Intelligent ship monitoring method and system

Abstract

The invention relates to the technical field of intelligent navigation and control, and discloses a ship intelligent monitoring method which comprises the following steps: data acquisition: acquiring an initial state variable and real-time environment data of a ship; dynamic modeling: establishing a nonlinear dynamic model of the ship based on the collected ship state variable and control input; optimization target construction: on the basis of the dynamic model, constructing an optimization target function, and generating an optimal control strategy; solving numerical values; the invention also discloses a ship intelligent monitoring system, and the system comprises a data collection module which is used for collecting the initial state variable and the real-time environment data of the ship; a dynamic modeling module; an optimization module; a numerical solution module; a real-time monitoring module; and a simulation verification module. According to the invention, through multi-objective optimization, real-time dynamic adjustment and nonlinear dynamic modeling, optimization of high efficiency, adaptability, safety and economical efficiency of ship path planning in a complex marine environment is realized.

Description

Technical Field

[0001] The present invention relates to the field of intelligent navigation and control technology, and in particular to a method and system for intelligent monitoring of ships. Background Art

[0002] With the rapid development of the global shipping industry and the increasing importance of ships in international logistics and transportation, the demand for intelligent ship navigation and path planning has become increasingly significant. In actual navigation, ships need to cross complex marine environments and face unpredictable external interference such as ocean currents, wind and waves, and dynamic obstacles. How to improve fuel efficiency, shorten sailing time, and optimize the stability of ship operation while ensuring navigation safety has become a core issue that needs to be solved urgently. The present invention uses intelligent monitoring technology to make real-time dynamic adjustments to the ship's path planning, which can balance fuel consumption, path time, obstacle avoidance capability, and navigation stability under multi-objective optimization conditions, providing an innovative solution for the intelligent navigation of modern ships.

[0003] Research on ship path planning has made considerable progress. Some path planning methods, through static preset paths, can achieve lower fuel consumption and shorten voyage times in ideal ocean environments. Furthermore, some solutions significantly improve navigation efficiency in specific environments by optimizing a single objective. For example, some systems utilize predicted wind and current data to achieve preliminary optimization of fuel economy. These solutions perform well in relatively stable ocean environments, demonstrate considerable practicality, and lay a foundation for the field of path planning.

[0004] However, the shortcomings of the existing technology are obvious, and it is difficult to meet the actual needs of complex marine environments. The present invention proposes solutions to these problems. First, the existing technology usually focuses on the optimization of a single goal, and fails to effectively balance fuel efficiency, obstacle avoidance capability and navigation stability, resulting in it often being difficult to take into account multiple needs during actual operation. Secondly, the dynamic modeling is too simplified, and the nonlinear dynamic characteristics and external disturbances of the ship are not sufficiently considered, which makes the path planning results lack adaptability when facing complex environments. In addition, most existing solutions lack deep real-time adjustment capabilities when the environment changes. Usually, they can only perform path offsets or local corrections, and cannot re-optimize the control strategy according to the new environmental conditions. Finally, traditional numerical solution methods are inefficient in complex optimization problems, and the iterative convergence speed is slow, making it difficult to meet real-time requirements. These shortcomings greatly limit the practical application of existing solutions in dynamic and complex environments. The present invention effectively solves these problems through innovative technical means. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a method and system for intelligent ship monitoring, which solves the problems in the existing technology of insufficient single-objective optimization of path planning, poor adaptability due to simplified dynamic modeling, lack of real-time dynamic adjustment capabilities, and low efficiency of numerical solution.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for intelligent monitoring of ships, comprising the following steps: Data collection: Collect the ship's initial state variables and real-time environmental data, including the ship's speed, position, heading, surrounding ocean environment parameters, and navigation obstacle information; Dynamic modeling: Based on the collected ship state variables and control inputs, a nonlinear dynamic model of the ship is established. The dynamic model describes the motion of the ship under different propulsion forces and heading angles. Optimization target construction: Based on the dynamic model, an optimization objective function is constructed, with time, energy consumption, obstacle avoidance capability, and navigation stability as optimization targets. The optimal path value function is determined using the Hamilton-Jacobi-Bellman equation; Optimal control strategy generation: Utilizing the Pontryagin extreme value principle, based on the optimization results in the value function, the optimal propulsion force and heading angle control inputs are generated to meet the optimization objectives of the ship path planning. Numerical solution: The optimization objective is discretized in time, the ship's state evolution is calculated through forward simulation, and the discrete control sequence and state trajectory of the optimal path are determined by combining backward optimization iterations; Real-time adjustment and monitoring: Based on real-time collected environmental data and ship state variables, the optimal control input is dynamically adjusted, the ship's path planning results are simulated and verified, the control strategy is updated, and the path is optimized in real time during actual operation.

[0007] Preferably, the initial state variables include: Collect the initial position data of the ship, including the lateral position and longitudinal position; Collect the ship's initial speed and heading angle data; Collect ocean current velocity components; All collected initial data are input into the dynamic modeling module to provide basic data support for subsequent optimization goals.

[0008] Preferably, in the dynamic modeling step, the nonlinear dynamic model of the ship established includes the following contents: The change in the transverse and longitudinal position of the ship, determined by its speed and heading angle; The speed of a ship is affected by the combined effects of resistance, propulsion force and ocean current speed; The change of the ship's heading angle is determined by the rudder angle control and the external navigation environment factors.

[0009] Preferably, the optimization target construction includes: The fuel consumption target is related to a nonlinear function of propulsion power and speed; The path planning time goal is described by the shortest time optimization model; The obstacle avoidance capability goal is optimized through real-time calculation of the distribution of environmental obstacles; The navigation stability goal is to optimize the smoothness of the hull motion under the influence of wind and waves.

[0010] Preferably, the construction of the value function includes: The value function is associated with the ship’s current position, speed, and heading angle; The cost function minimizes fuel consumption at each state point while satisfying the time and obstacle avoidance constraints of path planning; The optimal path value function is generated by numerically solving the HJB equation.

[0011] Preferably, the optimal control strategy includes: The control input for optimal propulsion is determined by optimizing the relationship between fuel consumption and ship speed; The control input of the optimal heading angle is used to adjust the angle through the servo to achieve dynamic tracking of the target path; The control inputs of propulsion force and heading angle are determined by the extreme value conditions of the Hamiltonian function and satisfy the physical constraints of speed, propulsion force and rudder angle.

[0012] Preferably, the numerical solution includes: Discretize the optimization objective function in time and transform the continuous optimization problem into an optimization problem with discrete time steps; Forward simulate the evolution of ship state variables and calculate discrete state trajectories; Backward optimization iteratively updates the control variables and updates the optimal control inputs of propulsion force and heading angle through the adjoint equations; The optimal path planning result is determined through numerical iteration of discrete time series.

[0013] Preferably, the real-time adjustment and monitoring includes: Update the value function through real-time collected environmental data; Dynamically adjust the constraints in the optimization objective function to ensure the real-time performance of path planning; The digital twin system is used to conduct real-time simulation verification of the ship's path planning scheme and update the control strategy during operation.

[0014] Preferably, in the real-time adjustment and monitoring, the simulation verification based on the digital twin system includes: Simulation verifies whether the ship's fuel consumption meets the optimization target; Simulation evaluation to determine whether the speed, heading and hull stability meet safety requirements; The optimization results are iteratively adjusted multiple times through simulation data to generate the final control strategy.

[0015] The present invention also provides a ship intelligent monitoring system, comprising: Data acquisition module, used to collect the ship's initial state variables and real-time environmental data; Dynamic modeling module, used to establish a nonlinear dynamic model of the ship based on the ship's state and environmental data; The optimization module is used to construct a cost function and generate the optimal control strategy of the ship through the Hamilton-Jacobi-Bellman equation and Pontryagin extreme value principle with the goal of minimizing fuel consumption; Numerical solution module, used to generate optimal path planning results through discretization, forward simulation and backward optimization; Real-time monitoring module, used to dynamically adjust speed, heading angle and propulsion force, and verify the optimized path in real time; The simulation verification module is used to verify the feasibility and economy of the optimization path through digital twin technology.

[0016] The present invention provides a method and system for intelligent ship monitoring, which has the following beneficial effects: 1. This invention utilizes a real-time adjustment and monitoring technology based on dynamic environmental data to achieve dynamic optimization of vessel path planning. This technology enables rapid response to external disturbances in complex and changing marine environments. Compared to existing solutions that lack the ability to correct path planning results in real time, this solution addresses the challenges of path deviation and increased safety risks caused by environmental changes.

[0017] 2. This invention accurately simulates the motion of a ship by constructing a nonlinear dynamic model and combining it with a multi-objective optimization function. This technology achieves the technical effect of balancing multiple objectives, including fuel efficiency, path time, obstacle avoidance, and navigation stability. Existing technical solutions often struggle to simultaneously address these multiple objectives. This solution addresses the issues of low operational efficiency and insufficient stability of ships in complex scenarios.

[0018] 3. This invention utilizes the Hamilton-Jacobi-Bellman equation and the Pontryagin extremum principle to generate an optimal control strategy, achieving precise control inputs for propulsion force and heading angle. This achieves the technical effect of optimizing path planning results while reducing computational complexity. Compared to existing technologies that rely on single-objective optimization, this overcomes the shortcomings of inflexible optimization results and lack of precision in control inputs.

[0019] 4. By combining time discretization with numerical solution methods and employing an iterative calculation strategy combining forward simulation with backward optimization, this invention generates highly accurate discrete control sequences and state trajectories. This technology achieves the technical effect of real-time adjustment of path planning in complex dynamic environments. Compared to existing discrete solution solutions with insufficient precision, this approach overcomes the problems of insufficiently detailed and inefficient path adjustment in dynamic calculations. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a flow chart of the method of the present invention; Figure 2 Schematic diagram of the system of the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the present specification. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0022] Please see the attached Figure 1 , an embodiment of the present invention provides a method for intelligent monitoring of ships, comprising the following steps: S1. Data collection: Collect the ship's initial state variables and real-time environmental data, including the ship's speed, position, heading, surrounding ocean environment parameters, and navigation obstacle information; The data acquisition module acquires the ship's initial state variables and relevant parameters of the surrounding environment in real time. These data, serving as input for dynamic modeling and optimization calculations, directly impact the accuracy of subsequent models and the effectiveness of path planning optimization. Generally, data acquisition must include information on the ship's operating status and external environmental interference, ensuring the integrity and real-time nature of the collected data to meet the input data requirements of dynamic modeling and optimization calculations.

[0023] Data collection includes the ship's initial state variables and real-time environmental data, specifically the ship's speed, position, heading, surrounding ocean environment parameters, and navigation obstacle information. This data is typically acquired through modern sensor systems, navigation equipment, and external data sources.

[0024] In some embodiments, the initial state variables are collected via a shipboard navigation system. Specifically, the ship's current position, including its lateral position x0 and longitudinal position y0, can be obtained via a global positioning system (GPS) module. The initial speed v0 is measured by an inertial navigation system (INS) or speedometer mounted on the ship. The initial heading angle θ0 is provided by the ship's electronic compass or inertial measurement unit (IMU). These data constitute a basic description of the ship's state at a given moment.

[0025] As an option, the acquisition of ocean current velocity components usually relies on external ocean observation systems or ship-borne ocean current velocity sensors. Specifically, the lateral velocity component u c and the longitudinal velocity component v c This can be measured using Doppler current profilers (ADCPs). In one possible implementation, these sensors are attached to the bottom of the vessel and measure the speed and direction of ocean currents by reflecting sound waves. Furthermore, near port areas, these measurements can be supplemented with real-time current data provided by the port.

[0026] The real-time environmental data collection also includes information on wind speed and wave height. s and wind direction angle φ w It is obtained through the ship's meteorological sensors. Specifically, wind speed data is used to estimate the ship's resistance during the modeling process. w It is measured by ship-borne wave height meters or radar systems to assess the ship's navigation stability.

[0027] In some embodiments, navigation obstacle information is collected using a shipboard radar system and imaging sensors. The radar system detects surrounding obstacles and returns information on their location, size, and direction of movement. This data is further processed using an image processing algorithm to generate a spatial distribution map of obstacles. This obstacle information is then input into the optimization module as part of the subsequent obstacle avoidance optimization objectives.

[0028] During the data integration process, the collected data are processed into a standardized format to facilitate subsequent dynamic modeling. Specifically, all initial state variables are represented as vectors: X0=[x0,y0,v0,θ0] in: x0: initial transverse position of the ship, in meters; y0: initial longitudinal position of the ship, in meters; v0: the initial speed of the ship, in meters per second; θ0: The initial heading angle of the ship, in radians.

[0029] Real-time environmental data is also represented in vector form: E=[u c ,v c ,w s ,φ w ,h w ] in: u c ,v c : the lateral and longitudinal velocity components of the ocean current, in meters per second; w s : wind speed, in meters per second; φ w : wind direction angle, in radians; h w : Wave height in meters.

[0030] In one possible implementation, the data acquisition module can dynamically adjust during operation. For example, when a vessel enters an area with a high density of obstacles, the radar's scanning frequency will automatically increase to obtain more detailed obstacle distribution data. Simultaneously, the frequency of real-time environmental data collection will also dynamically adjust with dramatic changes in the external environment, thus meeting the needs of real-time optimization.

[0031] Typically, the data acquisition module needs to seamlessly integrate with the dynamic modeling module. In this embodiment, the collected data is directly transmitted to the dynamic modeling module via a communication interface as input parameters for the modeling. For example, the ship's speed, position, and heading data are used to construct the ship's displacement equation, while ocean current velocity components and wind speed data are used to calculate the environmental resistance to the ship. Furthermore, obstacle information is used as part of the modeling boundary conditions to simulate the ship's dynamic behavior in complex environments.

[0032] In some embodiments, to improve data accuracy and reliability, the data acquisition module filters the input data. For example, GPS data may contain noise, so the location data can be smoothed using a Kalman filter. Wind speed and wave height measurement data can also be filtered using a low-pass filter to remove high-frequency noise. Furthermore, obstacle detection data is accumulated over multiple scans to improve obstacle detection accuracy.

[0033] S2. Dynamic Modeling: Based on the collected ship state variables and control inputs, a nonlinear dynamic model of the ship is established. The dynamic model describes the motion of the ship under different propulsion forces and heading angles. The dynamic modeling step establishes a nonlinear dynamic model of the ship under different propulsion and heading angles. Based on the state variables and real-time environmental parameters acquired during the data acquisition step, this model comprehensively considers the ship's motion patterns and external interference factors, describing characteristics such as displacement changes, speed dynamics, and heading changes. Generally, this model needs to fully reflect the complexities of actual operation while ensuring computational feasibility and accuracy. This step provides the necessary physical foundation for the subsequent construction of optimization objectives.

[0034] In this example, the dynamic modeling is based on the following state variables and control inputs: The state variables include the ship’s transverse position x(t), longitudinal position y(t), speed v(t), and heading angle θ(t); The control input includes the thrust force F thrust (t) and the rate of change of heading angle ω(t).

[0035] The motion of a ship in a two-dimensional plane can be described by the following nonlinear dynamic equation: in, and where represents the rate of change of the ship's position in the transverse and longitudinal directions, respectively. The ship's velocity v(t) and heading angle θ(t) determine the ship's trajectory. Alternatively, the above equations can be discretized to facilitate numerical computation.

[0036] In some embodiments, the dynamic change of the ship's speed is determined by the propulsion force and the environmental resistance, which can be specifically modeled as follows: m is the mass of the ship in kilograms; α is the nonlinear resistance coefficient of interaction with water, in Newton / (m / s) 2 ; β is the linear resistance coefficient, in Newton / (m / s); F thrust (t) is the propulsion force generated by the ship's engine, in Newtons.

[0037] In general, the values ​​of α and β can be calibrated according to the specific design parameters or experimental data of the ship. For example, when the ship is sailing in a calm sea, the resistance is mainly determined by the linear term βv(t), while in a high sea condition, the nonlinear term αv 2 The impact of (t) is even more significant.

[0038] As an option, the influence of ocean current velocity cannot be ignored in the ship speed modeling. In one possible implementation, the lateral component u of the ocean current velocity is c and the longitudinal component v c is added to the ship's velocity to obtain the actual relative velocity. At this point, the speed dynamic equation can be further described as: In this formula, v eff (t) represents the ship's actual effective speed relative to still water, measured in meters per second. Effective speed directly affects fuel consumption and resistance calculations.

[0039] Specifically, the change of the ship's heading angle is achieved by rudder angle control, and its dynamic equation can be modeled as: in: ω(t) is the rate of change of heading angle, in radians per second.

[0040] In some embodiments, the range of ω(t) is constrained by the physical limits of the steering gear. For example, for a typical medium-sized ship, the maximum rate of change of the rudder angle is typically between [-0.5 and 0.5] radians per second. By constraining ω(t), unrealizable control inputs in the model can be avoided.

[0041] In one possible implementation, the effects of wind speed and waves on ship motion can be modeled by external disturbance terms. Specifically, the wind speed w s and wind direction angle φ w Additional lateral and longitudinal forces will be introduced, thus changing the actual motion trajectory of the ship. When modeling, these disturbance forces can be expressed as: C w The drag coefficient is related to the surface area of ​​the ship and the wind direction, and the unit is Newton / (meter / second) 2 ; F wind,x and F wind,y are the horizontal and vertical components of the wind force, respectively, in Newtons.

[0042] These perturbation forces are added to the dynamic equations to improve the accuracy of the model.

[0043] In some embodiments, to improve computational efficiency and model applicability, the dynamic model can be simplified. For example, when a ship is operating in a stable ocean current region with minimal wind and wave impact, the disturbance term can be ignored, retaining only the basic motion and resistance relationship.

[0044] The output of the dynamic model is the ship's state evolution trajectory, specifically the time-varying changes in the lateral position x(t), longitudinal position y(t), speed v(t), and heading angle θ(t). These trajectories provide the foundational data for the subsequent optimization objectives.

[0045] Typically, the dynamics model needs to be closely integrated with the data acquisition and optimization modules. In this embodiment, the initial state variables and environmental parameters provided by the data acquisition module are directly input into the dynamics model for model initialization and disturbance term calculation. The optimization module then uses the state evolution equations generated by the dynamics model to construct a multi-objective optimization problem.

[0046] S3. Optimization Target Construction: Based on the dynamic model, an optimization objective function is constructed, with time, energy consumption, obstacle avoidance capability, and navigation stability as optimization targets. The optimal path value function is determined through the Hamilton-Jacobi-Bellman equation. Based on the dynamic model, the various needs of ship operation are comprehensively considered, and an optimization objective function is constructed to achieve a balance between time, energy consumption, obstacle avoidance capability, and navigation stability. Generally, these objectives are related to each other and may conflict, requiring a trade-off through an optimization algorithm. Ultimately, the value function of the optimal path is determined through the Hamilton-Jacobi-Bellman (HJB) equation, providing guidance for the subsequent generation of control strategies.

[0047] This step uses the dynamic modeling results as input, combined with the ship's motion characteristics and environmental factors, to gradually construct a multi-objective optimization function. This optimization objective function quantifies the ship's overall performance under different path planning conditions and ensures a global optimum within the constraints.

[0048] In this embodiment, the optimization objective function consists of the following sub-objectives: The first is the time objective, which is used to minimize the time it takes for a ship to travel from its starting point to its destination. The mathematical expression of the time objective is: in: J time : total navigation time, in seconds; T: The total sailing time of the ship from the starting point to the destination.

[0049] In general, time objectives play a crucial role in route planning, and are particularly significant in high-efficiency transport scenarios. Specifically, this objective is combined with energy consumption and obstacle avoidance objectives during the optimization process to ensure that the ship maintains high speed while meeting other performance requirements.

[0050] The second is the energy consumption objective, which is used to minimize the ship's fuel consumption. Fuel consumption is related to the ship's propulsion power and speed. The mathematical description of its objective function is: in: J fuel : total fuel consumption, in kilograms; κ1: proportionality coefficient between fuel consumption and propulsion force, in kg / N; κ2: The proportional coefficient of fuel consumption and speed cubed, in kg / (m 3 / Second); F thrust : The instantaneous propulsion force of the ship, in Newton; v(t): instantaneous speed of the ship, in meters per second.

[0051] In one possible implementation, to improve the accuracy of the fuel consumption model, the coefficients κ1 and κ2 can be calibrated based on specific ship operating data. Furthermore, numerical optimization can be used to find the optimal balance between propulsion power and speed, thereby reducing fuel consumption.

[0052] The third goal is the obstacle avoidance goal, which is used to ensure that the ship avoids collisions with obstacles during navigation. Generally, the distribution of obstacles is obtained by the data collection step, and the mathematical description of its optimization goal is: in: J obstacle : obstacle avoidance objective function; d(t): The minimum distance between the current position of the ship and the obstacle, in meters.

[0053] Specifically, when the ship approaches an obstacle, the distance d(t) decreases rapidly, thereby increasing the objective function value. This design can automatically increase the weight of obstacle avoidance during the optimization process, ensuring that the ship prioritizes staying away from obstacles. In some embodiments, in order to enhance the robustness of the model, a safety distance parameter d can also be introduced. safe , so that the optimization algorithm is close to the actual distance d safe The obstacle avoidance behavior is triggered.

[0054] Finally, there is the navigation stability objective, which is used to optimize the stability of the ship under the influence of wind and waves. Navigation stability is achieved by modeling the disturbance response of wind and waves. The mathematical description of the objective function is: in: J stability : navigation stability objective function; h w (t): The wave height disturbance at a certain moment, in meters.

[0055] Alternatively, the stability objective can further incorporate hull vibration characteristics, such as adding a response function for the hull roll or pitch angle. This improvement enables a more comprehensive assessment of the ship's dynamic stability.

[0056] In one possible implementation, in order to combine the above sub-goals into one optimization problem, the present invention uses a weighted method to integrate the objective function. The final optimization objective function is expressed as: J=w1J time +w2J fuel +w3J obstacle +w4J stability in: J: comprehensive optimization objective function; w1,w2,w3,w4: weight factors of each sub-goal, used to control the priority of different goals in the optimization process.

[0057] In general, the weighting factors can be adjusted based on the specific application scenario. For example, in an emergency transport mission, the weight of w1 can be increased, while in severe sea conditions, the priority of w4 may need to be increased.

[0058] In this embodiment, the optimization objective function is solved by the Hamilton-Jacobi-Bellman (HJB) equation. The HJB equation is a partial differential equation used to describe the dynamic changes of the value function V(X, t). Its mathematical form is: in: V(X,t): value function, which represents the minimum optimization target value from the current state X to the end point; X = [x, y, v, θ]: state variables; u=[F thrust ,ω]: control input; Hamiltonian function.

[0059] Specifically, the Hamiltonian function is defined as: in: The gradient of the value function; f(X,u): rate of change of state of the dynamic model.

[0060] By numerically solving the HJB equation, the value function of the optimal path can be generated and provide support for the subsequent optimal control strategy.

[0061] S4. Optimal control strategy generation: Using the Pontryagin extreme value principle, based on the optimization results in the value function, the optimal propulsion force and heading angle control inputs are generated to meet the optimization objectives of the ship path planning; Generating an optimal control strategy is crucial for achieving path planning optimization goals. By leveraging the Pontryagin extreme value principle and combining the optimization results of the value function, the optimal propulsion force and heading angle control inputs for the vessel are calculated. This allows the vessel to maximize fuel efficiency, minimize path time, and comprehensively improve navigation safety and stability under multi-objective optimization conditions. This step generally requires consideration of the constraints of the dynamic model and the multiple trade-offs within the optimization objective to ensure the generated control strategy is both practically feasible and physically plausible.

[0062] The core of this step is to calculate the optimal control input by constructing a Hamiltonian function and combining it with the Pontryagin extremum principle. Optionally, the control inputs include propulsion force and heading angle change rate, which directly determine the ship's speed, heading, and path planning results.

[0063] In this embodiment, the Hamiltonian function H is defined as a combination of the optimization objective function and the rate of change of the state variable, specifically: in: J: comprehensive optimization objective function, consisting of time, energy consumption, obstacle avoidance capability and navigation stability; λ x ,λ y ,λ v ,λ θ : accompanying variables, corresponding to the shadow prices of lateral position, longitudinal position, speed and heading angle respectively; The rate of change of the state variables is given by the dynamic model.

[0064] The Hamiltonian function combines the dynamic changes of state variables and the direct impact of control inputs on the optimization objective, and is the core mathematical tool for generating optimal control strategies.

[0065] Specifically, the propulsion force F thrust(t) and the heading angle change rate ω(t) are the two core control input parameters. Using the Pontryagin extremum principle, the optimal values ​​of these two parameters are obtained by differentiating the Hamiltonian function. In one possible implementation, the optimal solution for the propulsion force is calculated using the following equation: The solution is: in: Optimal propulsion force, in Newtons; m: ship mass, in kilograms; λ v : accompanying variable corresponding to speed; κ1: Proportional coefficient between fuel consumption and propulsion force.

[0066] The above formula shows that the optimal propulsion force is not only related to the fuel consumption coefficient, but also affected by the dynamic changes of the speed. In general, the accompanying variable λ v The calculation of needs to be iteratively updated through the adjoint equation.

[0067] The optimal solution for the heading angle change rate ω(t) is obtained by taking the derivative of the heading angle part of the Hamiltonian function. The specific expression is: The solution is: in: ω * : optimal heading angle change rate, in radians per second; λ θ : accompanying variable corresponding to the heading angle; I z : Ship's moment of inertia, in kilograms per meter 2 .

[0068] Alternatively, the range of ω(t) can be constrained according to the physical limits of the steering gear. For example, the rate of change of the heading angle is usually limited to [-0.5, 0.5] radians per second to ensure the physical feasibility of the ship's steering.

[0069] In some embodiments, in order to more accurately calculate the accompanying variable λ x ,λ y ,λ v ,λ θ , we need to introduce adjoint equations. These equations are determined by the partial derivatives of the Hamiltonian function with respect to the state variables, specifically: in: The rate of change of the adjoint variable over time.

[0070] These equations form a set of ordinary differential equations that are used to calculate the corresponding adjoint variables for each state variable. In one possible implementation, these equations are solved using a numerical iterative method, such as gradient descent or shooting order, until the solution converges.

[0071] The optimal control strategy depends not only on the calculated propulsion force and heading angle, but also on the ship's dynamic constraints and the weighting of the optimization objectives. For example, in adverse sea conditions, to improve navigation stability, the weight of the stability objective can be appropriately increased, and the influence of wave height and wind speed can be introduced into the Hamiltonian function. Specifically, increasing the stability objective weight ω4 directly affects the calculated optimal propulsion force and heading angle, allowing the ship to optimize fuel efficiency while maintaining safety.

[0072] As an implementation method, to ensure real-time control and accuracy, the optimal control strategy generation module can be integrated with the numerical solution module. In actual operation, control inputs are updated in real time to adapt to dynamically changing environmental data. For example, when a ship approaches an obstacle, the optimization algorithm prioritizes the obstacle avoidance objective and generates a temporary emergency avoidance path by adjusting propulsion force and heading angle.

[0073] S5. Numerical solution: Discretize the optimization target in time, calculate the ship's state evolution through forward simulation, and combine it with backward optimization iteration to determine the discrete control sequence and state trajectory of the optimal path; By discretizing the continuous-time optimization objective, complex mathematical models can be transformed into computable discrete optimization problems. Generally, a ship's state evolution and control inputs are continuous. Numerical solutions simulate state changes by discretizing the time step. Combining iterative calculations using forward simulation and backward optimization, the discrete control sequence and state trajectory for the optimal path are determined. This step is directly related to the accuracy and real-time performance of the ship's path planning and control strategies.

[0074] During the numerical solution process, the nonlinear characteristics of the dynamic model and environmental interference will affect the solution accuracy. Therefore, combining an appropriate discretization method with an efficient iterative solution algorithm is the core means to achieve the technical objectives of the present invention.

[0075] In this embodiment, the basis of time discretization is to divide the voyage time T into N time steps, with the step length of each time step being Δt = T / N. After time discretization, the ship's state variable X(t) and control input u(t) are expressed in discrete form at each time step: X k =[x k ,y k ,v k ,θ k ] T ,u k =[F thrust,k ,ω k ] T in: X k : state variable vector at the kth time step; u k : control input vector for the kth time step; k=0,1,…,N。

[0076] Generally speaking, the size of Δt directly affects the accuracy and efficiency of the discretized calculation. Alternatively, the value of Δt can be dynamically adjusted based on the ship's operating conditions and environmental complexity. For example, in areas with dense obstacles, the step size can be appropriately reduced to improve solution accuracy.

[0077] Specifically, the ship's state evolution is calculated through forward simulation. At each time step, the state variables are discretized according to the dynamic equations. Taking the ship's transverse and longitudinal positions as an example, its discretization expression is: x k+1 =x k +Δt·v k cosθ k y k+1 =y k +Δt·v k sinθ k The above formula shows that the next moment position of the ship depends on the current speed and heading angle, as well as the time step Δt.

[0078] The discretized expression of speed is: The discretization of the heading angle is: θ k+1 =θ k +Δt·ω k in: m is the mass of the ship; α and β are the nonlinear resistance coefficient and the linear resistance coefficient, respectively; F thrust,k and ω k are the propulsion force and heading angle change rate of the current time step respectively.

[0079] The above formula can be used to calculate the state changes of the ship at each time step in sequence, thereby obtaining a complete state evolution trajectory.

[0080] In the backward optimization phase of numerical solution, the update of the adjoint variable is realized through the discretized adjoint equation. In general, the adjoint variable is used to describe the sensitivity of the optimization objective function to the state variable, and its discretized update formula is: in: λ x,k ,λ y,k ,λ v,k ,λ θ,k are the adjoint variables of the kth time step respectively; H k is the Hamiltonian function at the kth time step.

[0081] The above formula calculates the value of the adjoint variable at each time step in turn through reverse iteration.

[0082] In one possible implementation, forward simulation and backward optimization are performed alternately to form a bidirectional iterative solution process. The specific process is as follows: First, starting from the initial state, the complete state evolution trajectory of the ship is calculated through forward simulation; Then, starting from the terminal state, backward optimization is performed to update the adjoint variables at each time step; Finally, according to the update results of the accompanying variables, adjust the control input u k =[F thrust,k ,ω k ] and recalculate the state evolution of the forward simulation.

[0083] The above iterative process continues until the changes in control input and state variables converge to a preset accuracy range.

[0084] As an option, a numerical optimization algorithm can be introduced to accelerate iterative convergence. For example, the control input can be updated using the gradient descent method, and its formula is: in: and are the control inputs for the nth and n+1th iterations respectively; η is the learning rate, which controls the step size of each update.

[0085] In some embodiments, the optimization efficiency can be further improved by dynamically adjusting the learning rate η. For example, when the gradient descent process is close to convergence, the learning rate can be appropriately reduced to avoid iterative oscillation.

[0086] In this embodiment, the numerical solution module outputs the calculated discrete control sequence and state trajectory of the optimal path and directly transmits them to the real-time monitoring module for path verification and control input adjustment. In practical applications, the numerical solution module can dynamically restart the optimization iteration process based on environmental changes, thereby generating an optimal path that adapts to the new external conditions.

[0087] S6. Real-time adjustment and monitoring: Based on real-time collected environmental data and ship state variables, dynamically adjust the optimal control input, simulate and verify the ship's path planning results, update the control strategy, and optimize the path in real time during actual operation.

[0088] Real-time adjustment and monitoring are key components of dynamic optimization. Dynamically adjusting optimal control inputs based on real-time collected environmental data and vessel state variables effectively addresses complex and changing marine environments. Typically, ships are affected by external factors such as ocean currents, wind and waves, and unexpected obstacles during operation. These factors can cause the original path planning results to not fully meet actual requirements. Therefore, through real-time adjustment and monitoring, control strategies can be dynamically revised, and paths can be continuously optimized during actual operation, improving system stability and robustness.

[0089] Combining the control sequence and state trajectory output by the numerical solution module, the feasibility of the path planning results is evaluated using simulation verification methods, and the control input is updated through a feedback mechanism, thereby achieving real-time coordination between path planning and actual operation.

[0090] The core of real-time adjustment is to update the optimization objectives and constraints based on the latest environmental data and state variables. The real-time environmental data of the ship includes the ocean current velocity component (u c ,v c ), wind speed w s 、wave height h w , and the dynamic distribution information of obstacles. Specifically, the real-time state variables are provided by the data acquisition module, including the current position of the ship (x, y), speed v, heading angle θ, and propulsion force F thrust and the current value of the heading angle change rate ω.

[0091] As an implementation method, the value function V(X,t) is dynamically updated by processing the current environmental data to adjust the control input. The specific updated mathematical expression is: in: X = [x, y, v, θ] is the current state variable; u=[F thrust ,ω] is the current control input; is the Hamiltonian function; Δt is the time step.

[0092] By adjusting the value function in real time, the control input can be re-optimized under new environmental conditions so that the path planning results of the ship can adapt to the actual operating environment.

[0093] Specifically, simulation verification of path planning results is achieved through digital twin technology. Generally speaking, digital twin technology builds a virtual model that is highly consistent with the actual operation of the ship. This model can simulate the ship's motion trajectory and response to control inputs in a virtual environment. The core of simulation verification is to evaluate the effectiveness of the current control strategy. The verification content includes but is not limited to: Whether fuel consumption meets the optimization target; Whether the speed and heading meet the path planning requirements; Is the navigation stability within an acceptable range?

[0094] In one possible implementation, the simulation results are used to generate feedback data through deviation analysis. For example, if the fuel consumption in the simulation results deviates from the expected value, the propulsion force F can be adjusted to thrust The control strategy is modified by the optimization weights of

[0095] As an option, when environmental conditions suddenly change, the real-time adjustment module will prioritize adjusting the control input through the emergency obstacle avoidance algorithm. In the obstacle avoidance scenario, the weight of the path planning will dynamically bias towards the obstacle avoidance target J obstacle , the specific adjustment method is: w obstacle →w obstacle +δ in: w obstacle is the weight of the obstacle avoidance target; δ>0 is the dynamic increment of weight.

[0096] Through this dynamic adjustment mechanism, the ship is able to prioritize avoiding obstacles while maintaining the optimization goal of the original path planning as much as possible.

[0097] In this embodiment, the updated control strategy is directly applied to the vessel's operation, while the numerical solution module also recalculates the control sequence for the next phase. For example, when an obstacle disappears or environmental conditions return to stability, the real-time adjustment module restores the original optimized target weight configuration to maximize fuel efficiency and shorten routing time.

[0098] The mathematical expression of real-time optimization can be further expanded into the following form: in: is the optimal control input at the current moment; λ is the adjoint variable, which is calculated through backward optimization.

[0099] The above optimization problem is solved through real-time updated value functions and dynamic models, and the final adjustment is completed by combining feedback data from simulation verification.

[0100] In some embodiments, to improve the efficiency of real-time adjustments, machine learning algorithms can be introduced to predict environmental changes. For example, by using a learning model based on historical environmental data, the likelihood of obstacles appearing or the changing trend of ocean currents can be predicted in advance, allowing control strategies to be pre-adjusted before environmental changes occur. This approach can significantly improve the system's responsiveness and adaptability.

[0101] In this embodiment, the real-time adjustment and monitoring steps can also provide operators with optimization suggestions for the current path plan via the onboard control panel. For example, if simulation results indicate that a certain control input may result in reduced fuel efficiency, the system will prompt the operator to adjust the speed or heading to optimize the actual operating performance of the ship.

[0102] The ship intelligent monitoring system described below and the ship intelligent monitoring method described above can refer to each other.

[0103] Please see the attached Figure 2 The present invention also provides a ship intelligent monitoring system, comprising: Data acquisition module, used to collect the ship's initial state variables and real-time environmental data; Dynamic modeling module, used to establish a nonlinear dynamic model of the ship based on the ship's state and environmental data; The optimization module is used to construct a cost function and generate the optimal control strategy of the ship through the Hamilton-Jacobi-Bellman equation and Pontryagin extreme value principle with the goal of minimizing fuel consumption; Numerical solution module, used to generate optimal path planning results through discretization, forward simulation and backward optimization; Real-time monitoring module, used to dynamically adjust speed, heading angle and propulsion force, and verify the optimized path in real time; The simulation verification module is used to verify the feasibility and economy of the optimization path through digital twin technology.

[0104] The system of this embodiment can be used to execute the above method embodiments, and its principles and technical effects are similar, so they will not be repeated here.

[0105] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent monitoring of ships, characterized in that: The following steps are involved: Data collection: Collect the ship's initial state variables and real-time environmental data, including the ship's speed, position, heading, surrounding ocean environment parameters, and navigation obstacle information; Dynamic modeling: Based on the collected ship state variables and control inputs, a nonlinear dynamic model of the ship is established. The dynamic model describes the motion of the ship under different propulsion forces and heading angles. Optimization target construction: Based on the dynamic model, an optimization objective function is constructed, with time, energy consumption, obstacle avoidance capability, and navigation stability as optimization targets. The optimal path value function is determined using the Hamilton-Jacobi-Bellman equation; Optimal control strategy generation: Utilizing the Pontryagin extreme value principle, based on the optimization results in the value function, the optimal propulsion force and heading angle control inputs are generated to meet the optimization objectives of the ship path planning. Numerical solution: The optimization objective is discretized in time, the ship's state evolution is calculated through forward simulation, and the discrete control sequence and state trajectory of the optimal path are determined by combining backward optimization iterations; Real-time adjustment and monitoring: Based on real-time collected environmental data and ship state variables, the optimal control input is dynamically adjusted, the ship's path planning results are simulated and verified, the control strategy is updated, and the path is optimized in real time during actual operation.

2. A ship intelligent monitoring method according to claim 1, characterized in that: The initial state variables include: Collect the initial position data of the ship, including the lateral position and longitudinal position; Collect the ship's initial speed and heading angle data; Collect ocean current velocity components; All collected initial data are input into the dynamic modeling module to provide basic data support for subsequent optimization goals.

3. A ship intelligent monitoring method according to claim 1, characterized in that: In the dynamic modeling step, the established ship nonlinear dynamic model includes the following contents: The change in the transverse and longitudinal position of the ship, determined by its speed and heading angle; The speed of a ship is affected by the combined effects of resistance, propulsion force and ocean current speed; The change of the ship's heading angle is determined by the rudder angle control and the external navigation environment factors.

4. A ship intelligent monitoring method according to claim 1, characterized in that: The optimization target construction includes: The fuel consumption target is related to a nonlinear function of propulsion power and speed; The path planning time goal is described by the shortest time optimization model; The obstacle avoidance capability goal is optimized through real-time calculation of the distribution of environmental obstacles; The navigation stability goal is to optimize the smoothness of the hull motion under the influence of wind and waves.

5. A ship intelligent monitoring method according to claim 1, characterized in that: The construction of the value function includes: The value function is associated with the ship’s current position, speed, and heading angle; The cost function minimizes fuel consumption at each state point while satisfying the time and obstacle avoidance constraints of path planning; The optimal path value function is generated by numerically solving the HJB equation.

6. A ship intelligent monitoring method according to claim 1, characterized in that: The optimal control strategy includes: The control input for optimal propulsion is determined by optimizing the relationship between fuel consumption and ship speed; The control input of the optimal heading angle is used to adjust the angle through the servo to achieve dynamic tracking of the target path; The control inputs of propulsion force and heading angle are determined by the extreme value conditions of the Hamiltonian function and satisfy the physical constraints of speed, propulsion force and rudder angle.

7. A ship intelligent monitoring method according to claim 1, characterized in that: The numerical solution includes: Discretize the optimization objective function in time and transform the continuous optimization problem into an optimization problem with discrete time steps; Forward simulate the evolution of ship state variables and calculate discrete state trajectories; Backward optimization iteratively updates the control variables and updates the optimal control inputs of propulsion force and heading angle through the adjoint equations; The optimal path planning result is determined through numerical iteration of discrete time series.

8. A ship intelligent monitoring method according to claim 1, characterized in that: The real-time adjustment and monitoring include: Update the value function through real-time collected environmental data; Dynamically adjust the constraints in the optimization objective function to ensure the real-time performance of path planning; The digital twin system is used to conduct real-time simulation verification of the ship's path planning scheme and update the control strategy during operation.

9. A ship intelligent monitoring method according to claim 1, characterized in that: In the real-time adjustment and monitoring, the simulation verification based on the digital twin system includes: Simulation verifies whether the ship's fuel consumption meets the optimization target; Simulation evaluation to determine whether the speed, heading and hull stability meet safety requirements; The optimization results are iteratively adjusted multiple times through simulation data to generate the final control strategy.

10. A ship intelligent monitoring system, according to a ship intelligent monitoring method according to any one of claims 1 to 9, characterized in that: include: Data acquisition module, used to collect the ship's initial state variables and real-time environmental data; Dynamic modeling module, used to establish a nonlinear dynamic model of the ship based on the ship's state and environmental data; The optimization module is used to construct a cost function and generate the optimal control strategy of the ship through the Hamilton-Jacobi-Bellman equation and Pontryagin extreme value principle with the goal of minimizing fuel consumption; Numerical solution module, used to generate optimal path planning results through discretization, forward simulation and backward optimization; Real-time monitoring module, used to dynamically adjust speed, heading angle and propulsion force, and verify the optimized path in real time; The simulation verification module is used to verify the feasibility and economy of the optimization path through digital twin technology.

Images