Ship route multi-objective optimization model for composite sea state scenarios

Abstract

The application discloses a ship route multi-objective optimization model for composite sea state scenarios and relates to the technical field of ship route management, and comprises the following steps: collecting ship sailing state data and marine environment data in real time by using marine monitoring equipment, a sensor network and ship self-equipment, and combining historical data to establish a ship motion model under different sea states; the application captures the dynamic changes of rolling, pitching and the like of a floating obstacle under the action of wind and waves in real time by constructing a floating obstacle dynamic posture model, and constructs a time-varying risk field in combination with a trajectory uncertainty ellipse. Compared with a traditional static obstacle processing mode, the model can accurately quantify the real-time changes of the dynamic windward area, the resistance coefficient and the influence range of the obstacle, avoids sailing route planning errors caused by neglecting the dynamic characteristics of the obstacle, and makes the ship avoid the risk area expanded due to the posture change of the obstacle in advance through real-time updating of the motion state of the obstacle by a dynamic posture vector, so that the collision probability is significantly reduced.

Description

Technical Field

[0001] This invention relates to the field of ship route management technology, and in particular to a multi-objective optimization model for ship routes under complex sea conditions. Background Technology

[0002] As one of the core technologies in the maritime field, the accuracy and safety of ship route planning directly affect shipping efficiency, fuel consumption, and navigation safety. In the complex marine environment, ships face dynamic changes in sea conditions such as wind, waves, currents, tides, and obstacles such as ice floes, abandoned vessels, and floating structures. Traditional route planning models are insufficient to fully adapt to the real-time risk assessment needs under complex sea conditions.

[0003] In existing technologies, ship route planning typically employs static obstacle handling, simplifying floating obstacles into fixed shapes or uniform motion models, neglecting dynamic attitude changes such as roll and pitch under the influence of wind and waves. This simplification makes it impossible to accurately quantify the real-time fluctuations in the windward area, drag coefficient, and impact range of obstacles, leading to misjudgments of risks during route planning. Traditional models do not consider the trajectory uncertainties caused by changes in the attitude of floating obstacles, making it difficult to construct dynamically expanding risk areas. Ships may face collision risks due to untimely avoidance. Therefore, how to construct adaptive route planning that considers the dynamic characteristics of floating obstacles and achieves multi-objective optimization has become an urgent problem to be solved in the field of ship route management technology. Summary of the Invention

[0004] The purpose of this invention is to address the problem that traditional models do not consider trajectory uncertainties caused by attitude changes of floating obstacles, making it difficult to construct dynamically expanding risk areas, and ships may face collision risks due to untimely avoidance. Therefore, a new multi-objective optimization model method and system for ship routes in complex sea state scenarios is proposed.

[0005] To achieve the above objectives, this invention employs the following technology for a multi-objective optimization model of ship routes under complex sea state scenarios, comprising the following steps:

[0006] By utilizing marine monitoring equipment, sensor networks, and ship-on-ship equipment, real-time data on ship navigation status and marine environment are collected, and historical data are combined to establish motion models of ships under different sea conditions.

[0007] The initial route is generated by taking the starting point and destination as input and combining basic information from the ocean map.

[0008] Determine if there are obstacles on the initial flight path. If there are:

[0009] To determine whether an obstacle is a floating obstacle, if it is:

[0010] Obtain the shape information of the obstacle, construct a dynamic attitude model of the floating obstacle, calculate the dynamic windward area, drag coefficient and radius of influence of the floating obstacle, and construct a time-varying risk field by combining the trajectory uncertainty ellipse;

[0011] The objective function is constructed by combining the time-varying risk field with changes in flight distance and heading, and the optimal route is calculated.

[0012] Furthermore, the dynamic attitude model of the floating obstacle is determined according to the following relationship:

[0013]

[0014] In the formula: Let θ be the attitude vector of the floating obstacle; θ(t) is the roll angle of the floating obstacle at time t. Let ω be the pitch angle of the floating obstacle at time t; θ (t) represents the roll angular velocity of the floating obstacle at time t; Let t be the pitch angular velocity of the floating obstacle.

[0015] Furthermore, the risk field is constructed through the following specific steps:

[0016] Based on the established dynamic attitude model, the dynamic windward area and dynamic drag coefficient of the obstacle are calculated;

[0017] The radius of influence of floating obstacles is determined based on dynamic windward area and drag coefficient;

[0018] Due to trajectory uncertainty caused by changes in obstacle attitude, an trajectory uncertainty ellipse is constructed.

[0019] By combining the calculated dynamic influence radius R(t) of the floating obstacle and the trajectory uncertainty ellipse parameters, a time-varying risk field is constructed.

[0020] Furthermore, the operational status of the floating obstacle is obtained, and it is detected whether there are fixed obstacles within the risk field of the floating obstacle. If so:

[0021] There is a risk of collision between floating obstacles and fixed obstacles. Potential collision points are constructed at the points where the risk fields of the fixed obstacle and the floating obstacle coincide. The collision direction of the floating obstacle at each collision point is obtained, and the comprehensive collision risk zone of the floating obstacle is calculated.

[0022] Calculate the new risk field for floating obstacles.

[0023] Furthermore, the specific steps for calculating the comprehensive collision risk zone of the floating obstacle are as follows:

[0024] Collect marine monitoring satellite imagery, shore-based radar monitoring data, ship automatic identification system data, and underwater sonar detection data;

[0025] Fixed obstacle information detected from different data sources is spatially matched and fused based on a geographic coordinate system to eliminate duplicate detections and errors, thereby constructing an accurate database of fixed obstacle locations.

[0026] Obtain the actual collision point from the potential collision point and calculate the collision direction of the floating obstacle at the moment of collision;

[0027] Using historical collision data and current floating and fixed obstacle states and environmental conditions, a machine learning model is trained; the relevant features of the collision point are used as input to the model, and the collision probability of that collision point is output; the number of times floating and fixed obstacles collide under the same conditions is counted, and the frequency of collision is calculated.

[0028] Obtain structural design drawings and material specifications for both fixed and floating obstacles, and establish a three-dimensional structural model; simulate the stress and strain distribution of the obstacles during the collision process to determine the weak points and failure modes of the structure; calculate the total energy released during the collision process based on the mass, velocity, and relative motion state of the floating obstacle at the moment of collision.

[0029] Based on the total energy combined with the discrete element method, the debris motion equation is established to calculate the trajectory of the debris in the marine environment and determine the impact range of debris splash. Based on the degree of damage to the obstacle after the collision, the remaining buoyancy, and the change in the center of gravity, combined with the ocean circulation model and real-time ocean current and tidal data, the drift path of the obstacle in the ocean is predicted; thus, the physical collision risk zone is obtained.

[0030] Using the obstacle state and marine environmental parameters at the moment of collision as boundary conditions, the changes in local water flow velocity and pressure fields caused by the collision are calculated. A mathematical model of ship motion is established, and the changes in local water flow velocity and pressure fields are input as external disturbance forces into the mathematical model of ship motion to calculate the ship's speed and heading deviation in the affected water flow. The degree of influence of water flow changes on ship handling is evaluated in combination with ship type, load conditions, and handling performance parameters. The degree of influence is divided into different levels, and different influence areas are defined accordingly. Thus, the hydrodynamic influence zone is obtained.

[0031] The calculation results of the physical collision risk zone and the fluid dynamics influence zone are spatiotemporally aligned, and a weighted fusion algorithm is used to merge the physical collision risk zone and the fluid dynamics influence zone.

[0032] Based on the degree of impact of different areas on ship navigation safety, different weights are assigned to the physical collision risk zone and the hydrodynamic influence zone to generate a comprehensive collision risk field.

[0033] Furthermore, the specific steps for calculating the new risk field of the floating obstacle are as follows:

[0034] The comprehensive collision risk zone of floating obstacles is superimposed and fused with the time-varying risk field;

[0035] The risk levels of the comprehensive collision risk zone and the time-varying risk field are weighted.

[0036] Based on the weighted results, a new risk field is constructed to address potential threats to ship navigation.

[0037] Furthermore, the specific steps for detecting whether there are fixed obstacles within the risk field of the floating obstacles are as follows:

[0038] For satellite imagery and underwater sonar images, edge detection algorithms are used to extract the contour edges of obstacles, and then shape features are obtained through shape descriptors to determine whether they are fixed obstacles.

[0039] For radar and AIS data, a Kalman filter algorithm is used to smooth the position and velocity data of obstacles and analyze their motion trajectory.

[0040] If the trajectory does not change significantly over a long period of time and its relative position to the seabed topography remains fixed, it is determined to be a fixed obstacle.

[0041] Furthermore, the specific steps for obtaining the actual collision point based on the potential collision point are as follows:

[0042] The time-varying risk field of floating obstacles is divided into small grids in three-dimensional space, and each small grid is assigned a unique number and spatial coordinates;

[0043] Based on the dynamic attitude model of the floating obstacle, its dynamic windward area, the radius of influence, and the trajectory uncertainty ellipse, the risk value of each small grid is calculated. The higher the risk value, the greater the probability of collision in that area.

[0044] The edges of fixed obstacles are discretized into multiple points, and each point is checked to see if it falls within a high-risk small grid of the risk field of floating obstacles.

[0045] If a point falls into a high-risk small grid, then that point is identified as a potential collision point;

[0046] For the potential collision points initially identified, the movement trajectories of the floating obstacles over a period of time are combined to determine whether there is an actual collision hazard at that point; if the movement trajectories of the floating obstacles intersect at that point, then there is an actual collision hazard at that point, and that point is confirmed as a real collision point.

[0047] Furthermore, the specific steps for calculating the collision direction of the floating obstacle at the moment of collision are as follows:

[0048] Simulate the collision process between a floating obstacle and a fixed obstacle at the point of impact, and calculate the direction of the resultant force on the floating obstacle at the instant of collision;

[0049] Considering the influence of environmental factors such as ocean currents and waves on the collision direction, an environmental factor correction model is established to adjust the calculated resultant force direction. The corrected resultant force direction is the collision direction.

[0050] Furthermore, the specific steps for determining the impact range of the debris splash are as follows:

[0051] Based on the discrete element method, floating obstacles and fixed obstacles are discretized into multiple elements to simulate the breaking and splashing of elements during the collision process;

[0052] The connection strength and fracture criteria between units are set, and the units that will break are determined based on the total energy released during the collision and the structural stress distribution.

[0053] By combining the driving force of ocean currents and waves on debris, a debris motion equation is established to calculate the trajectory of debris in the marine environment;

[0054] The trajectory of debris in the marine environment was simulated multiple times using debris motion equations. The probability of debris appearing in different areas was statistically analyzed, and a debris splash probability distribution cloud map was drawn to determine the impact range of debris splash.

[0055] In summary, due to the adoption of the above-mentioned technology in the multi-objective optimization model of ship routes under complex sea state scenarios, the beneficial effects of this invention are:

[0056] 1. This invention constructs a dynamic attitude model of floating obstacles to capture their dynamic changes such as roll and pitch under the influence of wind and waves in real time, and combines this with a trajectory uncertainty ellipse to construct a time-varying risk field. Compared with traditional static obstacle handling methods, this model can accurately quantify the real-time changes in the dynamic windward area, drag coefficient, and influence range of obstacles, avoiding route planning errors caused by ignoring the dynamic characteristics of obstacles. By updating the motion state of obstacles in real time through dynamic attitude vectors, ships can avoid the risk area expanded by the change in obstacle attitude in advance, significantly reducing the probability of collision.

[0057] 2. Based on the objective function constructed from a time-varying risk field, the voyage length, cumulative navigation risk value, and course change are weighted and integrated. The optimal route is then solved using a particle swarm optimization algorithm or a genetic algorithm. This multi-objective optimization mechanism can dynamically adjust the weights according to navigation needs, shortening the voyage, reducing course fluctuations, lowering fuel consumption and maneuverability while avoiding obstacles. In high-risk sea areas, increasing the risk weight coefficient can generate safer detour routes; in conventional sea areas, the voyage weight is emphasized to achieve efficient transportation.

[0058] 3. When there is a risk of collision between floating and fixed obstacles, this invention constructs a core risk zone, a debris risk zone, and an eddy current risk zone, and integrates them with the original time-varying risk field to form a new comprehensive risk assessment system. This mechanism not only considers the dynamic changes of the obstacle itself, but also covers the derivative risks such as debris diffusion and eddies that may occur after a collision, making route planning more aligned with the actual threats under complex sea conditions. For example, by calculating the water depth at the apex of the fixed obstacle to determine the radius of the eddy current risk zone, the impact of the water flow on the ship after a collision can be accurately predicted, avoiding secondary dangers.

[0059] 4. Utilizing marine monitoring equipment and sensor networks, the system collects navigation data and sea state information in real time, and dynamically updates the ship motion model and obstacle risk field by combining historical data. When deviations occur between actual navigation data and model predictions, the system can quickly restart the route optimization module to achieve dynamic route adjustments. This closed-loop feedback mechanism enables the model to continuously adapt to complex sea state changes, gradually improving the accuracy and reliability of route planning by iteratively optimizing the ship's motion parameters under different currents and wave heights. Attached Figure Description

[0060] Figure 1 A flowchart of the present invention is shown. Detailed Implementation

[0061] The following will describe, with reference to the accompanying drawings of the embodiments of the present invention, the multi-objective optimization model for ship routes under complex sea state scenarios. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0062] To more clearly and intuitively demonstrate the practical application effects and advantages of the multi-objective optimization model method for ship routes under complex sea state scenarios, and to verify its feasibility and effectiveness, the invention is further described below with reference to embodiments. Through specific scenario simulations and data calculations, the model is explained in detail how it functions in actual ship routes, helping readers better understand the technical details and practical value of the invention. The invention is further described below with reference to embodiments;

[0063] Example 1: Ship Navigation Data Acquisition and Analysis Module: First, a complete data acquisition system is established, utilizing marine monitoring equipment, sensor networks, and the ship's own equipment to acquire real-time ship navigation status data and marine environmental data. The ship navigation status data includes the ship's position (x, y), speed v, heading θ, and acceleration a. The marine environmental data includes wind speed v. w Wind direction θ w Ocean current speed v c and direction θ c Water depth h, wave height, and wave period are considered. In-depth analysis of historical navigation data and real-time collected data is conducted to establish motion models of the ship under different sea states. Without considering the effects of wind speed and ocean currents, the ship's equation of motion can be expressed as:

[0064]

[0065] In the formula, and These represent the rate of change of the ship's position (velocity component) in the x-axis and y-axis directions, respectively, used to describe the dynamic changes in the ship's position in a planar coordinate system; v represents the ship's speed, which is the distance the ship moves per unit time; θ is the ship's heading angle, that is, the angle between the ship's direction of travel and a certain reference direction (such as due north), and cosθ and sinθ are used to decompose the ship's velocity into the x-axis and y-axis directions; The rate of change of a ship's speed, or acceleration 'a', reflects how quickly a ship's speed increases or decreases. This represents the rate of change of the ship's heading angle, also known as the turning angular velocity w, reflecting how quickly the ship changes course; w is the ship's turning angular velocity. When considering the effects of wind speed and ocean currents, the ship's actual speed v... actual It can be represented as:

[0066] The above model provides basic data support for subsequent route optimization.

[0067] Based on the ship's starting point S(x) s ,y s ) and destination D(x d ,y d The information is obtained using an improved A algorithm. The evaluation function of the A algorithm is f(n) = g(n) + h(n), where g(n) is the actual cost from the starting point S to node n, and h(n) is the estimated cost from node n to the destination D. In the ocean map, the coordinates of node n are (x... n ,y n ),but:

[0068]

[0069] The algorithm starts from point S and continuously searches for reachable nodes in the surrounding area. It selects the optimal node based on the evaluation function f(n) and expands the search until the destination node is found, thus generating an initial route. Considering basic information from the ocean map (such as coastline and typical restricted areas like shoals), a preliminary route from the starting point to the destination is generated. This initial route serves as the basic framework for subsequent optimizations.

[0070] Determine if there are obstacles in the ship's shipping route:

[0071] Leveraging the real-time scanning capabilities of the radar system, it dynamically captures surface targets in the shipping lanes. Its electromagnetic wave detection characteristics can penetrate fog and darkness, forming a continuous spatial target map. Combined with an Automatic Identification System (AIS), it obtains information such as the position and course of surrounding vessels through ship signal interaction, constructing a dynamic obstacle avoidance database. Simultaneously, the underwater multibeam sonar system uses fan-shaped beams to cover the seabed topography, transforming underwater obstacles such as reefs and shipwrecks into a visualized three-dimensional terrain model, thereby determining the presence of obstacles in the ship's shipping route.

[0072] When there is an obstacle in the shipping route of a ship, determine whether the obstacle is a floating obstacle:

[0073] If a side-scan sonar detects that the bottom of an obstacle is not fixed to the seabed (e.g., there is a gap or the bottom shape changes with depth), then fixed obstacles (such as reefs or shipwrecks, which are fixed at the bottom if they are aground) are excluded, and the obstacle is a floating obstacle.

[0074] At this time, marine environmental data, including but not limited to wind speed, wind direction, wave height, and wave period, are collected in real time using marine monitoring equipment, sensor networks, and the ship's own equipment. Data on floating obstacles at sea is collected through radar detection, visual recognition, and sonar scanning technologies, obtaining data on the shape (e.g., cuboid, cylinder, irregular shape), mass, initial position, and attitude (including roll angle, pitch angle, roll rate, and pitch rate) of the floating obstacles. Simultaneously, the ship's initial position, target position, loading status (e.g., cargo weight, center of gravity position), and fuel consumption rate data are collected to provide basic data support for subsequent calculations and optimizations. Among these, marine environmental data reflects the external environmental conditions in which the ship navigates, and these conditions affect the movement of obstacles and the ship's navigation; obstacle data is used to understand the characteristics and state of the obstacles themselves; and ship data is used to plan routes in conjunction with the ship's own capabilities.

[0075] Dynamic attitude modeling steps for floating obstacles: Based on the collected shape information of the floating obstacles, construct the corresponding dynamic attitude model. Define the attitude vector:

[0076]

[0077] In the formula: Let θ be the attitude vector of the floating obstacle; θ(t) is the roll angle of the floating obstacle at time t, i.e., the rotation angle around the horizontal axis. Let ω be the pitch angle of the floating obstacle at time t, i.e., the rotation angle about the longitudinal axis; θ (t) represents the roll angular velocity of the floating obstacle at time t; Let t be the pitch angular velocity of the floating obstacle.

[0078] The attitude parameters (θ(t)) of obstacles are acquired in real time using multi-sensor fusion technology. ω θ (t), Specifically, it combines angular velocity information acquired by an inertial measurement unit (IMU) with attitude information acquired by visual simultaneous localization and mapping (SLAM) technology. An inertial measurement unit is a sensor array capable of measuring the motion state and spatial attitude of an object. It can accurately measure the angular velocity changes of obstacles in various directions and is widely used in aerospace, autonomous driving, robotics, virtual reality, and other fields. SLAM technology is an interdisciplinary technology integrating sensor perception, computer vision, robotics, and numerical optimization. It enables devices to infer their own trajectory and construct geometric and semantic maps of the environment in environments without prior maps, using continuous sensor data (such as visual images and laser point clouds). SLAM technology can acquire the attitude information of obstacles from a visual perspective. The combination of these two technologies updates the various parameters (referring to θ(t)) in the attitude vector in real time at a high sampling frequency. ω θ (t), This allows for a precise description of the dynamic attitude of obstacles. For example, when an obstacle rolls and pitches under the influence of wind and waves, the IMU senses the change in angular velocity, and SLAM technology captures the corresponding attitude changes. Through data fusion processing, θ(t) is updated in a timely manner. ω θ (t), The value of allows the dynamic attitude model to accurately reflect the attitude changes of the obstacle at different times, and thus describe the rolling and pitching attitude changes of the obstacle under the action of wind and waves.

[0079] Based on the collected shape information of the floating obstacle, a corresponding dynamic attitude model is established to describe the obstacle's roll and pitch attitude changes under the influence of wind and waves. Through multi-sensor fusion technology, the obstacle's attitude parameters are acquired in real time, enabling a precise description of the obstacle's dynamic attitude. The dynamic attitude model is like creating a "digital model" of the obstacle's attitude changes; by observing the changes in attitude parameters, the attitude of the obstacle at each moment can be clearly understood, such as the angle of roll and the degree of pitch.

[0080] Dynamic influence range calculation steps: Based on the constructed dynamic attitude model, calculate the dynamic windward area A(t) of the floating obstacle. The formula for calculating the dynamic windward area for floating obstacles of different shapes is as follows:

[0081]

[0082] In the formula: A(t) is the dynamic windward area of ​​the floating obstacle; A0 is the static reference windward area, calculated based on the projection of the obstacle's geometric center; α is the attitude influence coefficient, determined by the shape of the obstacle. For example, for a cuboid-shaped floating obstacle, α takes a value between 0.3 and 0.4, while for a spherical floating obstacle, α takes a smaller value, approximately 0.1 to 0.15; φ θ , The phase parameter is determined by the initial attitude of the obstacle.

[0083] Based on the dynamic windward area and the motion characteristics of the obstacle in seawater, the dynamic drag coefficient C is calculated. d (t), its calculation formula is:

[0084]

[0085] In the formula: C d0 β is the static reference drag coefficient, which is related to the material and surface roughness of the obstacle; β is the dynamic influence coefficient, which is determined by the surface roughness of the obstacle. For example, the value of β for smooth metal obstacles is between 0.2 and 0.25, while the value of β for rough wooden obstacles is between 0.3 and 0.35. Let t be the attitude vector of the floating obstacle; t is time. The modulus of the attitude change rate reflects the speed at which the change in the obstacle's attitude affects the change in its frontal area.

[0086] Static reference drag coefficient C d0The method of obtaining the drag coefficient is as follows: Starting from the fundamental theories of fluid mechanics, for objects with regular geometric shapes (such as spheres and cylinders), a drag model can be constructed analytically. Focusing on the force analysis of the object in the fluid, and combining boundary layer theory and the equations of motion for viscous fluids, expressions for the drag coefficient applicable to specific flow field conditions (such as laminar and turbulent flow) are derived. For example, when a fluid flows around a static object, the relationship between fluid viscosity, the object's frontal area, and drag is described by theoretical formulas, thereby determining a reference value for the drag coefficient under standard operating conditions.

[0087] Then, based on the dynamic windward area and drag coefficient, determine the radius R(t) of the floating obstacle's influence range. The calculation formula is as follows:

[0088]

[0089] In the formula: R0 is the static safety radius, which is determined based on the initial position of the obstacle and the safe avoidance distance of the ship; γ is the expansion coefficient, which is determined by the size of the obstacle. For large obstacles (such as abandoned ships), the value of γ is between 0.4 and 0.6, and for small obstacles (such as small buoys), the value of γ is between 0.1 and 0.2.

[0090] Simultaneously, considering the trajectory uncertainty caused by the change in obstacle attitude, an trajectory uncertainty ellipse is constructed, whose semi-major axis a(t) and semi-minor axis b(t) are calculated by integration, i.e. Where σ a ,σ b The uncertainty coefficient is determined by the hydrodynamic characteristics of the obstacle; t is the current time; τ is the integral variable (dummy variable), representing any moment in the process from 0 to t; θ is the pitch angle of the floating obstacle, and θ is the roll angle of the floating obstacle;

[0091] Based on the established dynamic attitude model, the dynamic windward area and dynamic drag coefficient of the floating obstacle are calculated. These two parameters adjust dynamically with changes in the floating obstacle's attitude. The radius of influence of the floating obstacle is determined based on the dynamic windward area and drag coefficient, and this radius dynamically expands with the rate of change of the obstacle's attitude. Simultaneously, considering the trajectory uncertainty caused by changes in the floating obstacle's attitude, a trajectory uncertainty ellipse is constructed to quantify the uncertainty of the floating obstacle's trajectory. The dynamic windward area changes in the size and direction of the area in contact with the wind due to changes in the obstacle's attitude; the dynamic drag coefficient correspondingly affects the magnitude of the resistance to the obstacle's movement in seawater; the radius of influence expands or shrinks with the rate of change of the floating obstacle's attitude; and the trajectory uncertainty ellipse uses specific parameters to represent the uncertain regions that may occur in the obstacle's trajectory, allowing ships to better predict the obstacle's movement range.

[0092] Risk field construction steps: Combining the calculated dynamic influence radius R(t) of the floating obstacle and the trajectory uncertainty ellipse parameters a(t) and b(t) of the floating obstacle, a time-varying risk field R(x,y,t) is constructed. The formula for calculating the time-varying risk field is:

[0093]

[0094] In the formula: R(x,y,t) is the time-varying risk field; r(x,y,t) is the distance from (x,y) to the center of the floating obstacle at time t; The uncertainty radius reflects the range of uncertainty in the obstacle's trajectory; k1 and k2 are risk weight coefficients, which can be adjusted according to actual navigation requirements and ship performance. For example, in high-risk sea areas, the values ​​of k1 and k2 can be increased to improve the priority of obstacle avoidance.

[0095] By combining the calculated dynamic impact range and trajectory uncertainty ellipse of floating obstacles, a time-varying risk field is constructed to reflect the degree of risk impact of obstacles on the surrounding sea area at different times. The time-varying risk field integrates and visualizes the risk situation caused by obstacles to the surrounding sea area at different times and in different attitudes. Ships can use this risk field to intuitively understand which areas have a high collision risk at each time.

[0096] Route optimization steps: Construct an objective function J that incorporates distance, risk, and course changes to comprehensively consider the multi-objective requirements of ship navigation. The formula for calculating the objective function J is:

[0097]

[0098] In the formula: L is the route length, which is obtained by calculating the path length between the ship's initial position and the target position; This represents the total sailing time [t0, t] of the ship. f The cumulative risk value of the areas traversed reflects the safety during the voyage; The objective function J represents the cumulative change in the ship's heading angle ψ(t) during navigation, reflecting the difficulty of ship maneuvering and fuel consumption. w1, w2, and w3 are weighting coefficients, where w1 + w2 + w3 = 1. These coefficients are set according to different navigation tasks and requirements. For example, in tasks prioritizing rapid transportation, the value of w1 can be appropriately increased; in tasks emphasizing safety, the value of w2 can be increased. Intelligent optimization algorithms, such as particle swarm optimization and genetic algorithms, are used to solve the objective function J, obtaining the ship's optimal route while satisfying navigation constraints (such as avoiding obstacles and maintaining a safe speed).

[0099] The particle swarm optimization algorithm solves for the objective function J, and the specific steps are as follows:

[0100] Step 1: In the solution space of the route planning, a certain number of particles are randomly generated. Each particle represents a potential ship route, and the particle's position vector corresponds to the coordinates of the nodes passed by the route. At the same time, a velocity vector is initialized for each particle. The velocity determines the direction and step size of the particle's search in the solution space.

[0101] Step 2: Substitute each particle (i.e., potential route) into the objective function J to calculate the corresponding fitness value. The fitness value reflects the merits of the route under the comprehensive consideration of multiple objectives such as distance, risk, and course changes. The smaller the fitness value, the better the route.

[0102] Step 3: For each particle, record its own historical best position (i.e., the position corresponding to the best fitness obtained during the particle search process), called the individual best position. Simultaneously, find the position of the particle with the best fitness in the entire particle swarm, as the global best position;

[0103] Step 4: Update the particle's velocity and position according to the following formula:

[0104]

[0105] In the formula, and These are the velocity and position of particle i in the k-th iteration, respectively; ω is the inertia weight, adjusting the degree to which the particle inherits its previous velocity; c1 and c2 are learning factors, controlling the degree to which the particle learns towards its individual optimal and global optimal positions; r1 and r2 are random numbers between [0,1]. When updating the position, ensure that the generated route meets the ship's navigation constraints, such as avoiding obstacles by determining whether the route passes through the dynamic influence range of obstacles, and maintaining a safe navigation speed by limiting the speed to within the safe navigation speed range. If the constraints are not met, adjust the particle position, for example, by regenerating some route node coordinates;

[0106] Step 5: Repeat steps 2 to 4, continuously updating the particle's velocity and position until the preset termination condition is met, such as reaching the maximum number of iterations or the global optimal solution no longer significantly improving within a certain number of iterations. At this point, the global optimal position is the desired optimal route.

[0107] The particle swarm optimization algorithm solves for the objective function J, and the specific steps are as follows:

[0108] Step 1: Randomly generate an initial population containing several individuals (each individual represents a shipping route). The individuals can be encoded using real number encoding or path encoding. The encoded gene string corresponds to the relevant parameters of the shipping route (such as node coordinates).

[0109] Step 2: Similarly, substitute each individual into the objective function J and calculate its fitness value to evaluate the individual's performance.

[0110] Step 3: Based on the fitness values, select a certain number of individuals from the current population using methods such as roulette wheel selection or tournament selection to serve as parents for the next generation. Individuals with higher fitness are more likely to be selected, thus ensuring that excellent flight path characteristics are preserved in the next generation.

[0111] Step 4: Perform a crossover operation on the selected parent individuals. This involves randomly selecting two parent individuals and exchanging partial gene segments of them at a certain crossover probability (e.g., 0.7-0.9) to generate new offspring individuals. During the crossover process, it is essential to ensure that the generated offspring's course meets ship navigation constraints. For example, check whether the new course will collide with obstacles; if a collision occurs, the crossover operation is repeated.

[0112] Step 5: Mutate the genes of some offspring individuals with a small mutation probability (e.g., 0.01-0.1), that is, randomly change some gene values ​​in the gene string to introduce new flight path features, increase the diversity of the population, and avoid the algorithm getting trapped in local optima. Similarly, after mutation, it is necessary to check whether the flight path meets the constraints; if not, it needs to be corrected.

[0113] Step Six: Repeat steps two through five to continuously evolve the population until the preset termination conditions are met, such as reaching the maximum number of generations or the population fitness value stabilizing. At this point, the route corresponding to the individual with the highest fitness in the population is the optimal route.

[0114] An objective function incorporating distance, risk, and course change is constructed to comprehensively consider the multi-objective requirements of ship navigation. An intelligent optimization algorithm is employed to solve the objective function, obtaining the optimal route for the ship while satisfying navigation constraints. The objective function comprehensively evaluates several important aspects of ship navigation; routes with shorter distances, lower risks, and smoother course changes score higher in the objective function evaluation. The intelligent optimization algorithm, through computational methods, finds the optimal route from numerous options that meets the requirements of the objective function.

[0115] During ship navigation, real-time data on actual navigation and the latest sea state are continuously collected. The actual data is compared and analyzed with model predictions. If deviations are found, such as sudden and significant changes in wind speed or deviations from the optimized course, a real-time adjustment mechanism is activated. Based on the latest data, the course optimization module is rerun to dynamically adjust the course, ensuring the ship navigates optimally under the current complex sea state. Simultaneously, the actual navigation data and optimization results are fed back to the ship navigation data acquisition and analysis module to continuously optimize the ship's motion model and course optimization algorithm, improving the model's accuracy and adaptability.

[0116] Example 2:

[0117] It collects multi-source information, including marine monitoring satellite imagery, shore-based radar monitoring data, Automatic Identification System (AIS) data, and underwater sonar detection data. Satellite imagery undergoes radiometric and geometric correction to improve image clarity; radar data is filtered and denoised to remove clutter interference; AIS data format is standardized and missing values ​​are filled in; and echo signal processing is performed on sonar data to extract obstacle contour information.

[0118] For satellite imagery and underwater sonar images, edge detection algorithms (such as the Canny operator) are used to extract the contour edges of obstacles, and shape features are then obtained using shape descriptors (such as Fourier descriptors) to determine whether they are fixed obstacles. For radar and AIS data, a Kalman filter algorithm is used to smooth the position and velocity data of obstacles and analyze their motion trajectories. If the trajectory does not change significantly over a long period of time and its relative position to the seabed topography is fixed, it is determined to be a fixed obstacle. Fixed obstacle information detected from different data sources is spatially matched and fused based on a geographic coordinate system to eliminate duplicate detections and errors, and to construct an accurate database of fixed obstacle locations.

[0119] The time-varying risk field of floating obstacles is divided into small grids in three-dimensional space, with each grid assigned a unique number and spatial coordinates. Based on the dynamic attitude model of the floating obstacles, their dynamic windward area, radius of influence, and trajectory uncertainty ellipse, the risk value of each small grid is calculated. A higher risk value indicates a greater probability of collision in that area.

[0120] 3D mesh risk value function:

[0121] R(i,j,k,t)=w1R A (i,j,k,t)+w2R U (i,j,k,t)+w3R E (i,j,k,t)

[0122] In the formula: R(i,j,k,t) is the risk value of the grid with coordinates (i,j,k) in three-dimensional space at time t; w1, w2, and w3 are weighting coefficients; R A (i,j,k,t) represents the dynamic influence range risk value of grid point (i,j,k) at time t; R U (i,j,k,t) represents the trajectory uncertainty risk value of grid point (i,j,k) at time t; R E (i,j,k,t) represents the environmental impact risk value of grid point (i,j,k) at time t;

[0123] The calculations for each component are as follows:

[0124]

[0125] In the formula: R A (i,j,k,t) represents the dynamic influence range risk value of grid point (i,j,k) at time t; d(i,j,k,t) represents the distance from grid point (i,j,k) to the center of the floating obstacle at time t; R1(t) represents the dynamic influence radius.

[0126]

[0127] In the formula, A(t) is the dynamic windward area of ​​the floating obstacle at time t, R0 is the static safety radius; ω1 and ω2 are phase parameters used to describe the periodicity of the dynamic windward area change; ψ1 and ψ2 are phase parameters used to adjust the starting phase of the dynamic windward area change; α1 is the expansion coefficient, which is determined by the obstacle size, with a value between 0.4 and 0.6 for large obstacles and between 0.1 and 0.2 for small obstacles; A0 is the static reference windward area. The attitude influence function reflects the roll angular velocity of the floating obstacle. and pitch angular velocity Impact on windward area;

[0128] Trajectory uncertainty risk:

[0129]

[0130] In the formula, R U (i,j,k,t) represents the trajectory uncertainty risk value of grid point (i,j,k) at time t; Σ is the covariance matrix of the trajectory uncertainty ellipse, describing the uncertainty distribution of the obstacle trajectory; (x,y,z) are the three-dimensional coordinates of the grid point, corresponding to (i,j,k); T represents transpose, used to interchange the rows and columns of a matrix or vector.

[0131] Environmental impact risks:

[0132]

[0133] In the formula, R E (i,j,k,t) represents the environmental impact risk value of grid point (i,j,k) at time t; v c (t) represents the ocean current velocity vector at time t, n represents the ocean current velocity vector at time t, and v th This is the speed threshold for environmental impact risk.

[0134] The edge of the fixed obstacle is discretized into multiple points, and each point is checked to see if it falls into a high-risk small grid in the risk field of the floating obstacle. If a point falls into a high-risk small grid, it is identified as a potential collision point.

[0135] For the initially screened potential collision points, the predicted trajectories of floating and fixed obstacles over a future period are used (based on a high-precision fluid-structure coupled dynamics model and a deep learning prediction model) to determine whether there is an actual collision hazard at that point. If the trajectories of the floating obstacles intersect at that point, then there is an actual collision hazard, and that point is confirmed as a real collision point.

[0136] Based on a fluid-structure coupled dynamics model, the collision process between a floating obstacle and a fixed obstacle at the collision point is simulated, and the direction of the resultant force on the floating obstacle at the instant of collision is calculated. Considering the influence of environmental factors such as ocean currents and waves on the collision direction, an environmental factor correction model is established to adjust the calculated resultant force direction. For example, if the angle between the ocean current direction and the initially calculated collision direction is less than a certain threshold, the component of the ocean current velocity in the collision direction is added to the collision force, and the collision direction is recalculated. The corrected resultant force direction is the collision direction.

[0137] Fluid-structure coupling force:

[0138] F coupling =F hydro +F structure

[0139] In the formula, F coupling The fluid-structure coupling force is the resultant force of the fluid forces and the structural forces; F hydro For fluid forces; F structure For structural forces;

[0140] Among them, the fluid force F hydro Solving the Navier-Stokes equations using CFD yields the following:

[0141]

[0142] In the formula, ρ is the fluid density; u is the fluid velocity field vector; p is the fluid pressure; μ is the dynamic viscosity of the fluid; Fb These are volume forces, such as gravity.

[0143] Collision direction after environmental correction:

[0144]

[0145] In the formula, d collision v is the corrected collision direction unit vector; enu β1 is the environmental factor velocity vector, such as the ocean current velocity vector; β1 is the environmental factor correction coefficient, adjusted according to the angle θ between the environmental velocity and the initial collision direction.

[0146]

[0147] (θ is the angle between the ambient velocity and the initial collision direction, θ) th Angle threshold corrected for environmental factors).

[0148] Using historical collision data and current floating and fixed obstacle states and environmental conditions, a machine learning model (such as Support Vector Machine (SVM) or Random Forest model) is trained. Relevant features of the collision point (such as distance from the obstacle center, relative velocity, attitude differences, and environmental parameters) are used as model input, and the collision probability of that point is output. Monte Carlo simulation is then used to perform numerous random simulations of the motion trajectories of floating and fixed obstacles, counting the number of collisions under the same conditions and calculating the collision frequency. This serves as a supplementary verification and adjustment basis for the collision probability.

[0149] Collision probability calculation

[0150] Machine learning model output probability:

[0151] P ML =σ(w1x1+w2x2+…+w n x n +b)

[0152] In the formula, P ML The collision probability is the output of the machine learning model; σ is the sigmoid function that maps the input to probability values ​​between 0 and 1; w1, w2, ..., w n These are the weight coefficients of each feature in the machine learning model; x1, x2, ..., x n 'b' represents the relevant feature vectors of the collision point, such as the distance from the center of the obstacle, relative velocity, attitude difference, environmental parameters, etc.; 'b' represents the bias term of the machine learning model.

[0153] Monte Carlo simulation frequency probability:

[0154]

[0155] In the formula, P MC N represents the collision probability obtained through Monte Carlo simulation. collision N represents the number of collisions that occur in the Monte Carlo simulation. simulation The total number of Monte Carlo simulations.

[0156] Overall collision probability:

[0157] P final =λP ML +(1-λ)P MC

[0158] Among them, P final The final collision probability is calculated by combining the results of the machine learning model and the Monte Carlo simulation; P ML P represents the collision probability output by the machine learning model. MC λ represents the collision probability obtained through Monte Carlo simulation; λ is the weighting coefficient of the machine learning model probability in the comprehensive collision probability calculation.

[0159] Obtain structural design drawings and material specifications for both fixed and floating obstacles, and establish three-dimensional structural models. Using finite element analysis software (such as ABAQUS), simulate the stress and strain distribution of the fixed and floating obstacles during the collision process to identify weak points and failure modes. Calculate the total energy released during the collision based on the mass, velocity, and relative motion state of the floating obstacle at the moment of impact.

[0160] Elastic equilibrium equations in stress-strain calculations using finite element analysis:

[0161]

[0162] in, σ is the divergence of the stress tensor, reflecting the variation of stress in space; σ is the stress tensor, describing the stress state at various points inside the object; f is the volume force vector, such as the force acting on the object like gravity.

[0163] Constitutive relation (linear elasticity):

[0164] σ=Dε;

[0165]

[0166] In the formula, D is the elasticity matrix, which describes the elastic properties of the material and is related to parameters such as the elastic modulus and Poisson's ratio; ε is the strain tensor, which describes the deformation of the object; and u is the displacement field vector, which describes the displacement of each point on the object.

[0167] Calculation of total energy released during the collision:

[0168]

[0169] In the formula, E collision v is the total energy released during the collision; m is the mass of the floating obstacle, v rel Let be the relative velocity vector at the instant of collision between the floating obstacle and the fixed obstacle.

[0170] This study considers the conversion of energy between different forms (such as kinetic energy, elastic potential energy, and thermal energy) and analyzes the destructive effect of energy on obstacle structures. Based on the Discrete Element Method (DEM), floating and fixed obstacles are discretized into multiple elements, and the fragmentation and splashing of elements during collisions are simulated. The connection strength and fracture criteria between elements are set, and the elements that will break are determined based on the total energy released during the collision and the structural stress distribution.

[0171] Element motion equations in discrete element method debris splash simulation:

[0172]

[0173] Where, m i Let r be the mass of the i-th unit. i Let F be the position vector of the i-th unit. ij Let F be the position vector of the i-th unit. ext,i This is the vector of external forces acting on the i-th unit, such as ocean currents or wave forces.

[0174] Element fracture criterion: Plastic deformation begins when the equivalent stress on an element exceeds the yield strength of the material; if it further exceeds the ultimate strength, the element fractures.

[0175] Considering the driving effects of ocean currents and waves on debris, a debris motion equation is established to calculate the debris's trajectory in the marine environment. Multiple simulations of the debris's trajectory are performed using this equation, and the probability of debris appearing in different regions is statistically analyzed. A debris splash probability distribution cloud map is then plotted to determine the impact range of debris splash. Based on the degree of obstacle damage, remaining buoyancy, and changes in the center of gravity after collision, combined with ocean circulation models (such as the ROMS model) and real-time ocean current and tidal data, the obstacle's drift path in the ocean is predicted. Particle swarm optimization or genetic algorithms are used to optimize the parameters of the prediction model, improving the accuracy of drift path prediction; thus, the physical collision risk zone is obtained.

[0176] Consider the debris motion equations of ocean currents:

[0177]

[0178] In the formula, m is the mass of the fragment; v1 is the velocity vector of the fragment; t is time; F dragThe drag force vector acting on the debris is related to factors such as the relative velocity between the debris and the water flow; F buoyancy F is the vector of the buoyant force acting on the fragment. current F is the force vector of the ocean current on the debris; wave This is the force vector of the wave acting on the debris;

[0179] Among them, drag force: (F drag The drag force vector acting on the fragment; ρ is the fluid density; C D V is the drag coefficient; A is the projected area of ​​the fragment perpendicular to the direction of motion; v water (This is the velocity vector of the water flow.)

[0180] Buoyancy: F buoyancy =ρgVk(F buoyancy (where is the buoyancy vector acting on the fragment; g is the acceleration due to gravity; V is the volume of the fragment; k is the unit vector in the vertical direction);

[0181] Ocean current force: (F current The vector of the force exerted by the ocean current on the debris; m is the mass of the debris; v current The vector of ocean current velocity output by an ocean circulation model (such as the ROMS model); t is time.

[0182] Wave force: F wave =ρgA wave sin(ωt)k(F wave The force vector of the wave on the debris; ρ is the fluid density; g is the acceleration due to gravity; A wave ω is the equivalent area of ​​the wave acting on the debris; k is the angular frequency of the wave; t is the unit vector in the vertical direction; and t is time.

[0183] Drift path prediction of floating obstacles or debris, based on drift velocity using ocean circulation models:

[0184] v drift =+v current +v wind_drag +v buoyancy

[0185] In the formula, v drift v is the obstacle drift velocity vector; current The current velocity vector output by an ocean circulation model (such as the ROMS model); v wind_drag v is the drag velocity vector of the wind on the obstacle; buoyancy The velocity vector caused by the buoyancy force acting on the obstacle;

[0186] Among them, v currentOutput of the ROMS model: the drag velocity vector v of the wind on the obstacle. wind_drag :

[0187]

[0188] (v wind_drag Let C be the drag velocity vector of the wind on the obstacle. w v is the wind drag coefficient. wind Let v be the wind speed vector. ref (For reference wind speed).

[0189] Particle swarm optimization parameter correction:

[0190]

[0191] Where, θ i is the i-th parameter in the model parameter vector; w is the inertia weight, which controls the influence of the particle's previous velocity on its current velocity; k is the number of iterations; c1 and c2 are learning factors, which control the degree to which the particle learns from the individual optimal solution and the global optimal solution; r1 and r2 are random numbers in the interval [0,1]. This is the parameter vector of the individual optimal solution at the k-th iteration; The parameter vector of the global optimal solution at the k-th iteration.

[0192] Using computational fluid dynamics (CFD) software (such as OpenFOAM), the Reynolds-averaged Navier-Stokes equations are solved with the obstacle state and ocean environmental parameters (current velocity, wave parameters, etc.) at the moment of collision as boundary conditions to simulate the changes in local water flow velocity and pressure fields caused by the collision. The Volume of Fluid (VOF) method is used to capture the deformation of the free surface of the water flow, considering the coupling effect of waves and water flow to more realistically simulate the changes in water flow. A mathematical model of ship motion is established, and the changes in local water flow velocity and pressure fields (such as changes in flow velocity and direction) obtained from the CFD simulation are input as external disturbance forces into the mathematical model of ship motion to calculate the ship's speed and heading deviation in the affected water flow. Taking into account the ship type (cargo ship, passenger ship, fishing vessel, etc.), load conditions, and handling performance parameters (rudder effectiveness, inertia, etc.), the degree of influence of water flow changes on ship handling is evaluated. The degree of influence is divided into different levels (such as slight, moderate, severe), and correspondingly different influence areas are defined; thus, the fluid dynamics influence zone is obtained.

[0193] The calculation results of the physical collision risk zone and the hydrodynamic influence zone are spatiotemporally aligned, and a weighted fusion algorithm is used to merge the two zones. Based on the degree of impact of different areas on ship navigation safety, different weights are assigned to the physical collision risk zone and the hydrodynamic influence zone to generate a comprehensive collision risk field. Geographic Information System (GIS) technology is used to visualize the comprehensive collision risk field, intuitively presenting the potential threat range to ship navigation.

[0194] Fundamental equations of computational fluid dynamics

[0195] Reynolds-averaged Navier-Stokes equations:

[0196]

[0197] Where ρ is the fluid density; The average velocity vector, For average pressure, Let F be the Reynolds stress tensor, describing the stress caused by turbulent fluctuations; μ is the dynamic viscosity of the fluid; F b This is a volume force vector.

[0198] VOF free surface capture:

[0199]

[0200] Where α2 is the fluid volume fraction, α2 = 1 indicates that the region is fluid, and α2 = 0 indicates that it is air; u is the fluid velocity vector.

[0201] Mathematical model of ship motion:

[0202]

[0203] In the formula, m is the mass of the ship; These are the acceleration components of the ship in the x and y directions, respectively; r is the angular velocity of the ship. X represents the angular velocity component of the ship; u and v represent the velocity components of the ship in the x and y directions, respectively; X hull Y hull N hull These represent the hydrodynamic forces acting on the hull in the x and y directions, and the torque about the z-axis, respectively; X rudder Y rudder N rudder These are the forces in the x and y directions and the torque about the z-axis, respectively; X current Y current N current These represent the forces exerted by the ocean current in the x and y directions, and the torque about the z-axis, respectively; I zThe moment of inertia of the ship about the z-axis; the subscript indicates the source.

[0204] Integration of the physical collision risk zone and the fluid dynamics influence zone:

[0205] R combined (x,y,z,t)=α3R physical (x,y,z,t)+β3R hydro (x,y,z,t)

[0206] In the formula, R combined (x,y,z,t) represents the risk value after merging the physical collision risk zone and the fluid dynamics influence zone; R physical (x,y,z,t) represents the risk value of the physical collision risk zone at coordinates (x,y,z) and time t; R hydro (x,y,z,t) represents the risk value of the fluid dynamics influence zone at coordinates (x,y,z) and time t; α3 and β3 are the fusion weight coefficients of the physical collision risk zone and the fluid dynamics influence zone, respectively.

[0207] The weighting coefficients α3 and β3 are calculated as follows:

[0208]

[0209] (Ε collision I represents the total energy released during the collision. hydro (Fluid influence index).

[0210] The comprehensive collision risk zone of floating obstacles is superimposed and fused with the time-varying risk field;

[0211] The risk levels of the comprehensive collision risk zone and the time-varying risk field are weighted.

[0212] Based on the weighted results, a new risk field is constructed to address potential threats to ship navigation.

[0213] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the multi-objective optimization model of ship routes for complex sea conditions and the inventive concept, should be covered within the scope of protection of the present invention.

Claims

1. A multi-objective optimization model for ship routes under complex sea state scenarios, characterized in that, Includes the following steps: By utilizing marine monitoring equipment, sensor networks, and ship-on-ship equipment, real-time data on ship navigation status and marine environment are collected, and historical data are combined to establish motion models of ships under different sea conditions. The initial route is generated by taking the starting point and destination as input and combining basic information from the ocean map. Determine if there are obstacles on the initial flight path. If there are: To determine whether an obstacle is a floating obstacle, if it is: Obtain the shape information of the obstacle, construct a dynamic attitude model of the floating obstacle, calculate the dynamic windward area, drag coefficient and radius of influence of the floating obstacle, and construct a time-varying risk field by combining the trajectory uncertainty ellipse; The objective function is constructed based on the time-varying risk field and the changes in flight distance and heading, and the optimal route is calculated. The dynamic attitude model of the floating obstacle is determined according to the following relationship: ; In the formula: Let the floating obstacle be the attitude vector. Let t be the roll angle of the floating obstacle. Let t be the pitch angle of the floating obstacle. Let t be the roll angular velocity of the floating obstacle. Let t be the pitch angular velocity of the floating obstacle. The specific steps for constructing a risk field are as follows: Based on the established dynamic attitude model, the dynamic windward area and dynamic drag coefficient of the obstacle are calculated; The radius of influence of floating obstacles is determined based on dynamic windward area and drag coefficient; Due to trajectory uncertainty caused by changes in obstacle attitude, an trajectory uncertainty ellipse is constructed. By combining the calculated dynamic influence radius R(t) of the floating obstacle and the trajectory uncertainty ellipse parameters, a time-varying risk field is constructed.

2. The multi-objective optimization model for ship routes under complex sea state scenarios according to claim 1, characterized in that, Obtain the operational status of the floating obstacle, detect whether there are fixed obstacles within the risk field of the floating obstacle, and if so: There is a risk of collision between floating obstacles and fixed obstacles. Potential collision points are constructed at the points where the risk fields of the fixed obstacle and the floating obstacle coincide. The collision direction of the floating obstacle at each collision point is obtained, and the comprehensive collision risk zone of the floating obstacle is calculated. Calculate the new risk field for floating obstacles.

3. The multi-objective optimization model for ship routes under complex sea state scenarios according to claim 2, characterized in that, The specific steps for calculating the comprehensive collision risk zone of the floating obstacle are as follows: Collect marine monitoring satellite imagery, shore-based radar monitoring data, ship automatic identification system data, and underwater sonar detection data; Fixed obstacle information detected from different data sources is spatially matched and fused based on a geographic coordinate system to eliminate duplicate detections and errors, and to build an accurate fixed obstacle location database. Obtain the actual collision point from the potential collision point and calculate the collision direction of the floating obstacle at the moment of collision; A machine learning model is trained using historical collision data and current floating and fixed obstacle states and environmental conditions. The model takes the relevant features of the collision point as input and outputs the collision probability of that point. The frequency of collisions between floating and fixed obstacles under the same conditions is calculated by counting the number of collisions. Obtain structural design drawings and material specifications for both fixed and floating obstacles, and establish 3D structural models; simulate the stress and strain distribution of the obstacles during the collision process to identify weak points and failure modes; calculate the total energy released during the collision based on the mass, velocity, and relative motion state of the floating obstacles at the moment of collision. Based on the total energy combined with the discrete element method, the debris motion equation is established to calculate the trajectory of the debris in the marine environment and determine the impact range of debris splash. Based on the degree of damage to the obstacle after the collision, the remaining buoyancy, and the change in the center of gravity, combined with the ocean circulation model and real-time ocean current and tidal data, the drift path of the obstacle in the ocean is predicted; thus, the physical collision risk zone is obtained. Using the obstacle state and marine environmental parameters at the moment of collision as boundary conditions, the changes in local water flow velocity and pressure fields caused by the collision are calculated. A mathematical model of ship motion is established, and the changes in local water flow velocity and pressure fields are input as external disturbance forces into the mathematical model of ship motion to calculate the ship's speed and heading deviation in the affected water flow. The degree of influence of water flow changes on ship handling is evaluated in combination with ship type, load conditions, and handling performance parameters. The degree of influence is divided into different levels, and different influence areas are defined accordingly. Thus, the hydrodynamic influence zone is obtained. The calculation results of the physical collision risk zone and the fluid dynamics influence zone are spatiotemporally aligned, and a weighted fusion algorithm is used to merge the physical collision risk zone and the fluid dynamics influence zone. Based on the degree of impact of different areas on ship navigation safety, different weights are assigned to the physical collision risk zone and the hydrodynamic influence zone to generate a comprehensive collision risk field.

4. The multi-objective optimization model for ship routes under complex sea state scenarios according to claim 2, characterized in that, The specific steps for calculating the new risk field of the floating obstacle are as follows: The comprehensive collision risk zone of floating obstacles is superimposed and fused with the time-varying risk field; The risk levels of the comprehensive collision risk zone and the time-varying risk field are weighted. Based on the weighted results, a new risk field is constructed to address potential threats to ship navigation.

5. The multi-objective optimization model for ship routes under complex sea state scenarios according to claim 2, characterized in that, The specific steps for detecting whether there are fixed obstacles within the risk field of floating obstacles are as follows: For satellite imagery and underwater sonar images, edge detection algorithms are used to extract the contour edges of obstacles, and then shape features are obtained through shape descriptors to determine whether they are fixed obstacles. For radar and AIS data, a Kalman filter algorithm is used to smooth the position and velocity data of obstacles and analyze their motion trajectory. If the trajectory does not change significantly over a long period of time and its relative position to the seabed topography remains fixed, it is determined to be a fixed obstacle.

6. The multi-objective optimization model for ship routes under complex sea state scenarios according to claim 3, characterized in that, The specific steps for obtaining the actual collision point based on the potential collision point are as follows: The time-varying risk field of floating obstacles is divided into small grids in three-dimensional space, and each small grid is assigned a unique number and spatial coordinates; Based on the dynamic attitude model of the floating obstacle, its dynamic windward area, the radius of influence, and the trajectory uncertainty ellipse, the risk value of each small grid is calculated. The higher the risk value, the greater the probability of collision in that area. The edges of fixed obstacles are discretized into multiple points, and each point is checked to see if it falls within a high-risk small grid of the risk field of floating obstacles. If a point falls into a high-risk small grid, then that point is identified as a potential collision point; For the potential collision points initially identified, the movement trajectories of the floating obstacles over a period of time are combined to determine whether there is an actual collision hazard at that point; if the movement trajectories of the floating obstacles intersect at that point, then there is an actual collision hazard at that point, and that point is confirmed as a real collision point.

7. The multi-objective optimization model for ship routes under complex sea state scenarios according to claim 3, characterized in that, The specific steps for calculating the collision direction of the floating obstacle at the moment of collision are as follows: Simulate the collision process between a floating obstacle and a fixed obstacle at the point of impact, and calculate the direction of the resultant force on the floating obstacle at the instant of collision; Considering the influence of environmental factors such as ocean currents and waves on the collision direction, an environmental factor correction model is established to adjust the calculated resultant force direction. The corrected resultant force direction is the collision direction.

8. The multi-objective optimization model for ship routes under complex sea state scenarios according to claim 3, characterized in that, The specific steps for determining the impact range of the debris splash are as follows: Based on the discrete element method, floating obstacles and fixed obstacles are discretized into multiple elements to simulate the breaking and splashing of elements during the collision process; The connection strength and fracture criteria between units are set, and the units that will break are determined based on the total energy released during the collision and the structural stress distribution. By combining the driving force of ocean currents and waves on debris, a debris motion equation is established to calculate the trajectory of debris in the marine environment; The trajectory of debris in the marine environment was simulated multiple times using debris motion equations. The probability of debris appearing in different areas was statistically analyzed, and a debris splash probability distribution cloud map was drawn to determine the impact range of debris splash.

Images