A digital decision-making method for ship navigation simulation integrating human factors

By integrating the digital decision-making method of ship navigation simulation with human factors, combining semi-physical simulation platform and human factors big data model, the problem of human factors neglect in the existing technology is solved, and high-precision ship motion forecasting in complex environments is achieved Safe navigation decisions.

CN118366067BActive Publication Date: 2024-11-08TAIHU LAB OF DEEPSEA TECH SCI +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410521248.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-28
Publication Date
2024-11-08
Estimated Expiration
2044-04-28

AI Technical Summary

Technical Problem

The existing ship motion forecasting methods ignore this key factor in humans, which leads to unrational manipulation in complex environments. Traditional virtual simulations cannot reflect individual differences in humans and cannot be applied to manipulation decisions in harsh environments and conflicting scenarios.

Method used

Using a digital decision-making method for ship navigation simulation integrating human factors, based on a semi-physical simulation platform, perceive human manipulation behavior, combines ship motion mechanics model and human factors big data model to form an intelligent navigation control strategy, and use convolutional neural network for forecasting and optimization .

Benefits of technology

It improves the forecast accuracy of hull motion response, broadens the application scenarios of semi-physical simulation technology, ensures safe navigation and operation of ships in complex environments, and provides high-precision navigation assisted decision-making.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118366067B_ABST
    Figure CN118366067B_ABST
Patent Text Reader

Abstract

The present invention discloses a digital decision-making method for ship navigation simulation integrating human factors, and relates to the technical field of ship motion prediction. The method comprises: creating a ship navigation scene; determining the initial motion state of the ship under the corresponding navigation environment in the scene; sensing human manipulation behavior and obtaining human manipulation device instructions; inputting the ship motion state at the current moment, the measured navigation environment signal, and the human manipulation device instructions into the ship motion mechanics model to obtain the ship motion response result; matching the human manipulation device instructions and the ship motion response result in real time, and deducing the ship motion response result of a certain human manipulation device instruction under the corresponding navigation environment at a future moment based on the matching result, thereby forming an intelligent navigation control strategy. The method fully considers the human manipulation factors and the impact of the environment on people, provides navigation auxiliary decision-making, and ensures the safety of ship navigation and operation at sea.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of ship motion prediction, and in particular to a digital decision-making method for ship navigation simulation integrating human factors characteristics. Background Art

[0002] Ship safety is affected by the coupling of various factors in the complex system of "people-ship-environment", among which people are in a dominant position and control the ship, while the navigation environment is the medium between people and ships. Statistics show that about 80% of ship safety accidents are related to crew members' wrong judgment and control, especially in high-risk working conditions such as harsh environments and conflict scenarios, people often lack experience and make irrational control. However, traditional virtual simulation only considers the relationship between the environment and the ship, and does not take the role of people into consideration. It does not reflect the individual differences of people and cannot be applied to control decisions in complex navigation environments.

[0003] At present, although semi-physical simulation technology that takes human control behavior into consideration has been applied to the ship and sea fields, the current level only takes human control as physical input, does not explore the deep reasons behind the behavior, and has unclear understanding of the relationship between people, ships, and the environment. This determines that this mode belongs to a qualitative simulation based on "image" and is mostly used for junior crew driving training.

[0004] In addition, at a time when human-machine collaboration has become an inevitable development trend, we have not fully utilized the powerful data processing and analytical computing functions brought about by the rapid development of computing power, nor have we combined people's professional knowledge, perception ability, and decision-making logic. This has wasted a large amount of control data based on different personnel as samples, and has failed to present a deeper application value. Summary of the invention

[0005] In view of the need for safe navigation and successful operation of ships in complex environments, and the defect of existing ship motion prediction methods that ignore the key factor of human factors, the present invention proposes a digital decision-making method for ship navigation simulation that integrates human factors characteristics. Based on the constructed semi-physical simulation platform, it perceives human control behavior in the real marine environment and forms an intelligent navigation control strategy, which can be used for auxiliary decision-making for ship navigation and operation at sea.

[0006] The technical solution of the present invention is as follows:

[0007] A digital decision-making method for ship navigation simulation integrating human factors characteristics comprises the following steps:

[0008] Create ship sailing scenes;

[0009] In the ship navigation scenario, select to manipulate the ship and determine its initial motion state in the corresponding navigation environment;

[0010] Perceive human manipulation behavior and obtain instructions for human-controlled equipment;

[0011] The current ship motion state, measured navigation environment signals, and human control equipment instructions are input into the ship motion mechanics model to obtain the ship motion response results, including various forces on the ship and the six-degree-of-freedom motion of the hull;

[0012] The human-controlled device instructions and the ship motion response results are matched in real time, and based on the matching results, the ship motion response results of certain human-controlled device instructions in the corresponding navigation environment at a future moment are deduced, thereby forming an intelligent navigation control strategy;

[0013] Re-execute perceived human maneuvers until the end of the voyage.

[0014] Its further technical solution is to create a ship navigation scene, including:

[0015] According to the longitude and latitude coordinate data corresponding to the geographical location of the selected navigation area, the hydrological data under the corresponding wave level are retrieved from the real sea area wave database, the light parameters and weather display characteristics in the screen are changed according to the set time and weather conditions, and the passive target objects in the sea area are generated using the random distribution function algorithm. The display data is synchronously updated in the radar chart, electronic chart and AIS equipment through different interfaces and protocols to complete the creation of the initial ship navigation scene in the digital space.

[0016] A further technical solution is to determine the initial motion state of the maneuvering ship in the corresponding navigation environment, including:

[0017] The three-dimensional ship model including propellers and rudders is called from the ship performance database to generate waves radiating outward from the center of the hull in the initial ship navigation scenario, and the effects of wind, waves and currents on the ship are converted into digital signals that match the phase of the ship's motion as the initial state of the ship in the corresponding navigation environment.

[0018] A further technical solution is to generate waves radiating outward from the center of the hull in the initial ship navigation scenario, including:

[0019] The instantaneous wave surface data of the actual sea area around the hull is obtained, and the ocean water color, current direction, and wave height obtained based on real ocean satellite observations are used as the external wave surface data of the hull. The two types of wave surface data are matched and the wave-making characteristics are superimposed around the ship to obtain the waves radiating outward from the center of the hull under the initial ship navigation scenario.

[0020] A further technical solution is that the force of wind on the ship is expressed as:

[0021]

[0022] Among them, ρ ais the air density; F Xw 、F Yw 、N w are longitudinal wind force, lateral wind force and yaw moment respectively; C xw , C yw , C xyw Corresponding to the longitudinal wind force coefficient, lateral wind force coefficient and bow wind moment coefficient; A T , A L are the bow and lateral wind receiving areas respectively; L represents the length of the ship; α wr V is the relative angle of attack of the wind; wr is the relative wind speed;

[0023] The force and moment of waves on the ship are expressed as:

[0024]

[0025] Among them, X wave , Y wave 、N wave represents the longitudinal force, transverse force and pitch moment of the wave acting on the ship; Δp is the dynamic pressure of the ship in the wave; τ is the flooded volume of the ship; x and y correspond to the longitudinal and transverse directions in the coordinate system respectively;

[0026] The force and moment of the flow on the ship are expressed as:

[0027]

[0028] Where, ρ is the density of water; d is the draft of the ship; Y L 、N L represents the lateral force and yaw moment of the flow acting on the ship; V r is the relative speed of the ship to the water; parameters a and b are expressed as:

[0029]

[0030] Among them, Y v ', Y r '、N v '、N r ' is the dimensionless value of the corresponding fluid dynamic derivative; r represents the yaw angular velocity; v r It is the lateral projection of the relative speed of the ship to the water.

[0031] Its further technical solution is that the perception of human manipulation behavior includes taking different operators as samples:

[0032] By using physical control equipment to reflect human's manipulation of the ship, the human control equipment instructions under the influence of the current navigation environment and the ship's motion are read and converted into digital signals as a type of human factor feature sample;

[0033] With the help of the camera array and motion capture equipment arranged in the cabin, the action videos of different operators and the reaction time from judging the environment to operating the clock and rudder shaft are collected under the condition of constantly updating the navigation environment, and the image feature points are compared to convert the action features and reaction time into digital signals as another type of human factor feature samples. The action features include the position of the operator's hands, blinking frequency, breathing frequency and control speed.

[0034] Two types of human factor feature samples are used to train the human factor big data model, so that it can output the probability of the operator's manipulation behavior under the current physiological state.

[0035] Its further technical solution is to perceive human manipulation behavior, and also includes:

[0036] With the help of the camera array and motion capture equipment arranged in the cabin, the displacement and rotation data of the orbital plane relative to the cockpit are also collected as the displacement and rotation parameters of the scene acquisition camera, so as to realize the field of view tracking in the simulation cabin.

[0037] Its further technical solution is that the method further comprises:

[0038] After obtaining the ship motion response results after a certain human-controlled device instruction, the time series fusion method of the convolutional neural network is used to offline learn the expert judgment data and accident report data to supplement and optimize the predicted human-controlled device instructions.

[0039] The beneficial technical effects of the present invention are:

[0040] (1) Taking into full consideration the human control factors and the impact of the environment on humans, the behavioral data of different operators are used as samples to construct a high-precision ship motion response module that integrates human factors. This overcomes the defect that traditional prediction models are not suitable for dangerous navigation environments, complex operating conditions, and conflict confrontation scenarios. The prediction accuracy of the hull motion response is improved. Based on the true restoration of the navigation environment, the real-time capture of human control, and the correct response of the hull motion, the application scenarios and implementation functions of semi-physical simulation technology are broadened.

[0041] (2) A high-precision ship motion response module used for simulation and prediction and a learning module based on the time series fusion method of convolutional neural network are combined into a dual-line drive control strategy training system. The system takes into account human operating habits, draws lessons from accidents and learns from expert experience, thereby supplementing and optimizing the predicted human control equipment instructions, conforming to the development trend of human-machine collaboration, providing navigation assistance decisions, and ensuring the safety of ships' navigation and operations at sea.

[0042] (3) With the help of hidden camera arrays and motion capture equipment, the changes in the observation plane caused by the active behavior of the operator can be captured without destroying the simulation environment, and the update of the visual content can be affected by the hull motion response, further enhancing the immersion of the simulation process and improving the training effect. The camera array and motion capture equipment can also be used to collect behaviors during the process of pushing carts, steering, issuing flag signals, etc., making the semi-physical simulation platform have the development potential of human factors engineering research.

[0043] (4) Different from the traditional model, quantitative methods are no longer used to evaluate control behavior. Instead, based on the observation, identification, and analysis of dynamic behavior, digital methods are used to describe the interactive events in the digital space of human-ship-environment. By utilizing powerful computing resources and considering the impact of the environment on the ship and the response of the ship to humans, human control behavior can be calculated, thus forming a digital navigation decision-making technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is a flow chart of a digital decision-making method for ship navigation simulation integrating human factors provided in this application.

[0045] Figure 2 This is a schematic diagram of the architecture of a high-fidelity environment ship navigation semi-physical simulation platform provided by this application.

[0046] Figure 3 This is a schematic diagram of constructing a human factors big data model provided by this application.

[0047] Figure 4 It is a digital decision training flow chart provided by this application. DETAILED DESCRIPTION

[0048] The specific implementation of the present invention will be further described below in conjunction with the accompanying drawings.

[0049] Combination Figure 1 , Figure 2 As shown, the present application provides a digital decision-making method for ship navigation simulation integrating human factors characteristics. The method is based on Figure 2 The implementation of the high-fidelity ship navigation semi-physical simulation platform shown in the figure specifically includes the following steps:

[0050] Step 1: Create a ship sailing scene.

[0051] In the navigation simulation system, a ship navigation scene including geographical location, wave environment, time and weather, etc. is set. Specifically, according to the latitude and longitude coordinate data corresponding to the geographical location of the selected navigation area, the hydrological data under the corresponding wave level is retrieved from the real sea area wave database, and the light parameters, weather (such as fog / rain / snow) and other display features in the screen are changed according to the set time and weather conditions. The passive target objects in the sea area are generated using the random distribution function algorithm, and the display data is synchronously updated in the radar chart, electronic chart and AIS equipment through different interfaces and protocols to complete the creation of the initial ship navigation scene in the digital space.

[0052] In this embodiment, the sea area selected is Qingdao Port, corresponding to the longitude and latitude coordinates of 120°19'E, 36°04'N, the sailing time is 8 am, the foggy weather with a visibility of 200 meters is set, and a passive target object (such as a container ship) is generated with a speed of 18kn heading towards the port. The hydrological and meteorological data such as wind, waves, currents, and tides at the corresponding time are retrieved, and after digital analysis, they are connected to the semi-physical simulation platform to create an initial navigation scene. It is ensured that after operations such as zooming the screen and switching the perspective, the spatial structure, color texture, brightness and other features within the display range of the simulation platform are consistent with the real scene.

[0053] Step 2: In the ship navigation scenario, select the ship to be manipulated and determine its initial motion state in the corresponding navigation environment.

[0054] In the ship navigation scenario, the active target object (i.e. the maneuvering ship) is selected, and the three-dimensional ship model including the propeller and rudder is called from the ship performance database to generate waves radiating outward from the center of the hull in the initial ship navigation scenario. The effects of wind, waves, and currents on the ship are converted into digital signals that match the phase of the ship's motion as the initial state of the ship in the corresponding navigation environment, and then written into the real-time data transmission channel.

[0055] Among them, generating waves radiating outward from the center of the hull in the initial ship navigation scenario includes: obtaining instantaneous wave surface data of the actual sea area around the hull, using the ocean water color, current direction, and wave surface height obtained according to real ocean satellite observations as external wave surface data of the hull, matching the two types of wave surface data and superimposing wave-making characteristics around the ship to obtain waves radiating outward from the center of the hull in the initial ship navigation scenario.

[0056] Among them, the effects of wind, waves and currents on ships are specifically manifested as follows. Assuming that the speed and direction of the sea breeze are slowly varying parameters, the force of wind on the ship is expressed as:

[0057]

[0058] Among them, ρ a is the air density; F Xw 、FYw 、N w are longitudinal wind force, lateral wind force and yaw moment respectively; C xw , C yw , C xyw The corresponding longitudinal wind force coefficient, transverse wind force coefficient and bow wind moment coefficient can be determined by empirical formula according to wind direction angle, loading conditions and bow shape; A T , A L are the bow and lateral wind receiving areas respectively; L represents the length of the ship; α wr V is the relative angle of attack of the wind; wr is the relative wind speed.

[0059] The forces and moments of waves on ships include first-order wave forces and second-order wave forces, which can be uniformly expressed as follows:

[0060]

[0061] Among them, X wave , Y wave 、N wave represents the longitudinal force, transverse force and yaw moment of the wave acting on the ship; Δp is the dynamic pressure of the ship in the wave; τ is the flooded volume of the ship; x and y correspond to the longitudinal and transverse directions in the coordinate system respectively.

[0062] The linear fluid transverse force and moment of the flow on the ship are expressed as:

[0063]

[0064] Where, ρ is the density of water; d is the draft of the ship; Y L 、N L represents the lateral force and yaw moment of the flow acting on the ship; V r is the relative speed of the ship to the water; parameters a and b are expressed as:

[0065]

[0066] Among them, Y v ', Y r '、N v '、N r ' is the dimensionless value of the corresponding fluid dynamic derivative; r represents the yaw angular velocity; v r is the horizontal projection of the relative speed of the ship to the water, which can be understood as V r The velocity component of .

[0067] In this embodiment, the active target object is selected as an ONR (Office Of Naval Research) inboard ship, with a vertical length (i.e., ship length) of 154m, a maximum waterline width of 18.78m, a draft of 5.49m, an initial position with a longitudinal distance of 500m from the passive target object, and a lateral distance of 30m. The two ships are heading in opposite directions, and there is a risk of collision if they are not maneuvered to avoid in time. From the current moment, the entire domain displays the continuous development and change process of wind, waves, currents, tides, sea fog, etc. According to the visual refresh frequency of 200Hz, the meteorological satellite observation data is decomposed into time calendar data, and written into the real-time data transmission channel every 5ms. At the same time, the wave-making features such as opening waves and wakes are superimposed on the original wave field near the ship in the simulated visual scene to ensure that the navigation environment constructed after adding the target ship takes into account both physical and mechanical characteristics.

[0068] Step 3: Sense human manipulation behavior and obtain instructions for human-controlled equipment.

[0069] like Figure 3 As shown in the figure, in order to reflect the individual differences of different operators, this method uses different operators as samples, uses physical control equipment such as car bells and rudder shafts to reflect people's manipulation of ships, reads the instructions of human control equipment under the influence of the current navigation environment and hull motion, and converts the physical signals into digital signals, which are transmitted to the ship motion mechanics model in real time as input. The instructions of human control equipment can be car commands and rudder commands. The control time, positive and negative direction, and amplitude are separated from the converted digital signals and stored in the manipulation behavior database as a type of human factor feature samples.

[0070] At the same time, with the help of the camera array and motion capture equipment arranged in the simulation cabin, under the condition of constantly updating the navigation environment, the action videos of different operators in the process of pushing the car, steering, issuing flags, etc. and the reaction time of people from judging the environment to controlling the car clock and rudder axis are collected, and the image feature points are compared. The action features and reaction time are converted into digital signals and stored in the manipulation behavior database as another type of human factor feature samples, where the action features include the position of the personnel's hands, blinking frequency, breathing frequency and manipulation speed. Finally, the two types of human factor feature samples are used to train the human factor big data model to output the probability of the manipulation behavior of the operator under the current physiological state.

[0071] Optionally, the displacement and rotation data of the orbital plane relative to the cockpit can be collected with the help of a camera array and motion capture equipment arranged in the cabin, and used as the displacement and rotation parameters of the view acquisition camera to achieve field tracking in the simulation cabin.

[0072] In this embodiment, mixed reality technology is used to compare the feature points of image data through the depth sensor and camera array of the motion capture device, and to determine the acceleration and angular velocity of the person's hand relative to the real space positioning point in the three coordinate axes of XYZ. According to the coordinate conversion relationship obtained in advance, the real position information is transformed into the visual coordinate system, and the blinking frequency, breathing frequency, manipulation speed and other data are recorded. The personal manipulation behavior characteristics of the person in the whole process of perceiving the outside world, processing information and responding are converted into digital signals to train the human factors big data model and obtain the behavior probability of the person manipulating the person under the current physiological state.

[0073] Step 4: Calculation of ship motion response: Input the current ship motion state, measured navigation environment signals, and human control equipment instructions into the ship motion mechanics model to obtain the ship motion response results.

[0074] Specifically, the ship propeller and rudder information and performance parameters (such as the ship's maneuverability hydrodynamic derivative, drag coefficient, roll damping coefficient, propeller and rudder installation method, ship propeller and rudder interference coefficient, etc.) in the ship performance database are called, and the navigation environment signal that changes with the ship in the real-time data transmission channel is read. The update frequency is adjusted according to the calculation time step, and the propeller and rudder control data of the ship by the operator is connected at the same time. Based on the propeller and rudder response curve, the hydrodynamic force, wind force, wave force, propeller force, rudder force and other forces and the six-degree-of-freedom motion of the hull that change with time during the navigation process are calculated, and the navigation scene in the current motion state is synchronously updated through coordinate conversion. During the navigation process, since the relative position of the ship and the wave will change in real time, the various forces and hull motion conditions after the relative position of the ship and the wave changes are recalculated and written into the real-time data transmission channel to cover the data of the previous moment, so as to realize the iteration and update of the data.

[0075] The ship motion mechanics model selected in this embodiment is implemented based on an existing mechanics model and will not be described in detail here.

[0076] Step 5: Training of AI decision making.

[0077] like Figure 4 As shown in the figure, after all interactive events in the digital space composed of people, ships and environment are expressed in digital form, the delay effect is considered, and the human control equipment instructions in step 3 and the ship motion response results in step 4 are matched in real time. The dynamic control behavior in the human factor feature sample is analyzed based on the rich computing power, and the ship motion response results of a certain human control equipment instruction in the corresponding navigation environment at a future moment are deduced in combination with the matching results. The online prediction speed is achieved in seconds, and the time series fusion method of the convolutional neural network is used to learn the expert judgment data and accident report data offline, so as to supplement and optimize the predicted human control equipment instructions.

[0078] In this embodiment, 1.2P FP32 high-performance parallel computing power and 24PFP16 artificial intelligence training computing power are built based on the Ascend 910A card, which are used for online training prediction and offline training learning respectively. According to the expert's driving experience of active target objects and the analysis results of accident cases such as ship collisions in foggy weather, the decision forecast of the current operator is supplemented and optimized, and an "online + offline" dual-line drive control strategy training system is adopted to form intelligent navigation decisions based on a semi-physical simulation platform.

[0079] Step 6: Formation of intelligent navigation control strategy.

[0080] Re-execute the perceived human manipulation behavior until the end of the voyage, that is, repeat steps 3 to 5. Based on a large amount of sample data consisting of a large number of human-controlled equipment instruction sets and the calculation results of the high-precision ship motion response module, a self-growing and self-learning intelligent navigation decision-making algorithm is formed. For passive target objects in the environment, the human manipulation speed and the ship's response time are considered, and the coupling performance of the ship's own speed, maneuverability, seakeeping, etc. is combined. Operation prompts are given in time through the visualization system to avoid navigation accidents caused by improper manipulation, excessive speed, slow avoidance and other decision-making errors. The real-time ship response after this manipulation is continuously included in the sample library, and the manipulation strategy training system is continuously optimized.

[0081] In this embodiment, according to the current operator's habit of steering first and then operating the car clock in a critical state, combined with the speed and relative position of the active target object and the passive target object, the rudder response time and change trend are considered to provide safe navigation decisions and specific operation time, and a text prompt is given in the vision through the decision prompt visualization module. If the personnel still does not take avoidance measures, a voice prompt is issued before the limit operation time, and the control behavior is continuously converted into a digital signal, which is included in the dual-line drive control strategy training system for continuous learning.

[0082] The digital decision-making method for ship navigation simulation integrating human factors proposed in this application utilizes abundant computing power and combines human factors to express interactive scenarios, control behaviors and response results in a digital way. Through technologies such as artificial intelligence and machine learning, it processes and analyzes multi-source fusion data including simulation data, expert judgment data, accident report data, etc., to form a training system under a dual-line drive control strategy, which is used to calculate human control decisions during navigation, deduce ship motion under different control strategies in the same scenario, provide navigation auxiliary decision-making, and ensure the safety of ship navigation and operation at sea.

[0083] The above is only a preferred embodiment of the present application, and the present invention is not limited to the above embodiments. It is understood that other improvements and changes directly derived or associated by those skilled in the art without departing from the spirit and concept of the present invention should be considered to be included in the protection scope of the present invention.

Claims

1. A digital decision-making method for ship navigation simulation integrating human factors, characterized in that: include: Create ship sailing scenes; In the ship navigation scenario, selecting a maneuverable ship and determining its initial motion state in a corresponding navigation environment; Perceive human manipulation behavior and obtain instructions for human-controlled equipment; The current ship motion state, measured navigation environment signals, and human control equipment instructions are input into the ship motion mechanics model to obtain the ship motion response results, including various forces on the ship and the six-degree-of-freedom motion of the hull; The human control device instruction and the ship motion response result are matched in real time, and based on the matching result, the ship motion response result after a certain human control device instruction is deduced in a corresponding navigation environment at a future moment, so as to form an intelligent navigation control strategy; Re-execute the above-mentioned perceived human manipulation behavior until the end of the voyage; Among them, determining the initial motion state of the ship under the corresponding navigation environment includes: The three-dimensional ship model including propellers and rudders is called from the ship performance database to generate waves radiating outward from the center of the hull in the initial ship navigation scenario, and the effects of wind, waves and currents on the ship are converted into digital signals that match the phase of the ship's motion as the initial state of the ship in the corresponding navigation environment; The generating of waves radiating outward from the center of the hull in the initial ship navigation scene includes: Obtain instantaneous wave surface data of the real sea area around the hull, use the ocean water color, current direction, and wave surface height obtained from real ocean satellite observations as the external wave surface data of the hull, match the two types of wave surface data and superimpose the wave-making characteristics around the ship to obtain the waves radiating outward from the center of the hull in the initial ship navigation scenario; Among them, the force of wind on the ship is expressed as: Among them, ρ a is the air density; F Xw 、F Yw 、N w are longitudinal wind force, lateral wind force and yaw moment respectively; C xw , C yw , C xyw Corresponding to the longitudinal wind force coefficient, lateral wind force coefficient and bow wind moment coefficient; A T , A L are the bow and lateral wind receiving areas respectively; L represents the length of the ship; α wr V is the relative angle of attack of the wind; wr is the relative wind speed; The force and moment of waves on the ship are expressed as: Among them, X wave , Y wave 、N wave represents the longitudinal force, transverse force and pitch moment of the wave acting on the ship; Δp is the dynamic pressure of the ship in the wave; τ is the flooded volume of the ship; x and y correspond to the longitudinal and transverse directions in the coordinate system respectively; The force and moment of the flow on the ship are expressed as: Where, ρ is the density of water; d is the draft of the ship; Y L 、N L represents the lateral force and yaw moment of the flow acting on the ship; V r is the relative speed of the ship to the water; parameters a and b are expressed as: Among them, Y v ', Y r '、N v '、N r ' is the dimensionless value of the corresponding fluid dynamic derivative; r represents the yaw angular velocity; v r is the lateral projection of the relative speed of the ship to the water.

2. The digital decision-making method for ship navigation simulation integrating human factors according to claim 1 is characterized in that: The creating of the ship navigation scene comprises: According to the longitude and latitude coordinate data corresponding to the geographical location of the selected navigation area, the hydrological data under the corresponding wave level are retrieved from the real sea area wave database, the light parameters and weather display characteristics in the screen are changed according to the set time and weather conditions, and the passive target objects in the sea area are generated using the random distribution function algorithm. The display data is synchronously updated in the radar chart, electronic chart and AIS equipment through different interfaces and protocols to complete the creation of the initial ship navigation scene in the digital space.

3. The digital decision-making method for ship navigation simulation integrating human factors according to claim 1 is characterized in that: The sensing of human manipulation behavior includes taking different operators as samples: By using physical control equipment to reflect human's manipulation of the ship, the human control equipment instructions under the influence of the current navigation environment and the hull motion are read and converted into digital signals as a type of human factor feature sample; With the help of the camera array and motion capture equipment arranged in the cabin, the action videos of different operators and the reaction time of people from judging the environment to operating the clock and rudder shaft are collected under the condition of constantly updating the navigation environment, and image feature points are compared to convert the action features and reaction time into digital signals as another type of human factor feature samples. The action features include the position of the person's hand, blinking frequency, breathing frequency and operation speed. Two types of human factor feature samples are used to train the human factor big data model, so that it can output the probability of the operator's manipulation behavior under the current physiological state.

4. The digital decision-making method for ship navigation simulation integrating human factors according to claim 3 is characterized in that: The sensing of human manipulation behavior further includes: With the help of the camera array and motion capture equipment arranged in the cabin, the displacement and rotation data of the orbital plane relative to the cockpit are also collected as the displacement and rotation parameters of the scene acquisition camera, so as to realize the field of view tracking in the simulation cabin.

5. The digital decision-making method for ship navigation simulation integrating human factors according to any one of claims 1 to 4, characterized in that: The method further comprises: After obtaining the ship motion response results after a certain human-controlled device instruction, the time series fusion method of the convolutional neural network is used to offline learn the expert judgment data and accident report data to supplement and optimize the predicted human-controlled device instructions.