A shipboard intelligent navigation assistance method and system

Abstract

The application provides a kind of shipborne intelligent navigation auxiliary method and system, belongs to ship intelligent navigation and automatic driving technology.The method is by collecting multi-source perception information, based on international maritime collision avoidance rules to generate multiple candidate navigation strategies, using comprehensive evaluation model to quantitatively evaluate each candidate navigation strategy, determine the final execution strategy through man-machine cooperation mode, and continuously predict risks and actively intervene in safety based on digital twin system during execution.The application adopts the above-mentioned shipborne intelligent navigation auxiliary method and system, solves the problem of single decision target, insufficient man-machine cooperation and limited safety warning capability in the prior art, realizes the comprehensive optimization of navigation safety, efficiency and rule compliance, and improves the intelligent level and safety guarantee capability of ship navigation.

Description

Technical Field

[0001] This invention relates to the field of intelligent navigation and automatic driving technology for ships, and in particular to a shipborne intelligent navigation assistance method and system. Background Technology

[0002] Intelligent navigation assistance technologies for ships are crucial for improving maritime navigation safety and efficiency. However, existing technologies have the following shortcomings in achieving effective human-machine intelligent integration and reliable active safety: First, the isolated and static nature of decision-making models makes it difficult to cope with the complex multi-objective optimization and dynamic changes in navigation. Existing systems typically employ isolated and singular decision-making logic, considering only geometric collision avoidance or emphasizing only rule compliance, lacking a comprehensive evaluation mechanism that can simultaneously quantify and balance safety, economic efficiency, and rule compliance. This results in generated strategies that often address one aspect while neglecting another, failing to meet the demand for a comprehensive optimal solution in actual navigation. More importantly, the decision-making models of such systems are statically preset and completely unable to perceive and respond to key human factors, such as changes in crew fatigue and distraction. This leads to a disconnect between the system's auxiliary strategies and the actual needs and capabilities of the crew, preventing intelligent adaptive collaboration.

[0003] Second, the superficial nature of system interaction and monitoring leads to insufficient human-machine trust and a lack of safety redundancy. On the one hand, the interaction modes of existing systems are polarized: either the automated decision-making process is opaque, making it difficult for crew members to understand and trust; or it provides simple information prompts, placing the entire decision-making burden on the crew. The lack of a transparent process that allows crew members to deeply participate, compare and adjust multiple strategies, and ultimately make collaborative decisions makes the system difficult to effectively trust and use. On the other hand, existing technologies rely heavily on static risk assessments of the current environment, lacking the ability to predict risks proactively based on high-fidelity ship dynamics models. The system can only issue passive warnings when danger is imminent, unable to anticipate and proactively intervene when risks are still developing, making the system's final safety defense line very weak.

[0004] Therefore, there is an urgent need in this field for an intelligent navigation assistance method that can fundamentally solve the above problems and achieve multi-objective dynamic optimization decision-making and deep human-machine collaborative active safety. Summary of the Invention

[0005] The purpose of this invention is to provide a shipborne intelligent navigation assistance method and system, which comprehensively improves the decision-making quality, collaborative efficiency and safety assurance level of ship navigation by establishing a multi-objective comprehensive evaluation model, realizing a human-factor adaptive weight adjustment mechanism, constructing a high-fidelity digital twin prediction system, and combining a transparent human-machine collaborative decision-making process.

[0006] To achieve the above objectives, the present invention provides a shipborne intelligent navigation assistance method, comprising the following steps: Step S1: Collect the ship's sensor information; Step S2: Based on the perception information, generate one or more candidate navigation strategies; Step S3: Using a preset comprehensive evaluation model, quantitatively evaluate the candidate navigation strategies to obtain the evaluation value of each candidate navigation strategy. Step S4: Provide the candidate navigation strategies and their evaluation values ​​to the crew, and determine the final navigation strategy based on the crew's feedback; Step S5: Execute the final navigation strategy and continuously predict risks during the execution process. When potential risks are predicted, trigger proactive safety intervention.

[0007] Preferably, the perceived information in step S1 includes: External environment information: Dynamic information of other ships acquired by Automatic Identification System (AIS) and radar, including the position, course, and speed of other ships; Static obstacle information acquired by electronic charts, including information on reefs, shoals, and fixed bridges; Ship status information: position, heading, and speed of the ship obtained from the positioning and attitude determination system; load information obtained from the ship's load line system, including displacement and load status; Crew status information: information on the crew's line of sight direction obtained from in-cabin vision sensors.

[0008] Preferably, the generation of candidate navigation strategies in step S2 is based on the dynamic information of other ships and the ship's position, course, and speed, and includes the following steps: Step S21: Based on the relative motion relationship between the ship and other ships, classify the situation as a head-on encounter, a cross encounter, or an overtaking encounter. Step S22: Generate an initial strategy for the encounter situation based on the International Maritime Collision Prevention Regulations. The initial strategy is to maintain direction and speed, or turn right or left. Step S23: Generate candidate navigation strategies using the strategy perturbation formula. : ; in, Indicates the heading disturbance angle; This indicates the amount of speed disturbance.

[0009] Preferably, the comprehensive evaluation model in step S3 is used to evaluate candidate navigation strategies. The objective function for determining superiority or inferiority is expressed as follows: ; in, Indicate candidate navigation strategies The overall assessment value; Indicate candidate navigation strategies Risk indicators; Indicate candidate navigation strategies Efficiency indicators; Indicate candidate navigation strategies Rule compliance index; , , This indicates the weight of each indicator.

[0010] Preferred risk indicators The calculation formula is as follows: ; in, This represents the traversal of each other ship and static obstacle obtained; Indicates the execution of candidate navigation strategies At that time, the ship was close to the target The prediction is that the nearest distance will be encountered; Indicates the execution of candidate navigation strategies At that time, the ship was close to the target The prediction is that it will encounter time soon; Indicate candidate navigation strategies The risk of stranding; , , Represents the normalization coefficient; The formula for calculating efficiency indicators is as follows: ; in, Indicates the execution of candidate navigation strategies The estimated total distance traveled at that time; Indicates the execution of candidate navigation strategies The estimated total sailing time; This indicates the total distance of the originally planned route; This indicates the estimated total time for the originally planned route; , This indicates the preset weighting coefficients; The formula for calculating the rule compliance index is as follows: ; in, Indicate candidate navigation strategies Recommended course; Indicate candidate navigation strategies Recommended speed; Indicates the initial policy The course; Indicates the initial policy The speed of the ship; The normalization factor representing the heading deviation; The normalization factor representing the speed deviation; , This indicates the preset weighting coefficient.

[0011] Preferably, crew status information is used to dynamically adjust the weight coefficients in the objective function, and the specific process is as follows: Calculate the distraction index based on crew member's line of sight information. The calculation formula is: ; in, Indicates the duration of the statistics; This indicates the total time during which the crew's line of sight deviated from the preset forward channel area within the statistical period; Adjust the weighting coefficients using the following formula: ; ; in, This indicates the preset adjustment gain coefficient; This indicates the adjusted risk indicator weights; This indicates the adjusted weights of the efficiency indicators.

[0012] Preferably, step S4 includes the following steps: Step S41: Provide a graphical human-computer interaction interface to display all candidate navigation strategies to the crew in a list and route overlay format, and clearly indicate the comprehensive evaluation value of each strategy. Risk indicators Efficiency indicators and rule compliance indicators ; Step S42: The system listens for and receives feedback from the crew, including one or more of the following: a) Directly select a candidate navigation strategy; b) Manually modify the heading or speed parameters of a candidate navigation strategy; c) Ignore all candidate navigation strategies and execute autonomous manual navigation; Step S43: Determine the final navigation strategy to be executed based on the feedback: If the feedback is a), then the selected strategy will be executed directly as the final strategy. If the feedback is b), the system will use the modified strategy as the new candidate strategy, re-execute step S3 for quantitative evaluation, and use the new strategy after evaluation as the final strategy. If the feedback is c), the system enters monitoring mode, and the final navigation strategy is the crew's real-time manual operation instructions; If within the preset time If no feedback is received, the comprehensive evaluation value will be automatically applied. The optimal candidate navigation strategy.

[0013] Preferably, the continuous risk prediction in step S5 includes the following steps: Step S51: Construct a digital twin system. The construction process is as follows: Step S511: A parameterized ship maneuvering mathematical model is pre-established as the kernel. This model adopts a separable mathematical model, and its core motion equations include: Longitudinal motion equation: ; Lateral motion equations: ; Equation of the initial rocking motion: ; in, Indicates the actual mass of the ship; , These represent the longitudinal and transverse additional mass of the ship, respectively. Indicates that the ship is around Moment of inertia of the shaft; Indicates that the ship is around Additional moment of inertia of the shaft; , These represent the ship's longitudinal and lateral speeds, respectively. Indicates the ship's angular velocity of pitch; , These represent the longitudinal and lateral accelerations of the ship, respectively. Indicates the ship's roll angle acceleration; , , These represent the resultant longitudinal force, resultant lateral force, and resultant circumferential force acting on the ship, respectively. The resultant torque of the shaft; Step S512: At the start of each voyage, using the ship's loading information, calculate the ship's actual displacement and center of gravity position, and then dynamically update the equations of motion. , And hydrodynamic derivatives, to match the model with the current ship state; Step S513: Use the real-time acquired wind, wave, and current data as environmental disturbance vectors. By using wind pressure model, wave force model, and flow force model, the net external force in the equation of motion is calculated. , With resultant torque ; Step S514: Use the ship's position, heading, and speed as the initial state vector. To achieve real-time synchronization between the digital twin system and the physical ship; Step S52: Use the currently executed strategy as the control input. The constructed digital twin system performs forward rolling simulation, numerically solves the motion equations, and predicts the future. The ship's trajectory within a time period, i.e., the state sequence. ,in express Longitude of time express Time and latitude express Constant direction express time , express time , express time ; Step S53: Based on the predicted ship state sequence and real-time updated dynamic information of other ships and static obstacle information, calculate the time at any future moment. Risk indicators; Step S54: Preset safety threshold When predicting any future moment satisfy When this happens, proactive safety intervention is triggered.

[0014] Preferred proactive safety interventions include: Immediately issue visual and auditory alarms to the crew; The emergency strategy regeneration process is initiated in parallel, generating an emergency strategy with the lowest risk index according to the candidate navigation strategy generation logic. ; Preset emergency response time If the crew is If the internal controls are not taken over via the operation panel, the system will automatically execute the emergency strategy. .

[0015] The present invention also provides a shipborne intelligent navigation assistance system, comprising: The sensing information acquisition module is used to collect the sensing information of the ship; The strategy generation module is used to generate one or more candidate navigation strategies based on perception information; The strategy evaluation module is used to quantitatively evaluate candidate navigation strategies using a preset comprehensive evaluation model, and obtain an evaluation value for each strategy. The human-machine collaborative decision-making module is used to provide candidate navigation strategies and their evaluation values ​​to the crew, and to determine the final navigation strategy to be executed based on the crew's feedback. The execution and safety monitoring module is used to execute the final navigation strategy and continuously predict risks during the execution process. When potential risks are predicted, it triggers proactive safety intervention.

[0016] Therefore, the present invention employs the above-described shipborne intelligent navigation assistance method and system, and the beneficial technical effects are as follows: (1) This invention constructs a comprehensive evaluation model that includes risk indicators, efficiency indicators and rule compliance indicators, and uses an objective function for quantitative evaluation. It can generate a navigation strategy that achieves a balance between safety, economy and rule compliance, thus overcoming the limitation of the single decision-making objective of the existing system. (2) This invention calculates distraction level by collecting crew member line-of-sight information and dynamically adjusts the weight coefficients of the evaluation model accordingly, enabling the system to respond to changes in crew member status. When crew member is distracted, the system automatically increases safety weights, thereby providing more protective auxiliary strategies and achieving intelligent adaptation between the system and crew member status; (3) This invention establishes a transparent collaborative decision-making process by clearly presenting the quantitative assessment details of each strategy to the crew and providing multiple feedback channels such as selection and modification. At the same time, forward rolling simulation based on a digital twin model can predict potential risks earlier and trigger proactive intervention, thereby enhancing the system's proactive safety defense level. Attached Figure Description

[0017] Figure 1 This is a flowchart of a shipborne intelligent navigation assistance method according to the present invention; Figure 2 This is a schematic diagram of the comprehensive evaluation model; Figure 3 This is an architectural diagram of a shipborne intelligent navigation assistance system according to the present invention. Detailed Implementation

[0018] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0019] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0020] Example 1 like Figure 1 As shown, a shipborne intelligent navigation assistance method includes the following steps: Step S1: Collect the ship's sensor information.

[0021] This step involves data acquisition using a multi-source sensor system. External environment perception: The Automatic Identification System (AIS) is used to obtain dynamic data such as the identity, position, course, and speed of other ships; the navigation radar scans the surrounding waters to detect small vessels and floating objects that are not equipped with AIS; and the electronic chart system is used to obtain accurate geographic information of static obstacles such as channel boundaries, reefs, shoals, and bridges.

[0022] Ship status monitoring: The ship's position (latitude and longitude), true heading, and speed over land are obtained in real time using a high-precision GPS / INS integrated navigation system; the current displacement, draft, and load distribution status are obtained through the ship load line monitoring system.

[0023] Crew status awareness: Non-contact vision sensors (near-infrared cameras) are installed in the cockpit to track the crew’s head posture and line of sight in real time through computer vision algorithms, and to calculate the angle and duration of the shift of the crew’s gaze focus relative to the channel area ahead.

[0024] The implementation steps of computer vision algorithms are as follows: Facial and eye landmark localization: A deep learning-based object detection model (YOLOv5), trained on the WIDER FACE dataset, was used to robustly detect the facial regions of crew members from video streams. After locating the face, a dedicated facial landmark detection model (such as an HRNet-based architecture) was used to locate the two-dimensional pixel coordinates of key landmarks, including the outer and inner corners of the left and right eyes, the upper and lower eyelids, the pupil center, and the facial contours.

[0025] Head pose estimation: Matching the acquired 2D facial feature points with a general 3D face model. This is achieved by solving the perspective... The point positioning method calculates the rotation and translation matrix of the face model relative to the camera in the camera coordinate system, thereby solving the three-dimensional attitude angles of the head relative to the camera (i.e., yaw angle, pitch angle, and roll angle).

[0026] Gaze direction estimation: The direction of the eyeball is estimated by analyzing eye feature points. First, the outline of the eyeball is fitted based on the located upper and lower eyelid points. Then, the rotation angle of the eyeball is determined by calculating the relative position of the pupil center with respect to the inner and outer corners of the eye. Specifically, a coordinate system is established with the line connecting the inner and outer corners of the eye as the reference. By calculating the normalized position of the pupil center in this coordinate system and combining it with known physiological structural parameters of the eyeball (such as the eyeball radius), the two-dimensional deflection angle of the gaze is estimated. Finally, this two-dimensional deflection angle is fused with the previously calculated three-dimensional head pose to obtain the three-dimensional gaze direction of the crew member in the real-world coordinate system.

[0027] Step S2: Based on the perception information, generate one or more candidate navigation strategies.

[0028] Step S21: Calculate the nearest encounter distance (DCPA) and the time to closest point of arrival (TCPA) based on the relative positions, heading angles, and velocity vectors of the vessel and other vessels. When the DCPA is less than 2 nautical miles and the TCPA is less than 15 minutes, a collision risk is automatically identified, and specific encounter situations such as head-on encounter, cross encounter, or overtaking are classified according to the International Regulations for Preventing Collisions at Sea (ICP-100).

[0029] Step S22: Generate an initial strategy for the identified encounter situation. Initial strategies include maintaining direction and speed, turning right, or turning left.

[0030] Step S23: Generate candidate navigation strategies using the strategy perturbation formula. : ; in, This represents the heading disturbance angle, which takes values ​​within the range of [-30°, +30°]. This represents the speed disturbance, and its value is taken in the range of [-4 knots, 0 knots] (a knot is a unit of ship speed, 1 knot ≈ 1.852 km / h).

[0031] Step S3: Utilize a pre-set comprehensive evaluation model (such as...) Figure 2 As shown in the figure, the candidate navigation strategies are quantitatively evaluated to obtain the evaluation value of each strategy.

[0032] The comprehensive evaluation model is used to evaluate candidate navigation strategies. The objective function for determining superiority or inferiority is expressed as follows: ; in, Indicate candidate navigation strategies The overall assessment value; Indicate candidate navigation strategies Risk indicators; Indicate candidate navigation strategies Efficiency indicators; Indicate candidate navigation strategies Rule compliance index; , , In this embodiment, the weights of each indicator are represented. , , Set them to 0.5, 0.3, and 0.2 respectively.

[0033] The crew status information collected in step S1 is used to dynamically adjust the weight coefficients in the objective function. The specific process is as follows: Calculate the distraction index based on crew member's line of sight information. The calculation formula is: ; in, Indicates the duration of the statistics; This indicates the total time during which the crew's line of sight deviated from the preset forward channel area within the statistical period; Adjust the weighting coefficients using the following formula: ; ; in, This indicates the preset adjustment gain coefficient; This indicates the adjusted risk indicator weights; This indicates the adjusted weights of the efficiency indicators; The weight of the rule compliance index remains unchanged.

[0034] Risk indicators The calculation formula is as follows: ; in, This represents traversing each other ship and static obstacle obtained in step S1; Indicates the execution of candidate navigation strategies At that time, the ship was close to the target Predicted nearest encounter distance (to other ships or static obstacles); Indicates the execution of candidate navigation strategies At that time, the ship was close to the target The prediction is that it will encounter time soon; Indicate candidate navigation strategies The risk of grounding (the draft calculated based on the ship's condition information obtained in step S1, and the candidate navigation strategies on the electronic chart obtained in step S1). The water depth matching relationship under the route shows that when the draft is greater than the water depth... (Value increases) , , This represents the normalization coefficient.

[0035] The formula for calculating efficiency indicators is as follows: ; in, Indicates the execution of candidate navigation strategies The estimated total distance traveled at that time; Indicates the execution of candidate navigation strategies The estimated total sailing time; This indicates the total distance of the originally planned route; This indicates the estimated total time for the originally planned route; , This represents the preset weighting coefficient, in this embodiment , Set them to 0.6 and 0.4 respectively.

[0036] The formula for calculating the rule compliance index is as follows: ; in, Indicate candidate navigation strategies Recommended course; Indicate candidate navigation strategies Recommended speed; Indicates the initial policy The course; Indicates the initial policy The speed of the ship; The normalization factor representing the heading deviation is taken as the maximum reasonable turning angle during the ship's navigation (preset to 90°). The normalization factor representing the speed deviation is taken as the maximum reasonable speed change during ship navigation (preset to 10 knots). , This represents the preset weighting coefficient, in this embodiment , Set them to 0.7 and 0.3 respectively.

[0037] Step S4: Provide the candidate navigation strategies and their evaluation values ​​to the crew, and determine the final navigation strategy to be implemented based on the crew's feedback.

[0038] Step S41: Provide a graphical human-computer interaction interface (Electronic Chart Display System, ECDIS) to display all candidate navigation strategies to the crew in a list and route overlay format, and clearly indicate the comprehensive evaluation value of each strategy. Risk indicators Efficiency indicators and rule compliance indicators ; Step S42: The system listens for and receives feedback from the crew, including one or more of the following: a) Directly select a candidate strategy; b) Manually modify the heading or speed parameters of a candidate strategy; c) Ignore all candidate strategies and execute autonomous manual navigation; Step S43: Determine the final navigation strategy to be executed based on the feedback: If the feedback is a), then the selected strategy will be executed directly as the final strategy. If the feedback is b), the system will use the modified strategy as the new candidate strategy, re-execute step S3 for quantitative evaluation, and use the new strategy after evaluation as the final strategy. If the feedback is c), the system enters monitoring mode, and the final navigation strategy is the crew's real-time manual operation instructions; If within the preset time If no feedback is received within 30 seconds, the comprehensive evaluation value will be automatically applied. The optimal candidate navigation strategy.

[0039] Step S5: Execute the final navigation strategy and continuously predict risks during the execution process. When potential risks are predicted, trigger proactive safety intervention.

[0040] Continuous risk forecasting includes the following steps: Step S51: Construct a digital twin system. The construction process is as follows: Step S511: A parameterized ship maneuvering mathematical model is pre-established as the kernel. This model adopts a split mathematical model (MMG model), and its core motion equations include: Longitudinal motion equation: ; Lateral motion equations: ; Equation of the initial rocking motion: ; in, Indicates the actual mass of the ship; , These represent the longitudinal and transverse additional mass of the ship, respectively. Indicates that the ship is around Moment of inertia of the axis (the axis perpendicular to the water surface); Indicates that the ship is around Additional moment of inertia of the shaft; , These represent the ship's longitudinal and lateral speeds, respectively. Indicates the ship's angular velocity of pitch; , These represent the longitudinal and lateral accelerations of the ship, respectively. Indicates the ship's roll angle acceleration; , , These represent the resultant longitudinal force, resultant lateral force, and resultant circumferential force acting on the ship, respectively. The resultant torque of the shaft; Step S512: At the start of each voyage, using the ship's loading information, calculate the ship's actual displacement and center of gravity position, and then dynamically update the equations of motion. , And hydrodynamic derivatives, to match the model with the current ship state; Step S513: Use the real-time acquired wind, wave, and current data as environmental disturbance vectors. By using wind pressure model, wave force model, and flow force model, the net external force in the equation of motion is calculated. , With resultant torque This is existing technology and will not be described in detail here; Step S514: Use the ship's position, heading, and speed as the initial state vector. To achieve real-time synchronization between the digital twin system and the physical ship; Step S52: Use the currently executed strategy as the control input. The constructed digital twin system performs forward rolling simulation, numerically solves the motion equations, and predicts the future. The ship's trajectory within a time period (10 minutes), i.e., the state sequence. ,in express Longitude of time express Time and latitude express Constant direction express time , express time , express time ; Step S53: Based on the predicted ship state sequence and real-time updated dynamic information of other ships and static obstacle information, calculate the time at any future moment. Risk indicators; Step S54: Preset safety threshold (Based on the preset risk level of the navigation waters, such as near bridge areas) (taking smaller values) when predicting any future moment satisfy When this happens, proactive safety intervention is triggered.

[0041] Proactive safety interventions include: Immediately issue the highest level of visual and auditory alarm to the crew (such as a red flashing light and a buzzer alarm with a sound level greater than 90 decibels). The emergency strategy regeneration process is initiated in parallel, generating an emergency strategy with the lowest risk index according to the candidate navigation strategy generation logic. ; Preset emergency response time If the crew is If the internal controls are not taken over via the operation panel, the system will automatically execute the emergency strategy. .

[0042] Example 2 like Figure 3 As shown, a shipborne intelligent navigation assistance system includes: The sensing information acquisition module is used to collect the sensing information of the ship; The strategy generation module is used to generate one or more candidate navigation strategies based on perception information; The strategy evaluation module is used to quantitatively evaluate candidate navigation strategies using a preset comprehensive evaluation model, and obtain an evaluation value for each strategy. The human-machine collaborative decision-making module is used to provide candidate navigation strategies and their evaluation values ​​to the crew, and to determine the final navigation strategy to be executed based on the crew's feedback. The execution and safety monitoring module is used to execute the final navigation strategy and continuously predict risks during the execution process. When potential risks are predicted, it triggers proactive safety intervention.

[0043] It is worth noting that all contents not described in detail in this invention are existing technologies and are well known to those skilled in the art.

[0044] Therefore, the present invention adopts the above-mentioned shipborne intelligent navigation assistance method and system, which provides a more reliable and practical solution for intelligent ship navigation through multi-objective integrated decision-making, human-factor adaptive collaboration and safety monitoring.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A shipborne intelligent navigation assistance method, characterized in that, Includes the following steps: Step S1: Collect the ship's sensor information; Step S2: Based on the perception information, generate one or more candidate navigation strategies; Step S3: Using a preset comprehensive evaluation model, quantitatively evaluate the candidate navigation strategies to obtain the evaluation value of each candidate navigation strategy. Step S4: Provide the candidate navigation strategies and their evaluation values ​​to the crew, and determine the final navigation strategy based on the crew's feedback; Step S5: Execute the final navigation strategy and continuously predict risks during the execution process. When potential risks are predicted, trigger proactive safety intervention.

2. The method for shipborne intelligent navigation assistance according to claim 1, characterized in that, The perceived information in step S1 includes: External environment information: Dynamic information of other ships acquired by Automatic Identification System (AIS) and radar, including the position, course, and speed of other ships; Static obstacle information acquired by electronic charts, including information on reefs, shoals, and fixed bridges; Ship status information: position, heading, and speed of the ship obtained from the positioning and attitude determination system; load information obtained from the ship's load line system, including displacement and load status; Crew status information: information on the crew's line of sight direction obtained from in-cabin vision sensors.

3. The method for shipborne intelligent navigation assistance according to claim 1, characterized in that, Step S2 generates candidate navigation strategies based on the dynamic information of other ships and the ship's position, heading, and speed, including the following steps: Step S21: Based on the relative motion relationship between the ship and other ships, classify the situation as a head-on encounter, a cross encounter, or an overtaking encounter. Step S22: Generate an initial strategy for the encounter situation based on the International Maritime Collision Prevention Regulations. The initial strategy is to maintain direction and speed, or turn right or left. Step S23: Generate candidate navigation strategies using the strategy perturbation formula. : ； in, Indicates the heading disturbance angle; This indicates the amount of speed disturbance.

4. The shipborne intelligent navigation assistance method according to claim 3, characterized in that, The comprehensive evaluation model in step S3 is used to evaluate candidate navigation strategies. The objective function for determining superiority or inferiority is expressed as follows: ； in, Indicate candidate navigation strategies The overall assessment value; Indicate candidate navigation strategies Risk indicators; Indicate candidate navigation strategies Efficiency indicators; Indicate candidate navigation strategies Rule compliance index; , , This indicates the weight of each indicator.

5. A shipborne intelligent navigation assistance method according to claim 4, characterized in that, Risk indicators The calculation formula is as follows: ； in, This represents the traversal of each other ship and static obstacle obtained; Indicates the execution of candidate navigation strategies At that time, the ship was close to the target The prediction is that the nearest distance will be encountered; Indicates the execution of candidate navigation strategies At that time, the ship was close to the target The prediction is that it will encounter time soon; Indicate candidate navigation strategies The risk of stranding; , , Represents the normalization coefficient; The formula for calculating efficiency indicators is as follows: ； in, Indicates the execution of candidate navigation strategies The estimated total distance traveled at that time; Indicates the execution of candidate navigation strategies The estimated total sailing time; This indicates the total distance of the originally planned route; This indicates the estimated total time for the originally planned route; , This indicates the preset weighting coefficients; The formula for calculating the rule compliance index is as follows: ； in, Indicate candidate navigation strategies Recommended course; Indicate candidate navigation strategies Recommended speed; Indicates the initial policy The course; Indicates the initial policy The speed of the ship; The normalization factor representing the heading deviation; The normalization factor representing the speed deviation; , This indicates the preset weighting coefficient.

6. The method for shipborne intelligent navigation assistance according to claim 5, characterized in that, Crew status information is used to dynamically adjust the weight coefficients in the objective function. The specific process is as follows: Calculate the distraction index based on crew member's line of sight information. The calculation formula is: ； in, Indicates the duration of the statistics; This indicates the total time during which the crew's line of sight deviated from the preset forward channel area within the statistical period; Adjust the weighting coefficients using the following formula: ； ； in, This indicates the preset adjustment gain coefficient; This indicates the adjusted risk indicator weights; This indicates the adjusted weights of the efficiency indicators.

7. The method for shipborne intelligent navigation assistance according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Provide a graphical human-computer interaction interface to display all candidate navigation strategies to the crew in a list and route overlay format, and clearly indicate the comprehensive evaluation value of each strategy. Risk indicators Efficiency indicators and rule compliance indicators ; Step S42: The system listens for and receives feedback from the crew, including one or more of the following: a) Directly select a candidate navigation strategy; b) Manually modify the heading or speed parameters of a candidate navigation strategy; c) Ignore all candidate navigation strategies and execute autonomous manual navigation; Step S43: Determine the final navigation strategy to be executed based on the feedback: If the feedback is a), then the selected strategy will be executed directly as the final strategy. If the feedback is b), the system will use the modified strategy as the new candidate strategy, re-execute step S3 for quantitative evaluation, and use the new strategy after evaluation as the final strategy. If the feedback is c), the system enters monitoring mode, and the final navigation strategy is the crew's real-time manual operation instructions; If within the preset time If no feedback is received, the comprehensive evaluation value will be automatically applied. The optimal candidate navigation strategy.

8. The method for shipborne intelligent navigation assistance according to claim 1, characterized in that, The continuous risk prediction in step S5 includes the following steps: Step S51: Construct a digital twin system. The construction process is as follows: Step S511: A parameterized ship maneuvering mathematical model is pre-established as the kernel. This model adopts a separable mathematical model, and its core motion equations include: Longitudinal motion equation: ； Lateral motion equations: ； Equation of the initial rocking motion: ； in, Indicates the actual mass of the ship; , These represent the longitudinal and transverse additional mass of the ship, respectively. Indicates that the ship is around Moment of inertia of the shaft; Indicates that the ship is around Additional moment of inertia of the shaft; , These represent the ship's longitudinal and lateral speeds, respectively. Indicates the ship's angular velocity of pitch; , These represent the longitudinal and lateral accelerations of the ship, respectively. Indicates the ship's roll angle acceleration; , , These represent the resultant longitudinal force, resultant lateral force, and resultant circumferential force acting on the ship, respectively. The resultant torque of the shaft; Step S512: At the start of each voyage, using the ship's loading information, calculate the ship's actual displacement and center of gravity position, and then dynamically update the equations of motion. , And hydrodynamic derivatives, to match the model with the current ship state; Step S513: Use the real-time acquired wind, wave, and current data as environmental disturbance vectors. Calculate the net external force in the equation of motion. , With resultant torque ; Step S514: Use the ship's position, heading, and speed as the initial state vector. To achieve real-time synchronization between the digital twin system and the physical ship; Step S52: Use the currently executed strategy as the control input. The constructed digital twin system performs forward rolling simulation, numerically solves the motion equations, and predicts the future. The ship's trajectory within a time period, i.e., the state sequence. ,in express Longitude of time express Time and latitude express Constant direction express time , express time , express time ; Step S53: Based on the predicted ship state sequence and real-time updated dynamic information of other ships and static obstacle information, calculate the time at any future moment. Risk indicators; Step S54: Preset safety threshold When predicting any future moment satisfy When this happens, proactive safety intervention is triggered.

9. A shipborne intelligent navigation assistance method according to claim 1, characterized in that, Proactive safety interventions include: Immediately issue visual and auditory alarms to the crew; The emergency strategy regeneration process is initiated in parallel, generating an emergency strategy with the lowest risk index according to the candidate navigation strategy generation logic. ; Preset emergency response time If the crew is If the internal controls are not taken over via the operation panel, the system will automatically execute the emergency strategy. .

10. A shipborne intelligent navigation assistance system, characterized in that, include: The sensing information acquisition module is used to collect the sensing information of the ship; The strategy generation module is used to generate one or more candidate navigation strategies based on perception information; The strategy evaluation module is used to quantitatively evaluate candidate navigation strategies using a preset comprehensive evaluation model, and obtain an evaluation value for each strategy. The human-machine collaborative decision-making module is used to provide candidate navigation strategies and their evaluation values ​​to the crew, and to determine the final navigation strategy to be executed based on the crew's feedback. The execution and safety monitoring module is used to execute the final navigation strategy and continuously predict risks during the execution process. When potential risks are predicted, it triggers proactive safety intervention.

Images