned surface ship, which comprises a processor 201 and a memory 202.

[0123] The memory 202 stores programs or instructions that can run on the processor 201. When the programs or instructions are executed by the processor 201, the steps of the path tracking control method of the unmanned surface ship are realized, and the same technical effects can be achieved. To avoid repetition, the present application will not be described again.

[0124] It should be understood that the processor 201 in the embodiment of the present application can be a central processing unit (CPU). The processor can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), ready-to-program gate arrays (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0125] It should also be appreciated that the memory 202 in embodiments of the present application can be volatile memory or nonvolatile memory, or can include both volatile and nonvolatile memory. In one embodiment, nonvolatile memory can be read-only memory (ROM), programmable ROM (PROM), erasable programmable ROM (EPROM), electrically EPROM (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), which acts as external cache. By way of example and not limitation, many forms of random access memory can be used, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double-data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), and direct Rambus RAM (DR RAM).

[0126] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system, or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, and the like) embodying computer-readable program code.

[0127] The present application is described in reference to the flowchart illustrations and / or block diagrams of the methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing system or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions of the flowchart illustrations and / or block diagrams specified in the flowchart illustrations and / or block diagrams. These computer program instructions can also be stored in a computer-usable or computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-usable or computer-readable memory produce an article of manufacture including instructions which implement the function / ies of the flowchart illustrations and / or block diagrams. Figure 1 one or more flow(s) and / or block(s) Figure 1means for performing the function specified by the block or blocks.

[0128] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the flow Figure 1 one or more flowcharts and / or blocks Figure 1 means for performing the function specified by the block or blocks.

[0129] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the flow Figure 1 one or more flowcharts and / or blocks Figure 1 means for performing the function specified by the block or blocks.

[0130] The specific embodiments have been shown and described for purposes of illustrating the application, and not limitation, the above described embodiments will come within the scope and protection of the application. Any modifications, and changes, which come within the scope and spirit of the application are intended to be part of this application.

Claims

1. A method for controlling an unmanned watercraft under path following, characterized by, The method comprises the following steps: constructing a dynamic model of the water unmanned ship; establishing a path tracking error equation in a Serret-Frenet coordinate system in combination with the dynamic model, the path tracking error equation being used to quantify a system error between an actual sailing state of the water unmanned ship and an expected path; constructing a path tracking controller based on a linear active disturbance rejection controller, estimating external disturbances faced by the water unmanned ship through the path tracking controller and compensating the external disturbances, and outputting an initial water unmanned ship control signal; inputting the system error and the initial water unmanned ship control signal into a fuzzy RBF neural network to minimize the system error, and outputting optimized control parameters of the path tracking controller; updating the path tracking controller according to the optimized control parameters, re-estimating external disturbances by using the updated path tracking controller, and outputting an optimized water unmanned ship control signal; controlling the water unmanned ship to travel along the expected path based on the optimized water unmanned ship control signal.

2. The water surface unmanned ship control method under path tracking according to claim 1, characterized in that, The specific steps of constructing the dynamic model of the water unmanned ship comprise: determining a position vector and a velocity vector of the water unmanned ship in a preset coordinate system according to Newton-Euler equations, and constructing the dynamic model in combination with the position vector and the velocity vector.

3. The water surface unmanned ship control method under path tracking according to claim 2, characterized in that, The preset coordinate system comprises a system coordinate system and a geostationary coordinate system. The system coordinate system has an origin at a center of the water unmanned ship, an x-axis pointing to a forward direction of the water unmanned ship, a y-axis perpendicular to the x-axis and pointing to a starboard of the water unmanned ship, and a z-axis determined based on the x-axis and the y-axis through a right-hand rule. The geostationary coordinate system has an origin at a center of the earth, an x-axis pointing to a geographic north pole, a y-axis pointing to a geographic east pole, and a z-axis vertically downward. 4.The method of claim 1, wherein, The system error comprises a lateral error and a heading error, the lateral error being a vertical deviation of the water unmanned ship from the expected path, and the heading error being an angular deviation of an actual heading of the water unmanned ship from a corresponding heading of the expected path. 5.The method of claim 1, wherein, The path tracking controller introduces a plurality of feedback control gains and a plurality of observer gains, estimates a current heading, a heading change rate and an external total disturbance of the water unmanned ship in real time through the observer gains, and adjusts control parameters in combination with the feedback control gains to realize compensation of the external disturbance. 6.The method of claim 1, wherein, The fuzzy RBF neural network comprises an input layer, a fuzzification layer, a fuzzy reasoning layer and an output layer connected in sequence, and the fuzzification layer and the output layer are directly connected.

7. The water surface unmanned ship control method according to claim 6, wherein The input layer receives the system error and the initial water unmanned ship control signal to transfer input data through a linear activation function; The fuzzification layer performs fuzzification processing on the input data to calculate fuzzy quantities of the system error and the initial water unmanned ship control signal belonging to different fuzzy sets respectively; The fuzzy reasoning layer performs a pairwise product operation on the fuzzy quantities of different fuzzy sets corresponding to each input data to output a reasoning result; The output layer receives the reasoning result output by the fuzzy reasoning layer to calculate initial optimized control parameters based on a preset connection weight matrix.

8. The water surface unmanned ship control method according to claim 7, wherein The fuzzy RBF neural network optimizes the connection weight matrix by introducing a learning momentum factor and a learning rate, calculates a performance index function according to the deviation between the ideal output and the actual output of the output layer at the current iteration number, and outputs the initial optimization control parameter as the final optimization control parameter when the value of the performance index function is less than a preset threshold; if not, the learning momentum factor and the learning rate are updated, the initial optimization control parameter is recalculated, and the optimization process is repeated.

9. The water surface unmanned ship control method according to claim 8, wherein, The optimization control parameter includes an observer bandwidth assumption value, an external disturbance influence estimation value, and a feedback control gain.

10. A water surface unmanned ship control system under path tracking, characterized in that, The method comprises: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, so that the one or more processors implement the path tracking under water unmanned ship control method as claimed in any one of claims 1~9.

Images