Ship formation control method and device, electronic equipment and computer storage medium

Abstract

This invention relates to a ship formation control method, apparatus, electronic device, and computer storage medium, belonging to the field of ship transportation technology. The ship formation control method includes: constructing a pilot ship dynamic model based on satellite positioning data and propeller / rudder data of a manned pilot ship; constructing a follower ship dynamic model based on the hydrodynamic parameters of an unmanned follower ship; constructing a mixed ship formation model based on the pilot ship dynamic model and the unmanned follower ship dynamic model; determining the multi-scale navigation intention of the manned pilot ship based on its historical navigation data and real-time maneuvering data; and performing formation navigation control of the unmanned follower ship based on the multi-scale navigation intention of the manned pilot ship and the mixed ship formation model, combined with a distributed model predictive control framework. This invention enables the control of a mixed ship formation of manned pilot ships and unmanned follower ships.

Description

Technical Field

[0001] This invention relates to the field of shipping technology, and in particular to a method, apparatus, electronic device, and computer storage medium for controlling ship formations. Background Technology

[0002] Ship formations can reduce navigation resistance, improve energy efficiency and navigation safety through coordinated navigation, which is of great significance for improving the level of shipping intelligence, increasing profitability and developing a green economy.

[0003] Existing research on the control of manned vessel-guided mixed vessel formations is still lacking. Related studies focus on the distributed collaborative control of fully autonomous vessel formations, where control behaviors and prediction information are directly shared among vessels. However, manned vessel-led formations require modeling and prediction of manned vessel behavior to achieve efficient collaboration, which is not applicable to existing research.

[0004] This demonstrates that existing technologies cannot achieve manned vessel-guided mixed vessel convoy control. Summary of the Invention

[0005] In view of this, it is necessary to provide a method, device, electronic equipment and computer storage medium for ship formation control, in order to solve the problem that the existing technology cannot achieve mixed ship formation control with manned ship guidance.

[0006] To address the aforementioned problems, in a first aspect, the present invention provides a method for controlling ship formations, comprising:

[0007] A dynamic model of the pilot ship is constructed based on satellite positioning data and propeller data of the manned pilot ship. A dynamic model of the follower ship is constructed based on hydrodynamic parameters of the unmanned follower ship. A mixed formation model of ships is constructed based on the dynamic models of the pilot ship and the unmanned follower ship. The multi-scale navigation intentions of the manned pilot vessel are determined based on its historical navigation data and real-time maneuvering data. This paper describes a distributed model predictive control framework for controlling the formation navigation of unmanned follower ships based on the multi-scale navigation intentions of manned pilot ships and the ship hybrid formation model.

[0008] In one possible implementation, the pilot ship's dynamics model is as follows:

[0009]

[0010]

[0011] in, Let K+1 be the state of the ship with a pilot. For the first fully connected deep neural network in a feedforward network l The weight matrix of the layer, For activation function, For the first fully connected deep neural network in a feedforward network l Layer bias vector, The position of the piloted ship at time k. Let K be the heading angle of the piloted vessel at time k. Let be the longitudinal velocity of the manned ship at time k. Let be the lateral velocity of the piloted ship at time k. Let be the bow roll rate of the manned vessel at time k. Let $k$ be the propeller speed of the manned ship at time $k$. The rudder angle of a manned ship at time k.

[0012] In one possible implementation, the follower ship dynamics model is as follows:

[0013] in, To determine the state of the following ship at time k+1, This is the rotation matrix used to transform the velocity vector in the hull coordinate system to the inertial coordinate system. Let M be the velocity vector of the following ship at time k+1, and M be the inertia matrix of the following ship. For the Coriolis and centripetal force matrices of the following ship, For the damping matrix of the following ship, The parameters of the following ship at time k are the hydrodynamic parameters.

[0014] In one possible implementation, the multi-scale navigation intentions of the manned pilot are determined based on the pilot's historical navigation data and real-time maneuvering data, including: Acquire historical satellite positioning data of the manned pilot in the current waters and historical maneuvering data corresponding to the historical satellite positioning data to obtain historical trajectory data of the manned pilot vessel; Acquire real-time maneuvering data and real-time trajectory status of manned pilot vessels, and combine them with historical trajectory data to determine the micro-operational behavior of manned pilot vessels. Extract historical trajectory features from historical trajectory data, and combine real-time manipulation data and historical trajectory features to cluster the future trajectories of manned pilot vessels to determine the macroscopic trajectory category of manned pilot vessels; By fusing micro-level operational behaviors and macro-level trajectory categories, we can obtain the multi-scale navigation intentions of manned piloted ships.

[0015] In one possible implementation, formation navigation control of the unmanned follower vessel is performed based on a distributed model predictive control framework combined with a multi-scale navigation intention model of the manned pilot vessel and a mixed formation model of ships, including: In the pre-defined distributed model predictive control framework, the objectives are to minimize the error between the ship's operating state and the reference trajectory state, and to keep the distance and angle difference between each adjacent unmanned follower ship within a preset range. The objective cost function of each unmanned follower ship is constructed by combining the multi-scale navigation intention of the manned pilot ship and the ship hybrid formation model. The contraction constraints of the ship mixed formation model are determined based on the objective cost function combined with the inverse step control law; A formation controller is designed for unmanned follower ships based on contraction constraints, and the formation navigation control of unmanned follower ships is performed based on the formation controller.

[0016] In one possible implementation, the objective cost function is:

[0017] in, For the first i The target cost function of the ships To predict the step size, N is the number of ships in the mixed fleet. Indicates the first i The ship at the moment k The state vector, Indicates the first i The ship at the predicted time k + s The state error relative to the reference trajectory, and They represent the first i The ship at the predicted time k + s The actual relative position vector and the estimated position vector, and They represent the first j The ship at the predicted time k + s The actual relative position vector and the estimated position vector, This represents the cost function used to avoid collisions between ships, with the input being the first... i Ship and the j Spatial position estimation of the ship at the predicted time. For the control input of the i-th ship at the predicted time k+s, It is a positive definite matrix. and This is a preset weighted matrix.

[0018] In one possible implementation, the controller for the ship mixed formation model is:

[0019] Among them, among them, Let be the discrete-time system dynamics equation for the i-th ship. Is this moment for k + s Predicted state at any given time Is this moment for k + s Predictive control input at time, Let k be the predicted state of the system at time k. The actual state of the system. and Indicates the upper and lower bounds of the input. For the ship i With the ship j The distance between them and For the ship i With the ship j Distance constraints between them Indicates state tracking error. This represents the first and current input in the action sequence. This is the rotation matrix used to transform the velocity vector in the hull coordinate system to the inertial coordinate system. Let be the velocity vector of ship i at time k. Let M be the Coriolis and centripetal force matrices of ship i, and M be the inertia matrix of ship i. Let i be the damping matrix of ship i. It is a positive definite matrix.

[0020] Secondly, the present invention also provides a ship formation control device, comprising: The model building module is used to build a pilot ship dynamic model based on satellite positioning data and propeller data of the manned pilot ship, build a follower ship dynamic model based on hydrodynamic parameters of the unmanned follower ship, and build a mixed formation model of ships based on the pilot ship dynamic model and the unmanned follower ship dynamic model. The manned pilot vessel intent estimation module is used to determine the multi-scale navigation intent of the manned pilot vessel based on its historical navigation data and real-time maneuvering data. The formation control module is used to control the formation navigation of unmanned follower ships based on the multi-scale navigation intentions of the manned pilot ship and the ship hybrid formation model combined with a distributed model predictive control framework.

[0021] Thirdly, the present invention also provides an electronic device, including a memory and a processor, wherein, Memory, used to store programs; A processor, coupled to memory, is used to execute programs stored in memory to implement the steps in the ship formation control method of any of the above implementations.

[0022] Fourthly, the present invention also provides a computer-readable storage medium for storing a computer-readable program or instructions, which, when executed by a processor, can implement the steps in the ship formation control method of any of the above implementations.

[0023] The beneficial effects of this invention are as follows: The ship formation control method provided by this invention constructs a dynamic model of the pilot ship based on satellite positioning data and propeller and rudder data of the manned pilot ship, constructs a dynamic model of the follower ship based on the hydrodynamic parameters of the unmanned follower ship, and constructs a mixed ship formation model based on the dynamic models of the pilot ship and the unmanned follower ship. By modeling the manned pilot ship and the unmanned follower ship separately, it can adapt to the navigation characteristics of the manned pilot ship and the unmanned follower ship, and construct a mixed ship formation model, which facilitates the subsequent control of the entire mixed ship formation. Based on historical navigation data and real-time maneuvering data of manned pilot vessels, the multi-scale navigation intentions of manned pilot vessels are determined. By predicting the multi-scale navigation intentions of manned pilot vessels, their short-term navigation trajectory can be determined. Based on the multi-scale navigation intentions of manned pilot vessels and a mixed formation model, combined with a distributed model predictive control framework, formation navigation control of unmanned follower vessels is carried out. A closed-loop stability strategy based on distributed model predictive control is designed to ensure the stable operation of manned pilot mixed vessel formations. This provides a solution for the formation control of mixed vessels with unpredictable relative positions of manually piloted and intelligent vessels in busy waterways, filling the research gap in mixed formation control of water transportation. As an intermediate transitional solution before the arrival of full-scale intelligentization, it is of great significance for reducing crew ratios and improving efficiency in water transportation. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A schematic flowchart of a ship formation control method provided in an embodiment of the present invention; Figure 2 A schematic diagram of a ship model with degrees of freedom provided in an embodiment of the present invention; Figure 3 A diagram illustrating a ship formation communication topology provided in an embodiment of the present invention; Figure 4A schematic diagram of a ship formation provided in an embodiment of the present invention; Figure 5 A flowchart illustrating a multi-scale navigation intent determination method provided in an embodiment of the present invention; Figure 6 A schematic diagram of a formation navigation control process provided in an embodiment of the present invention; Figure 7 This is a schematic diagram of the structure of a ship formation control device provided in an embodiment of the present invention; Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0026] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0027] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0028] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0029] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0030] A specific embodiment of the present invention, such as Figure 1 As shown, a method for controlling ship formations is disclosed, including: S101: A dynamic model of the pilot ship is constructed based on satellite positioning data and propeller data of the manned pilot ship, a dynamic model of the follower ship is constructed based on hydrodynamic parameters of the unmanned follower ship, and a mixed formation model of ships is constructed based on the dynamic models of the pilot ship and the unmanned follower ship.

[0031] In this embodiment of the invention, the provided ship formation control method is applicable to the navigation control of a mixed ship formation consisting of one manned pilot ship and multiple unmanned follower ships. First, typical navigation scenarios are analyzed, selecting typical inland waterway and coastal traffic scenarios. Using machine learning and other technologies, a data-driven, reliable, high-dimensional pilot ship model is constructed. Then, based on hydrodynamic analysis and rigid body motion models, a mechanism-driven general follower ship model is established. Finally, a mixed formation model is formed, providing model support for intent recognition and the acquisition of temporal prediction information for control. Specifically, by selecting inland waterway and coastal traffic scenarios, satellite positioning data and propeller / rudder data are extracted, cleaned, and interpolated at the second level. Using the input propeller / rudder data and output ship state information, a deep neural network is used to train and model the dynamics of the manned pilot ship.

[0032] Furthermore, for ship dynamics models, traditional modeling methods systematically describe the ship's motion behavior in six degrees of freedom: roll, pitch, sway, pitch, bow, and heave. When a ship is sailing on water, the main degrees of freedom are concentrated in the horizontal plane. Vertical motion is significantly affected by waves and is not significant for most inland waterway vessels or ocean-going vessels navigating in calm water. Therefore, ship motion models are usually simplified to three degrees of freedom in the horizontal plane: pitch, sway, and bow, as shown below. Figure 2 As shown, a follower ship dynamics model can be constructed based on this.

[0033] Furthermore, after determining the dynamics models of the manned pilot ship and the unmanned follower ship, a mixed formation model of ships is constructed by formally unifying the dynamics models of the manned pilot ship and the unmanned follower ship.

[0034] S102, determine the multi-scale navigation intention of the manned pilot vessel based on the historical navigation data and real-time maneuvering data of the manned pilot vessel.

[0035] In this embodiment of the invention, the manned pilot vessel is responsible for identifying the navigation situation and undertaking navigation tasks, and broadcasting its status information to the unmanned follower vessels. The unmanned follower vessels adjust their own motion state by receiving information from the manned pilot vessel or other unmanned follower vessels to maintain the formation structure. Therefore, the navigation intention of the manned pilot vessel is crucial. Satellite positioning data and historical maneuvering information can be used, combined with Markov decision methods, to identify the micro-behaviors of the pilot vessel, such as turning, deceleration, or acceleration, based on short-term trajectory information and observation data. Secondly, historical satellite positioning data is analyzed and clustered to establish a macro-trajectory probabilistic prediction model. By fusing the behavioral intention with the trajectory fitting model, the pilot vessel's maneuvering intention is predicted. The specific method for determining the multi-scale navigation intention of the manned pilot vessel will be described in detail later in this invention.

[0036] S103 uses a distributed model predictive control framework to control the formation navigation of unmanned follower ships based on the multi-scale navigation intentions of manned pilot ships and the ship hybrid formation model.

[0037] In this embodiment of the invention, after determining the multi-scale navigation intentions of the manned pilot vessel, its short-term navigation trajectory can be predicted. Then, based on the predicted trajectory and a ship formation model combined with a distributed model predictive control framework, formation navigation control is performed on the unmanned follower vessel. Figure 3 As shown, under the local information sharing mechanism, a distributed model predictive control framework is designed based on a nonlinear formation model. Based on the cooperative tracking and collision avoidance tasks, optimization objectives and coupling constraints are constructed. Using Lyapunov stability theory, a backstepping controller is designed, and the backstepping control law is transformed into a contraction constraint to constrain the output action, thereby ensuring system stability. Furthermore, recursive feasibility conditions are explored to ensure that the control method has a solution, thus realizing the control of mixed ship formations.

[0038] The ship formation control method provided by this invention constructs a dynamic model of the pilot ship based on satellite positioning data and propeller and rudder data of the manned pilot ship, constructs a dynamic model of the follower ship based on the hydrodynamic parameters of the unmanned follower ship, and constructs a mixed ship formation model based on the dynamic models of the pilot ship and the unmanned follower ship. By modeling the manned pilot ship and the unmanned follower ship separately, it can adapt to the navigation characteristics of the manned pilot ship and the unmanned follower ship, and construct a mixed ship formation model to facilitate subsequent control of the entire mixed ship formation. Based on historical navigation data and real-time maneuvering data of manned pilot vessels, the multi-scale navigation intentions of manned pilot vessels are determined. By predicting the multi-scale navigation intentions of manned pilot vessels, their short-term navigation trajectory can be determined. Based on the multi-scale navigation intentions of manned pilot vessels and a mixed formation model, combined with a distributed model predictive control framework, formation navigation control of unmanned follower vessels is carried out. A closed-loop stability strategy based on distributed model predictive control is designed to ensure the stable operation of manned pilot mixed vessel formations. This provides a solution for the formation control of mixed vessels with unpredictable relative positions of manually piloted and intelligent vessels in busy waterways, filling the research gap in mixed formation control of water transportation. As an intermediate transitional solution before the arrival of full-scale intelligentization, it is of great significance for reducing crew ratios and improving efficiency in water transportation.

[0039] In some possible embodiments of the present invention, the pilot ship dynamics model is as follows:

[0040]

[0041]

[0042] in, Let K+1 be the state of the ship with a pilot. For the first fully connected deep neural network in a feedforward network l The weight matrix of the layer, For activation function, For the first fully connected deep neural network in a feedforward network l Layer bias vector, The position of the piloted ship at time k. Let K be the heading angle of the piloted vessel at time k. Let be the longitudinal velocity of the manned ship at time k. Let be the lateral velocity of the piloted ship at time k. Let be the bow roll rate of the manned vessel at time k. Let $k$ be the propeller speed of the manned ship at time $k$. The rudder angle of a manned ship at time k.

[0043] In this embodiment of the invention, when constructing the dynamic model of a manned pilot vessel, the pilot vessel's propeller and rudder control inputs are:

[0044] in, Let $k$ be the propeller speed of the manned ship at time $k$. Let be the rudder angle of a manned ship at time k. For a ship with multiple propellers and multiple rudders, this is extended to a vector.

[0045]

[0046] in, The position of the piloted ship at time k. Let K be the heading angle of the piloted vessel at time k. Let be the longitudinal velocity of the manned ship at time k. Let be the lateral velocity of the piloted ship at time k. Let be the bow roll rate of the manned vessel at time k.

[0047] The dynamic model of the deep neural network is then:

[0048] in, For a manned, piloted ship, the nonlinear function is represented by a deep neural network. This is a set of network parameters.

[0049] An L-layer feedforward fully connected deep neural network is used, with the activation function being... Input layer and hidden layer l =1,2,..., L -1 represents:

[0050]

[0051] in, For the first fully connected deep neural network in a feedforward network l The weight matrix of the layer, For the first fully connected deep neural network in a feedforward network l The bias vector of the layer.

[0052] The output layer is:

[0053] The results of integration are as follows:

[0054] This invention combines deep neural networks to construct a dynamic model of a manned pilot vessel, which facilitates subsequent formation control.

[0055] In some possible embodiments of the present invention, the following ship dynamics model is as follows:

[0056] in, To determine the state of the following ship at time k+1, This is the rotation matrix used to transform the velocity vector in the hull coordinate system to the inertial coordinate system. Let M be the velocity vector of the following ship at time k+1, and M be the inertia matrix of the following ship. For the Coriolis and centripetal force matrices of the following ship, For the damping matrix of the following ship, The parameters of the following ship at time k are the hydrodynamic parameters.

[0057] In this embodiment of the invention, for the construction of the dynamic model of the unmanned following vessel, the following vessel adopts the classic Fossen model. Considering that the hydrodynamic parameters of the autonomous following vessel are known, its hydrodynamic model architecture is as follows:

[0058]

[0059] Where, vector This represents the ship's actual trajectory in the Earth's inertial coordinate system, consisting of the ship's position (x, y) and heading angle. Composition, vector This represents the velocity vector of the ship in the ship's coordinate system, with variables u, v, and r representing the forward velocity (swell), lateral velocity (sway), and angular velocity (bow roll), respectively. To control the input vector, The expression for transforming the velocity vector from the hull coordinate system to the inertial coordinate system is as follows:

[0060] Matrix M is the ship's inertia matrix, derived from the rigid body mass matrix. and additional mass matrix Composition, that is

[0061]

[0062]

[0063] Coriolis and centripetal force matrix Rigid body Coriolis force and centripetal moment generated by rotation of the hull coordinate system and the Coriolis and centripetal force matrix of the added mass Composition, can be written as:

[0064]

[0065] Damping matrix This describes the resistance effect between the ship's hull and the water. To simplify the model, the damping matrix is ​​used in this study. With appropriate simplification, retaining only the constant term that is linearly related to velocity, its expression is:

[0066] Where m is the mass of the ship. The coordinates of the ship's center of gravity on the x-axis of the ship's coordinate system. Let z be the moment of inertia of the ship about the z-axis at its center of gravity. , , , as well as This is the hydrodynamic derivative, used to characterize the linear effect of hydrodynamics on each degree of freedom of motion.

[0067] Furthermore, a mixed formation model is formed by manned pilot ship models and autonomous follower ship models. To ensure uniformity in form, the autonomous follower ship models can be standardized as follows:

[0068] The discrete form is:

[0069] in This represents a discretization mapping.

[0070] Based on the above method for establishing ship motion models, a mathematical model of a three-degree-of-freedom ship train formation system is constructed:

[0071] when i =0 means the ship is the pilot.

[0072] Under a leader-follower control strategy, a ship formation typically consists of a lead ship and several follower ships. The lead ship is responsible for identifying the navigational situation and providing navigation, broadcasting its status information to the follower ships. The follower ships adjust their own motion by receiving information from the lead ship and other follower ships to maintain the formation structure. The internal communication within the formation directly affects the efficiency of information transmission and the stability of the formation. The objective of formation control is to develop a formation control strategy for each ship, enabling each ship to track the preceding ship at a desired distance and angle, while ensuring that each follower ship remains within an energy-efficient range relative to the preceding ship and avoids collisions. Figure 4 As shown, the distance between two consecutive ships in a formation is defined. and angle As shown below:

[0073]

[0074] Where atan2 is the arctangent function. Using the above equation, we can obtain:

[0075]

[0076] This invention provides a specific construction process for a mixed ship formation model, providing a model foundation for subsequent ship formation control.

[0077] In some possible embodiments of the present invention, such as Figure 5 As shown, the multi-scale navigation intentions of a manned pilot are determined based on historical navigation data and real-time maneuvering data, including: S501, acquire historical satellite positioning data of the manned navigator in the current waters and historical maneuvering data corresponding to the historical satellite positioning data, and obtain historical trajectory data of the manned navigator; S502: Obtain real-time maneuvering data and real-time trajectory status of the manned pilot vessel, and combine historical trajectory data to determine the micro-operational behavior of the manned pilot vessel. S503, extract historical trajectory features from historical trajectory data, combine real-time manipulation data and historical trajectory features to cluster the future trajectory of manned pilot vessels, and determine the macroscopic trajectory category of manned pilot vessels; S504 integrates micro-operational behaviors and macro-trajectory categories to obtain the multi-scale navigation intentions of manned pilot ships.

[0078] In this embodiment of the invention, the identification of multi-scale navigation intentions of manned pilot vessels can be performed from multiple dimensions. One dimension is micro-behavioral identification, which involves acquiring historical satellite positioning data and corresponding manipulation information within the target waters, and then cleaning, removing anomalies, aligning the time, and interpolating the data at the second level to construct a trajectory data sequence at a unified sampling time. The resulting trajectory data of a single pilot vessel can be represented as follows:

[0079] in, Let t be the position of the ship. For the course, , as well as For speed information, This is a historical manipulation command.

[0080] Within a short time window, by utilizing changes in trajectory state and maneuvering commands, combined with the Markov decision process method, the micro-maneuvering behaviors of the pilot vessel, including turning, acceleration, deceleration, and maintaining course, are identified. The state of the pilot vessel is defined as:

[0081] Among them, actions This includes left turn, right turn, acceleration, deceleration, and holding, followed by building the transfer rate:

[0082] Identifying the current behavior type based on maximum a posteriori or maximum likelihood methods:

[0083] To obtain the micro-level operational behavior of the ship with a navigator at every moment.

[0084] Furthermore, feature extraction was performed on the satellite positioning trajectories of multiple historical voyages, including indicators such as flight path shape, rate of change of heading, rate of change of speed, and lateral offset. Clustering algorithms were then used to cluster the trajectories from multiple voyages, resulting in several typical macroscopic trajectory patterns, denoted as:

[0085] Obtain the macroscopic trajectory category of the pilot ship .

[0086] Based on the trajectory clustering results, a Gaussian mixture model of trajectory probabilistics is constructed to predict the probability of future trajectories.

[0087] in, The model outputs the trajectory probability distribution or predicted mean and variance over several future time steps.

[0088] By fusing micro-level behavioral labels with macro-level trajectory categories and trajectory probability models, a joint prediction model is constructed:

[0089] By using Bayesian updates to achieve joint decision-making between the behavior layer and the trajectory layer, the future maneuvering intention of the pilot ship can be obtained.

[0090] This invention analyzes the navigation intentions of a manned pilot vessel from the perspectives of microscopic operational intent and macroscopic trajectory categories, enabling precise identification of the pilot vessel's trajectory within a short period of time.

[0091] In some possible embodiments of the present invention, such as Figure 6 As shown, a distributed model predictive control framework is used to control the formation navigation of unmanned follower ships based on the multi-scale navigation intentions of manned pilot ships and a mixed formation model of ships, including: S601, in the preset distributed model predictive control framework, takes the minimum error between the ship's operating state and the reference trajectory state, and the distance and angle difference between each adjacent unmanned follower ship as the objective, and constructs the objective cost function of each unmanned follower ship by combining the multi-scale navigation intention of the manned pilot ship and the ship mixed formation model. S602, the contraction constraints of the ship mixed formation model are determined based on the objective cost function combined with the inverse step control law; S603 is a design of a formation controller for unmanned follower ships based on contraction constraints, and a formation navigation control for unmanned follower ships based on the formation controller.

[0092] In this embodiment of the invention, the objective function design for single-ship distributed model predictive control aims to minimize the error between the ship's motion state and the reference trajectory state, as well as the control input. For multi-ship distributed model predictive control, the objective function design also needs to consider collision avoidance costs and formation maintenance factors. Therefore, the objective cost function for each following ship i in simulation step k is defined as follows:

[0093] in, For the first i The target cost function of the ships To predict the step size, N is the number of ships in the mixed fleet. Indicates the first i The ship at the moment k The state vector, Indicates the first i The ship at the predicted time k + s The state error relative to the reference trajectory, and They represent the first i The ship at the predicted time k + s The actual relative position vector and the estimated position vector, and They represent the first j The ship at the predicted time k + s The actual relative position vector and the estimated position vector, This represents the cost function used to avoid collisions between ships, with the input being the first... i Ship and the j Spatial position estimation of the ship at the predicted time. For the control input of the i-th ship at the predicted time k+s, It is a positive definite matrix. and This is a preset weighted matrix.

[0094] For the i-th ship, using the above objective cost function, the optimization problem is designed as follows:

[0095] For the aforementioned distributed model predictive control, the pilot ship needs to obtain the predicted trajectory information through intent recognition to predict the state information in the time domain, and then input it into a deep learning network to obtain the state information in the time domain and share it with the following ships.

[0096] Based on the Fossen model of the following ship and the derivation principle of backstepping control, the backstepping control input can be derived as follows:

[0097]

[0098] in, and Since it is a positive definite matrix, it can be improved by adjusting appropriate parameters. and This is used solely for ship trajectory tracking. Furthermore, transforming the control law into a distributed model predictive control constraint ensures the global exponential stability of the tracking system, provided that all parameter matrices are in the positive definite range. The contraction constraint can be specifically expressed as:

[0099]

[0100]

[0101]

[0102] For the control of nonlinear systems, the traditional approach of ensuring the stability of distributed model predictive control (DMDC) by designing terminal constraint sets and terminal cost functions is highly complex, significantly increasing the computational complexity of solving the DMDC optimization problem and making real-time computation difficult. To overcome these shortcomings, a DMDC method based on Lyapunov theory is adopted. This method replaces terminal constraints with contraction constraints, directly applying stability constraints to the first step of DMDC implementation. The added constraints constitute sufficient conditions to guarantee the stability of the closed-loop system. By adjusting... and To ensure recursive feasibility, the stability collaborative control model is shown below:

[0103]

[0104] Among them, among them, Let be the discrete-time system dynamics equation for the i-th ship. Is this moment for k + s Predicted state at any given time Is this moment for k + s Predictive control input at time, Let k be the predicted state of the system at time k. The actual state of the system. and Indicates the upper and lower bounds of the input. For the ship i With the ship j The distance between them and For the ship i With the ship j Distance constraints between them Indicates state tracking error. This represents the first and current input in the action sequence. This is the rotation matrix used to transform the velocity vector in the hull coordinate system to the inertial coordinate system. Let be the velocity vector of ship i at time k. Let M be the Coriolis and centripetal force matrices of ship i, and M be the inertia matrix of ship i. Let i be the damping matrix of ship i. Let be a positive definite matrix, where, It is a stable function vector, whose function is to design virtual control inputs to stabilize the system's state variables at the desired values. It is a positive definite matrix. To track state errors, This represents the current state of ship i. For reference trajectory status, , These are the first and second derivatives of the reference trajectory state, respectively. The first equality constraint represents the dynamic constraint of the system. The state evolution of the controlled object is determined by the discrete-time system dynamics equations. Given that the second constraint equation specifies the initial conditions of the optimization problem, the third inequality is the control input constraint, and the last inequality is the Lyapunov stability constraint.

[0105] This invention addresses the challenges of strong uncertainty in the behavior of manned pilot vessels, significant time delays in ship maneuvering, and strong interference in convoy navigation by conducting research on distributed stability control based on the pilot vessel's intent perception. The convoy system can quantify and predict the behavior of manned vessels, enabling efficient imitation, following, and coordination of ship formations under manned pilot guidance without requiring upgrades to existing traditional vessels.

[0106] To better implement the ship formation control method in the embodiments of the present invention, based on the ship formation control method, correspondingly, such as Figure 7 As shown, this embodiment of the invention also provides a ship formation control device, the ship formation control device 700 including: Model building module 701 is used to build a pilot ship dynamic model based on satellite positioning data and propeller data of the manned pilot ship, build a follower ship dynamic model based on hydrodynamic parameters of the unmanned follower ship, and build a mixed formation model of ships based on the pilot ship dynamic model and the unmanned follower ship dynamic model. The manned pilot vessel intent estimation module 702 is used to determine the multi-scale navigation intent of the manned pilot vessel based on the historical navigation data and real-time maneuvering data of the manned pilot vessel. The formation control module 703 is used to control the formation navigation of unmanned follower ships based on the multi-scale navigation intentions of the manned pilot ship and the ship hybrid formation model combined with a distributed model predictive control framework.

[0107] The ship formation control device 700 provided in the above embodiments can realize the technical solutions described in the above ship formation control method embodiments. The specific implementation principles of each module or unit can be found in the corresponding content in the above ship formation control method embodiments, and will not be repeated here.

[0108] like Figure 8 As shown, the present invention also provides an electronic device 800. The electronic device 800 includes a processor 801, a memory 802, and a display 803. Figure 8 Only some components of the electronic device 800 are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.

[0109] In some embodiments, processor 801 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in memory 802 or process data, such as the ship formation control method of the present invention.

[0110] In some embodiments, processor 801 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 801 may be local or remote. In some embodiments, processor 801 may be implemented on a cloud platform. In some embodiments, the cloud platform may include private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, internal cloud, multi-cloud, etc., or any combination thereof.

[0111] In some embodiments, memory 802 may be an internal storage unit of electronic device 800, such as a hard disk or memory of electronic device 800. In other embodiments, memory 802 may also be an external storage device of electronic device 800, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 800.

[0112] Furthermore, the memory 802 may include both internal storage units of the electronic device 800 and external storage devices. The memory 802 is used to store application software and various types of data installed on the electronic device 800.

[0113] In some embodiments, display 803 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 803 is used to display information from electronic device 800 and to display a visual user interface. Components 801-803 of electronic device 800 communicate with each other via a system bus.

[0114] In some embodiments, when processor 801 executes the ship formation control program in memory 802, the following steps may be performed: A dynamic model of the pilot ship is constructed based on satellite positioning data and propeller data of the manned pilot ship. A dynamic model of the follower ship is constructed based on hydrodynamic parameters of the unmanned follower ship. A mixed formation model of ships is constructed based on the dynamic models of the pilot ship and the unmanned follower ship. The multi-scale navigation intentions of the manned pilot vessel are determined based on its historical navigation data and real-time maneuvering data. This paper describes a distributed model predictive control framework for controlling the formation navigation of unmanned follower ships based on the multi-scale navigation intentions of manned pilot ships and the ship hybrid formation model.

[0115] It should be understood that when the processor 801 executes the ship formation control program in the memory 802, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.

[0116] Furthermore, this embodiment of the invention does not specifically limit the type of electronic device 800 mentioned. Electronic device 800 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the invention, electronic device 800 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0117] Accordingly, this application also provides a computer-readable storage medium for storing computer-readable programs or instructions. When the programs or instructions are executed by a processor, they can implement the steps or functions of the ship formation control methods provided in the above-described method embodiments.

[0118] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0119] 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 changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for controlling ship formations, characterized in that, include: A pilot ship dynamics model is constructed based on satellite positioning data and propeller data of the manned pilot ship, a follower ship dynamics model is constructed based on hydrodynamic parameters of the unmanned follower ship, and a mixed formation model of ships is constructed based on the pilot ship dynamics model and the unmanned follower ship dynamics model. The multi-scale navigation intention of the manned pilot vessel is determined based on its historical navigation data and real-time maneuvering data. The unmanned follower vessel is controlled to navigate in formation based on the multi-scale navigation intentions of the manned pilot vessel and the ship hybrid formation model combined with a distributed model predictive control framework.

2. The ship formation control method according to claim 1, characterized in that, The pilot ship's dynamics model is as follows: in, Let K+1 be the state of the ship with a pilot. For the first fully connected deep neural network in a feedforward network l The weight matrix of the layer, For activation function, For the first fully connected deep neural network in a feedforward network l Layer bias vector, The position of the piloted ship at time k. Let K be the heading angle of the piloted vessel at time k. Let be the longitudinal velocity of the manned ship at time k. Let be the lateral velocity of the piloted ship at time k. Let be the bow roll rate of the manned vessel at time k. Let $k$ be the propeller speed of the manned ship at time $k$. The rudder angle of a manned ship at time k.

3. The ship formation control method according to claim 1, characterized in that, The following ship's dynamics model is: in, To determine the state of the following ship at time k+1, This is the rotation matrix used to transform the velocity vector in the hull coordinate system to the inertial coordinate system. Let M be the velocity vector of the following ship at time k+1, and M be the inertia matrix of the following ship. For the Coriolis and centripetal force matrices of the following ship, For the damping matrix of the following ship, The parameters of the following ship at time k are the hydrodynamic parameters.

4. The ship formation control method according to claim 1, characterized in that, The determination of the multi-scale navigation intention of the manned pilot vessel based on its historical navigation data and real-time maneuvering data includes: The historical satellite positioning data of the manned pilot vessel in the current waters and the historical maneuvering data corresponding to the historical satellite positioning data are obtained to obtain the historical trajectory data of the manned pilot vessel. The real-time manipulation data and real-time trajectory status of the manned pilot vessel are obtained, and the micro-operational behavior of the manned pilot vessel is determined by combining the historical trajectory data. Extract the historical trajectory features from the historical trajectory data, and combine the real-time manipulation data and the historical trajectory features to cluster the future trajectory of the manned pilot vessel to determine the macroscopic trajectory category of the manned pilot vessel; By fusing the micro-operational behaviors and the macro-trajectory categories, the multi-scale navigation intention of the manned pilot vessel is obtained.

5. The ship formation control method according to claim 1, characterized in that, The formation navigation control of the unmanned follower vessel based on the multi-scale navigation intention of the manned pilot vessel and the ship hybrid formation model combined with the distributed model predictive control framework includes: In the preset distributed model predictive control framework, the objectives are to minimize the error between the ship's operating state and the reference trajectory state, and to keep the distance and angle difference between each adjacent unmanned follower ship within a preset range. The objective cost function of each unmanned follower ship is constructed by combining the multi-scale navigation intention of the manned pilot ship and the ship hybrid formation model. The contraction constraints of the ship mixed formation model are determined based on the objective cost function combined with the backstepping control law; Based on the contraction constraint, a formation controller is designed for the unmanned follower vessel, and the formation controller is used to control the unmanned follower vessel's formation navigation.

6. The ship formation control method according to claim 5, characterized in that, The objective cost function is: in, For the first i The target cost function of the ships To predict the step size, N is the number of ships in the mixed fleet. Indicates the first i The ship at the moment k The state vector, Indicates the first i The ship at the predicted time k + s The state error relative to the reference trajectory, and They represent the first i The ship at the predicted time k + s The actual relative position vector and the estimated position vector, and They represent the first j The ship at the predicted time k + s The actual relative position vector and the estimated position vector, This represents the cost function used to avoid collisions between ships, with the input being the first... i Ship and the j Spatial position estimation of the ship at the predicted time. For the control input of the i-th ship at the predicted time k+s, This represents the communication connection between ship i and ship j; a value of 1 indicates a communication connection exists, while a value of 0 indicates no connection exists. It is a positive definite matrix. and This is a preset weighted matrix.

7. The ship formation control method according to claim 6, characterized in that, The formation controller is: in, Let be the discrete-time system dynamics equation for the i-th ship. Is this moment for k + s Predicted state at any given time Is this moment for k + s Predictive control input at time, Let k be the predicted state of the system at time k. The actual state of the system. and Indicates the upper and lower bounds of the input. For the ship i With the ship j The distance between them and For the ship i With the ship j Distance constraints between them Indicates state tracking error. This represents the first and current input in the action sequence. This is the rotation matrix used to transform the velocity vector in the hull coordinate system to the inertial coordinate system. Let be the velocity vector of ship i at time k. Let M be the Coriolis and centripetal force matrices of ship i, and M be the inertia matrix of ship i. Let i be the damping matrix of ship i. It is a positive definite matrix.

8. A ship formation control device, characterized in that, include: The model building module is used to build a pilot ship dynamic model based on satellite positioning data and propeller data of the manned pilot ship, build a follower ship dynamic model based on hydrodynamic parameters of the unmanned follower ship, and build a ship mixed formation model based on the pilot ship dynamic model and the unmanned follower ship dynamic model. A manned pilot vessel intent estimation module is used to determine the multi-scale navigation intent of the manned pilot vessel based on its historical navigation data and real-time maneuvering data. The formation control module is used to perform formation navigation control on the unmanned follower vessel based on the multi-scale navigation intentions of the manned pilot vessel and the ship hybrid formation model combined with a distributed model predictive control framework.

9. An electronic device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the ship formation control method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the ship formation control method according to any one of claims 1 to 7.

Images