Path tracking control method based on digital twin driven sea wave inversion observation

Abstract

The application discloses a fine path tracking control method for unmanned ships based on digital and intelligent double-driven sea wave inversion observation, which comprises the following steps: S1, identifying the ship motion response through the method of system identification; S2, obtaining an inversion wave spectrum algorithm of unknown waves to obtain information such as significant wave height, wave period and frequency; S3, modeling sea wave interference by using the obtained wave parameters, and applying real-time sea wave interference to the unmanned ship in the form of a mathematical model; S4, designing a bow angle speed and longitudinal speed controller to compensate for the interference in the current sea state in real time, thereby achieving better control effect; and S5, designing a disturbance observer based on a reduced-order extended state observer to estimate the error generated in the sea wave acquisition and the model uncertainty term, thereby offsetting the influence of external disturbances such as wind, waves and currents on the path tracking task of the unmanned water surface ship, and improving the robustness of the path tracking controller.

Description

Technical Field

[0001] This invention relates to the field of ocean wave simulation technology, and in particular to a refined path tracking control method for unmanned vessels based on digital and intelligent dual-drive ocean wave inversion observation. Background Technology

[0002] With the continuous development and breakthroughs of artificial intelligence and other high technologies, maritime navigation equipment is gradually moving towards intelligent decision-making, modular systems, and high autonomy. Among these, unmanned surface vessels (USVs), as intelligent maritime navigation equipment, play a crucial role in marine resource exploration and development, search and rescue of distressed vessels, and the maintenance of national maritime sovereignty due to their advantages such as small size, light weight, and ability to carry intelligent equipment. An unmanned surface vehicle (USV) is an intelligent, multifunctional surface platform that can be controlled remotely or autonomously. USVs can carry various functional modules, such as multiple intelligent sensors, detectors, search and rescue equipment, and weapons, playing diverse roles in different fields. Path tracking is one of the fundamental motion control problems in the autonomous control of USVs. Path tracking is defined as an USV following a predefined, time-independent path at a certain speed. Ocean waves refer to a complex, multidimensional, random fluctuation phenomenon caused by wind forcing and numerous factors such as atmospheric and oceanic topography.

[0003] The 21st century is the century of the ocean. To ensure the safety and efficiency of ships operating at sea, acquiring wave information has become particularly important. In ship motion control systems, it is necessary to utilize the thrust of the ship's propellers to overcome the interference forces of wind, waves, and currents in the environment. If wave force feedforward is incorporated, the system can respond sufficiently to the current wave forces, improving system control performance and reducing energy consumption. Currently, traditional ship motion control primarily relies on disturbance observers to estimate and compensate for current disturbances. However, disturbance observers cannot accurately estimate current disturbances when sea conditions are complex, thus hindering effective ship control. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention discloses a refined path tracking control method for unmanned surface vessels based on digital and intelligent dual-drive wave inversion observation, specifically including the following steps:

[0005] S1: Identify ship motion response using system identification methods;

[0006] S2: An algorithm for rapidly obtaining the inversion wave spectrum of unknown waves through adaptive optimization, acquiring information such as meaningful wave height, wave period, and frequency;

[0007] S3: Model the acquired wave parameters for wave interference, and apply the real-time wave interference to the unmanned surface vessel in the form of a mathematical model.

[0008] S4: Design a bow angular velocity and longitudinal velocity controller using the Backstepping and Lyapunov design methods to compensate for disturbances under the current sea state in real time, thereby achieving better control performance;

[0009] S5: Design a disturbance observer based on a reduced-order extended state observer to estimate the errors generated in wave acquisition and model uncertainties, thereby offsetting the impact of external disturbances such as wind, waves and currents on the path tracking task of unmanned surface vessels and improving the robustness of the path tracking controller.

[0010] In S1, an autoregressive model (AR model) is used to describe the wave-excited ship motion system. Its discrete model expression after Z-transformation is as follows:

[0011]

[0012] By using simulated ocean waves as excitation, the dynamic response time series of ships are obtained. The mechanism of the influence of different types and compositions of wave disturbances on the performance of the control system is summarized, and the dynamic response characteristics of ships under different operating conditions are studied when excited by wave disturbances.

[0013] To identify the ship amplitude response operator (RAO), and to reduce noise interference during the identification process, it is assumed that the observed noise signal and the wave signal are independent. Then, the autoregressive model parameters are automatically adjusted based on the dynamic changes of real-time measured data. When new data arrives, an equal amount of historical data is replaced, ensuring that the modeling data length remains constant and up-to-date. The parameters are optimized using a fast adaptive mutation particle swarm optimization algorithm to improve the identification rate. The AR model parameters are identified using the recursive least squares method. Let the model input data at time i be {x(i-1), x(i-2), x(i-3), ..., x(in)}, and let this discrete time-history sequence be a stationary time series with zero mean. In this paper, it is assumed that the model input data at time i < 1 is zero, and ε(i) is the desired output x(i) and the estimated output x(i). The difference can be derived using the AR model and the above assumptions.

[0014]

[0015] in, X(i)=[x(i-1),x(i-2),...,x(in)] T

[0016] The objective function is

[0017]

[0018] Here, β is the forgetting operator, and 0 < β(t,i) ≤ 1.

[0019] In S4, adaptive blind deconvolution is used to perform convolutional decomposition on the output observed signal, thereby obtaining the correlation characteristics between the excitation signal and the system transfer function. To ensure that deconvolution can be performed and that the decomposition is unique, a normalization approach is adopted, and constraints are added to the input based on prior knowledge of the ship's motion response and the offline database. This achieves a dual-driven approach of knowledge and data, which can improve the efficiency and speed of wave parameter acquisition.

[0020] The adaptive wave spectrum inversion system is driven by the mean square error of the input and inverted wave spectrum. The process of constructing the deconvolution kernel function using a Gaussian type function is as follows:

[0021] H(k,ω)=Cr1·h(ω)

[0022] h(i)=C a exp(-γ,i 2 )

[0023] Where H(k,ω) is the kernel function, Cr1 is the convolution kernel amplitude parameter, and C a The normalization constant is set to make max(h(i)) = 1, k is the iteration number, and γ is the convolution kernel fitting parameter.

[0024] In the adaptive iterative optimization process, the fitting function of the ship's motion frequency response and the wave spectrum inversion are iterated alternately, and the iteration termination criterion uses frequency domain error measurement.

[0025]

[0026] in, D(ω)=Y(ω).

[0027] The objective function for optimization can be chosen as J. i =0.5E i (k), i=1,2,3, thus forming an algorithm for rapidly obtaining the inversion wave spectrum of unknown waves, which can provide information such as meaningful wave height, wave period and frequency.

[0028] In S3, the acquired wave parameter information is used to perform interference modeling, and the obtained interference information is added to the ship in the form of a mathematical model.

[0029] The disturbance force of ocean waves on ships is not only related to the nature of the waves, but also to the ship's own motion. A ship in motion has its speed, course, and positional relationship and angle with the waves constantly changing. The ocean wave disturbance force model is defined and designed as follows:

[0030]

[0031] ρ is the density of seawater, L is the length of the ship, and B is the density of seawater. m For the width of the ship, T m It refers to the ship's draft, ψ α The encounter angle is the angle between the direction of the wave and the ship's heading. Since the energy of the waves is mainly concentrated in the lower-order waves, it is defined as follows:

[0032]

[0033] Where ω ei For the frequency of encounters, A i λ represents the amplitude of the ocean waves. i φ is the wavelength. i The phase is randomly distributed on [0, 2π).

[0034] In S4, the bow angular velocity and longitudinal velocity controllers are designed using backstepping and Lyapunov design methods:

[0035] This method uses a three-degree-of-freedom ship, whose kinematic and dynamic models are as follows:

[0036]

[0037]

[0038] The dynamic equation for the position error in unmanned surface vessel path tracking is as follows:

[0039]

[0040] The AILOS guidance law is designed as follows:

[0041]

[0042]

[0043] To accomplish the path tracking control task of the unmanned surface vessel, after obtaining the ideal heading angle through the guidance system, it is also necessary to design an effective dynamic controller to track the ideal heading angle and maintain the ideal longitudinal velocity of the unmanned surface vessel. Therefore, the control laws for the bow angular velocity subsystem and the longitudinal velocity control law are designed as follows:

[0044]

[0045]

[0046] We designed a disturbance observer based on a reduced-order extended state observer to estimate the errors and model uncertainties generated in wave acquisition, thereby offsetting the impact of external disturbances such as wind, waves and currents on the path tracking task of unmanned surface vessels and improving the robustness of the path tracking controller.

[0047] The basic idea is to observe the difference between the nominal model output and the actual system output caused by external environmental disturbances through a disturbance observer, and then use feedforward compensation in the controller design to offset the impact of disturbances on the system.

[0048] The error and disturbance observer design is as follows:

[0049]

[0050]

[0051] In this invention, to achieve better control during unmanned surface vessel (USV) navigation, a real-time wave-compensated path-tracking control method is studied. First, wave information is acquired through system identification based on the vessel's motion response, and then real-time compensation is applied during USV path-tracking control to achieve better control performance. The main contributions of this paper are as follows:

[0052] [1] This invention innovatively uses real-time acquired wave parameter information as wave interference in the form of a mathematical model on the unmanned vessel, and combines it with the path tracking control of the unmanned vessel to compensate for the current interference in real time, thereby achieving better control effect.

[0053] [2] Based on the ship motion response, the system identification and adaptive optimization method is innovatively applied to the acquisition of wave parameters for the design of unmanned surface vessel controller, which meets the needs of practical engineering.

[0054] [3] Compared with traditional unmanned vessel control, this invention considers the real wave interference encountered by the unmanned vessel during navigation and studies an unmanned vessel path tracking control based on wave parameter acquisition. The robustness of the path tracking controller is improved by designing the controller using Backstepping and Lyapunov.

[0055] Based on the above advantages, the method proposed in this application is more practical. This method has advantages such as strong real-time performance and good realism, and has broad application prospects in the fields of wave simulation and ship motion control. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 This is a three-degree-of-freedom unmanned surface vessel model according to the method of the present invention;

[0058] Figure 2 This is a schematic diagram of LOS guidance according to the present invention.

[0059] Figure 3 This is a block diagram of the path tracking controller of the present invention. Detailed Implementation

[0060] To make the technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention:

[0061] This invention proposes a refined path tracking control method for unmanned surface vessels (USVs) based on digitally intelligent dual-drive wave inversion observation. This method allows for real-time acquisition of sea state information under current operating conditions, enabling the vessel system to compensate for current wave disturbances in a timely manner, thereby achieving better control performance and improving system robustness. For most control systems, acquiring current wave information and performing real-time compensation not only achieves better control performance but also reduces energy consumption. For example, for ship roll reduction devices, incorporating real-time wave information into the roll reduction control system can compensate for related motions and avoid larger rolling motions; during carrier-based aircraft takeoff and landing, incorporating real-time wave information can more accurately estimate the ship's motion state, thus providing better timing for aircraft landing; during ship navigation, wave disturbances are generally treated as noise and incorporated into Kalman filters, but incorrect noise signals can reduce the filtering effect and cause navigation deviations. Incorporating real-time wave information can improve navigation accuracy. Accurate measurement of wave information is of great significance for ship navigation safety, offshore platform construction, and marine resource exploitation. Based on the above background, it is clear that this invention has significant practical implications.

[0062] The operating environment of unmanned surface vessels (USVs) is complex and varied, making consideration of current sea conditions crucial. External wave interference directly affects the design and effectiveness of observers and controllers. A good controller can enable USVs to navigate quickly and safely even in complex and changing sea conditions. Therefore, if wave information under current sea conditions can be obtained in real time before designing a ship motion controller, and the controller design can compensate for current wave interference, the controller will be more effective and robust.

[0063] like Figures 1 to 3 The refined path tracking control method for unmanned surface vessels based on digital and intelligent dual-drive wave inversion observation, as shown, includes the following steps:

[0064] S1: Identify ship motion response using system identification methods;

[0065] S2: An algorithm for rapidly obtaining the inversion wave spectrum of unknown waves through adaptive optimization, acquiring information such as meaningful wave height, wave period, and frequency;

[0066] S3: Model the acquired wave parameters for wave interference, and apply the real-time wave interference to the unmanned surface vessel in the form of a mathematical model.

[0067] S4: Design bow angular velocity and longitudinal velocity controllers using Backstepping and Lyapunov design methods to compensate for disturbances under the current sea state in real time, thereby achieving better control performance;

[0068] S5: Design a disturbance observer based on a reduced-order extended state observer to estimate the errors generated in wave acquisition and model uncertainties, thereby offsetting the impact of external disturbances such as wind, waves and currents on the path tracking task of unmanned surface vessels and improving the robustness of the path tracking controller.

[0069] The discrete model expression of the wave-excited ship motion system, after Z-transformation, is as follows: (Autoregressive model (AR model)

[0070]

[0071] By using simulated ocean waves as excitation, the dynamic response time series of ships are obtained. The mechanism of the influence of different types and compositions of wave disturbances on the performance of the control system is summarized, and the dynamic response characteristics of ships under different operating conditions are studied when excited by wave disturbances.

[0072] To identify the ship amplitude response operator (RAO), and to reduce noise interference during the identification process, it is assumed that the observed noise signal and the wave signal are independent. Then, the autoregressive model parameters are automatically adjusted based on the dynamic changes of real-time measured data. When new data arrives, an equal amount of historical data is replaced, ensuring that the modeling data length remains constant and up-to-date. The parameters are optimized using a fast adaptive mutation particle swarm optimization algorithm to improve the identification rate. The AR model parameters are identified using the recursive least squares method. Let the model input data at time i be {x(i-1), x(i-2), x(i-3), ..., x(in)}, and let this discrete time-history sequence be a stationary time series with zero mean. In this paper, it is assumed that the model input data at time i < 1 is zero, and ε(i) is the desired output x(i) and the estimated output x(i). The difference can be derived using the AR model and the above assumptions.

[0073]

[0074] in, X(i)=[x(i-1),x(i-2),...,x(in)] T

[0075] The objective function is

[0076]

[0077] Here, β is the forgetting operator, and 0 < β(t,i) ≤ 1.

[0078] An adaptive blind deconvolution is employed to perform convolutional decomposition on the output observed signal, thereby obtaining the correlation characteristics between the excitation signal and the system transfer function. To ensure that deconvolution is possible and that the decomposition is unique, a normalization approach is used to impose additional constraints on the ship's motion response and input based on prior knowledge (offline database). This dual-driven approach of knowledge and data improves the efficiency and speed of wave parameter acquisition.

[0079] The adaptive wave spectrum inversion system is driven by the mean square error of the input and inverted wave spectrum. The process of constructing the deconvolution kernel function using a Gaussian type function is as follows:

[0080] H(k,ω)=Cr1·h(ω)

[0081] h(i)=C a exp(-γ,i 2 )

[0082] Where H(k,ω) is the kernel function, Cr1 is the convolution kernel amplitude parameter, and C a The normalization constant is set to make max(h(i)) = 1, k is the iteration number, and γ is the convolution kernel fitting parameter.

[0083] In the adaptive iterative optimization process, the fitting function of the ship's motion frequency response and the wave spectrum inversion are iterated alternately, and the iteration termination criterion uses frequency domain error measurement.

[0084]

[0085] in, D(ω)=Y(ω).

[0086] The objective function for optimization can be chosen as J. i =0.5E i (k), i=1,2,3, thus forming an algorithm for quickly obtaining the inversion wave spectrum of unknown waves, which can provide information such as meaningful wave height, wave period and frequency.

[0087] The acquired wave parameter information is used to model interference, and the obtained interference information is added to the ship in the form of a mathematical model.

[0088] The disturbance force of ocean waves on ships is not only related to the nature of the waves, but also to the ship's own motion. A ship in motion has its speed, course, and positional relationship and angle with the waves constantly changing. The ocean wave disturbance force model is defined and designed as follows:

[0089]

[0090] ρ is the density of seawater, L is the length of the ship, and B is the density of seawater. m For the width of the ship, T m It refers to the ship's draft, ψ α The encounter angle is the angle between the direction of the wave and the ship's heading. Since the energy of the waves is mainly concentrated in the lower-order waves, it is defined as follows:

[0091]

[0092] Where ω ei For the frequency of encounters, A i λ represents the amplitude of the ocean waves. i φ is the wavelength. i The phase is randomly distributed on [0, 2π).

[0093] The unmanned surface vessel model will be based on the Cybership II unmanned surface vessel model, and its model parameters are shown in the table below:

[0094] Table 1. CybershipII Related Model Parameters

[0095]

[0096] Table 1 Parameters of Unmanned Surface Vessel Path Tracking Controller

[0097]

[0098] Design of bow angular velocity and longitudinal velocity controllers using backstepping and Lyapunov design methods:

[0099] This method uses a three-degree-of-freedom ship, whose kinematic and dynamic models are as follows:

[0100]

[0101]

[0102] The dynamic equation for the position error in unmanned surface vessel path tracking is as follows:

[0103]

[0104] The AILOS guidance law is designed as follows:

[0105]

[0106]

[0107] To accomplish the path tracking control task of the unmanned surface vessel, after obtaining the ideal heading angle through the guidance system, it is also necessary to design an effective dynamic controller to track the ideal heading angle and maintain the ideal longitudinal velocity of the unmanned surface vessel. Therefore, the control laws for the bow angular velocity subsystem and the longitudinal velocity control law are designed as follows:

[0108]

[0109]

[0110] We designed a disturbance observer based on a reduced-order extended state observer to estimate the errors and model uncertainties generated in wave acquisition, thereby offsetting the impact of external disturbances such as wind, waves and currents on the path tracking task of unmanned surface vessels and improving the robustness of the path tracking controller.

[0111] The error and disturbance observer design is as follows:

[0112]

[0113]

[0114] in, External disturbance D j The estimated value of (j=u,v) has an estimation error of ξ j (j=1,2) are the state variables of the observer, ωj (j=1,2) are the design parameters of the disturbance observer. M represents the positive definite symmetric inertial matrix including the added mass; C represents the Coriolis force and centripetal force matrices; D is the linear damping matrix, τ u , τ r These are the control laws for the longitudinal and lateral directions.

[0115] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A path tracking control method based on digital and intelligent dual-drive ocean wave inversion observation, characterized in that... include: The amplitude response operator of ship motion response is identified using a system identification method; An inversion wave spectrum algorithm for unknown waves is obtained through an adaptive optimization method, and the meaningful wave height, wave period and frequency information are obtained by an adaptive blind deconvolution method. The acquired wave parameters are used to model ocean wave interference, and the real-time ocean wave interference is applied to the unmanned surface vessel in the form of a mathematical model. A bow angular velocity and longitudinal velocity controller was designed using the Backstepping and Lyapunov design methods to compensate for disturbances under the current sea state in real time. We designed a disturbance observer based on a reduced-order extended state observer to estimate the errors and model uncertainties generated in wave acquisition, thereby offsetting the impact of external disturbances such as wind, waves and currents on the path tracking mission of unmanned surface vessels. The wave disturbance force model is defined and designed as follows: The density of seawater, He is the captain. For the width of the boat, It refers to the ship's draft. The angle between the direction of the waves and the ship's course. Let gravitational acceleration be the acceleration due to gravity, i.e., the encounter angle, and since the energy of ocean waves is mainly contained in low-order waves, the definition is as follows: in For the frequency of encounters, The amplitude of the ocean waves. For wavelength, for The phase is randomly distributed on the upper surface; When designing bow angular velocity and longitudinal velocity controllers: The kinematic and dynamic models of a ship employing three degrees of freedom are as follows: In the above formula: the position and attitude information of the unmanned surface vessel in the inertial coordinate system is provided by... The velocity information of the unmanned surface vessel in the carrier coordinate system is represented by... express; Let be the rotation matrix from the carrier coordinate system to the inertial coordinate system. This represents the positive definite symmetric inertia matrix that includes additional mass; Represents the Coriolis force and centripetal force matrices; It is a linear damping matrix; The input force and torque to the controller, To obtain wave disturbances through ship motion response, The error caused by the wave disturbance is to be obtained, as well as the uncertainty term not modeled in the model.

2. The method according to claim 1, characterized in that: An adaptive regression model is used to describe the ship motion system excited by ocean waves. The ship's dynamic response time series is obtained by simulating ocean waves as excitation. The mechanism of the influence of different types and compositions of wave disturbances on the performance of the control system is summarized, and the dynamic response characteristics of the ship under different operating conditions are studied when excited by wave disturbances. When identifying the ship amplitude response operator: it is assumed that the observed noise signal and the wave signal are independent of each other. The parameters of the adaptive regression model are automatically adjusted according to the dynamic changes of the real-time measured data. When new data enters, it will replace an equal amount of historical data. The parameters are optimized by a fast adaptive mutation particle swarm algorithm to improve the identification rate. Parameter identification of the adaptive regression model is performed using the recursive least squares method. The model input data at time t is The ship's dynamic response time series is a stationary time series with zero mean. Assuming... The model input data at time t is zero. It is hoped that the output will be... With estimated output The difference is obtained through an adaptive regression model and the above assumptions: in, , The objective function is: Where is the forgetting operator, and .

3. The method according to claim 1, characterized in that: An adaptive blind deconvolution method is used to perform convolutional decomposition on the observed signal of the ship's motion response output, thereby obtaining the correlation characteristics between the excitation signal and the system transfer function. Based on the prior knowledge of the ship's motion response and input, additional constraints are applied to the deconvolution process to realize a dual-driven process of knowledge and data. The adaptive wave spectrum inversion system is driven by the mean square error of the input and inverted wave spectrum. The process of constructing the deconvolution kernel function using a Gaussian type function is as follows: in, For kernel function, The convolution kernel amplitude parameter, For the normalization constant, , The iteration number, These are the parameters for fitting the convolution kernel; In the adaptive iterative optimization process, the fitting function of the ship's motion frequency response and the wave spectrum inversion are iterated alternately at the same time, and the iteration termination criterion uses frequency domain error measurement. in, , ; The optimization objective function is selected as This leads to the development of an inversion wave spectrum algorithm for quickly obtaining unknown waves, thereby acquiring meaningful wave height, wave period, and frequency information.

4. The method according to claim 1, characterized in that: The AILOS guidance law is designed as follows: In the above formula, Forward sight distance, forward sight distance. Forward sight distance, For design parameters, For drift angle The estimated value, the tangential angle of the path curve Lateral tracking error , Represents longitudinal velocity With drift speed The resultant velocity; To track the ideal heading angle and maintain the ideal longitudinal velocity of the unmanned surface vessel, the control laws for the bow angular velocity subsystem and the longitudinal velocity control law are designed as follows: In the above formula, , , For design parameters, , For model parameters, , Disturbed by the waves, , These are the inversion error and model uncertainty terms. For heading angle error, The ideal longitudinal velocity.

5. The method according to claim 1, characterized in that: The difference between the nominal model output and the actual system output caused by external environmental disturbances is observed by a disturbance observer, and the impact of disturbances on the system is offset by feedforward compensation in the controller design. The error and disturbance observer design is as follows: in, External disturbances The estimated value, with an estimation error of , For the observer's state variables, For the design parameters of the error and disturbance observer, This represents the positive definite symmetric inertia matrix that includes additional mass; Represents the Coriolis force and centripetal force matrices; It is a linear damping matrix. , These are the control laws for the longitudinal and lateral directions.

Images