Marine traffic risk assessment method

Abstract

The invention discloses a marine traffic risk assessment method, which belongs to the field of marine ship risk analysis, and comprises the following steps: S1, collecting and analyzing AIS data and environmental data of a ship and neighbor ships, and supplementing packet loss data; s2, correcting and calculating the dynamic track through an environment interference separation model; s3, screening neighbor ships, constructing a multi-ship interaction model, and quantifying the risk influence degree of the neighbor ships on the ship; s4, respectively constructing a single-ship comprehensive collision risk assessment model and a multi-ship comprehensive collision risk assessment model to obtain a single-ship comprehensive collision risk value and a multi-ship comprehensive collision risk value; and S5, the risk level is judged, and a global-local linkage risk judgment model is formed. By the adoption of the marine traffic risk assessment method, the problems of data deviation, risk false judgment and missed judgment and multi-ship effect neglecting of a traditional method are solved, model parameters can be dynamically optimized to adapt to complex scenes, and finally the accuracy of ship collision risk assessment and the safety and the performability of collision avoidance decision making are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of marine vessel risk analysis technology, and in particular to a method for assessing marine traffic risks. Background Technology

[0002] In existing technologies, maritime traffic risk assessment mainly relies on traditional models centered on Automatic Identification System (AIS), Distance to Closest Encounter (DCPA), and Time to Closest Encounter (TCPA). These models use linear motion assumptions and uniform ship parameters for trajectory prediction. However, since factors influencing maritime traffic risk assessment include, but are not limited to, ship-related factors, environmental factors, traffic flow factors, and the coupling effects of multiple factors, existing maritime traffic risk assessment methods suffer from the problem of fragmented analysis of multi-dimensional risk factors such as ships and the environment. They also have significant defects such as frequent risk misjudgments, delayed warnings, and risk omissions. This indicates that existing traditional methods are no longer sufficient to meet the collision risk assessment needs under complex traffic patterns. Summary of the Invention

[0003] The purpose of this invention is to provide a method for assessing maritime traffic risks and to solve the aforementioned technical problems.

[0004] To achieve the above objectives, the present invention provides a method for assessing maritime traffic risks, comprising the following steps: S1. Collect and analyze the static, dynamic and maneuvering messages of the ship and neighboring ships, supplement environmental data through meteorological sensors, and calculate and supplement the lost data by combining the effective data before and after the lost data period when AIS data is lost for a short time, so as to provide continuous basic data input for the subsequent trajectory prediction module. S2. Based on the AIS message data and environmental data from step S1, the active speed of the ship and environmental interference are separated by the environmental interference separation model, and the dynamic trajectory correction calculation is performed to obtain the active trajectory, which provides ship trajectory data reflecting the actual motion for the subsequent multi-ship interaction modeling module. S3. Filter neighboring vessels through AIS, construct a multi-vessel interaction model based on the vessel trajectory data in step S2, and combine the basic vessel data provided in step S1 to calculate the edge weights containing vessel type weights in order to quantify the risk impact of neighboring vessels on the vessel and provide data support for vessel collision risk warning or collision avoidance decision-making. S4. Based on the continuous data supplemented in step S1, the accurate trajectory corrected in S2, and the multi-ship interaction model constructed in S3, construct comprehensive collision risk assessment models for single ships and multiple ships respectively, and obtain comprehensive collision risk values ​​for single ships and multiple ships to promote the connection from data processing to risk warning. S5. Based on the correlation between the single-ship comprehensive collision risk value and the multi-ship comprehensive collision risk value in step S4, determine the risk level. With the overall risk of multiple ships as the basic framework and the risk of a single ship as a supplement to local anomalies, form a global-local linkage risk judgment model and output the risk level.

[0005] Preferably, in step S1, data samples of nearby vessels are collected through AIS. These data samples include static message data consisting of the vessel's basic dimensions, draft, and type; dynamic message data consisting of the vessel's real-time changing coordinates, heading, and speed; and maneuver message data consisting of the vessel's real-time rudder angle and propeller speed.

[0006] Preferably, the environmental dynamic data includes at least the current wind speed, wind direction, water flow speed, and water flow direction of the water area; When supplementing environmental interference data, it is also necessary to achieve time alignment with AIS data. That is, extract the timestamp corresponding to each message from the AIS dynamic messages, and use the message timestamp as a reference to perform time interpolation processing on the collected environmental dynamic data, so that the time nodes of the environmental dynamic data are completely matched with the time nodes of the AIS dynamic data, so as to ensure that the ship status data and environmental interference data are in the same time dimension.

[0007] Preferably, for short-term packet loss based on AIS data in step S1, including packet loss in linear motion segments and packet loss in curved motion segments, it is necessary to first distinguish between straight-line and turning navigation by the change in heading during the packet loss period. The determination method is as follows: When a packet is lost, the heading angle of the preceding valid point at the time of packet loss is: The heading angle of the subsequent valid points is The difference in heading angle between the two points is And set a critical value for the difference in heading angle that distinguishes between straight lines and turns. , If so, it is determined to be a straight-line journey, or If so, it is determined to be a turning maneuver; For straight-line navigation scenarios, based on the ship's uniform linear motion, the position and speed at the time of packet loss are estimated according to the time ratio of valid data before and after packet loss. For turning navigation scenarios, based on the characteristic of the heading changing continuously over time, the heading at the time of packet loss is first estimated, and then the speed is decomposed into northward and eastward components. Combined with time, the position at the time of packet loss is calculated to complete the AIS short-term packet loss data.

[0008] Preferably, in step S1, the navigation scenario is determined based on the data in the AIS system before and after packet loss, and the packet loss data is calculated accordingly. When the ship is determined to be moving at a constant velocity in a straight line, data packet loss is compensated for by linear interpolation. The formula for calculating the position data completion is as follows: ; in, , The coordinates of the packet loss point during straight-line navigation; , The coordinates of the preceding valid points; , For the longitude and latitude of the subsequent valid points; The time percentage coefficient is calculated using the following formula: ; in, For packet loss time in straight-line navigation scenarios; For the preceding valid point time; For subsequent valid point times; The formula for calculating speed data completion is: ; in, The ground speed when packet loss occurs during straight-line navigation; The ground velocity of the preceding valid point; This represents the ground velocity of the subsequent valid points; When the scenario is determined to be a turning navigation scenario, the ship's course changes. Segmented course-position interpolation is used to complete the data packet loss. The course data completion calculation formula is as follows: ; in, The heading angle when packet loss occurs during a turning navigation scenario; For the rate of change of heading, ; For packet loss time in a turning navigation scenario; The formula for calculating speed data completion is: ; in, The ground speed when the packet is lost during a turn; decomposition Its eastward velocity component was obtained. Northbound velocity component ; The location at the moment of packet loss is calculated based on the average velocity components before and after packet loss. The formula for calculating the location data is as follows: ; in, , The coordinates of the point where the packet was lost during the turning navigation; , The coordinates of the preceding valid points; The average northbound speed before and after the packet loss during the turn; The average eastward speed before and after the packet loss during the turn.

[0009] Preferably, step S2 specifically includes the following steps: S21. Based on the AIS dynamic data and environmental data throughout the entire time period in step S1, construct an environmental interference separation model to separate environmental interference from the ship's actual speed and obtain the ship's active speed. The formula is as follows: ; in, , The active velocity consists of the eastward and northward components; , These represent the eastward and northward components of the ship's speed. , These represent the eastward and northward components of the wind speed. , These represent the eastward and northward components of the water flow velocity; Ship active speed for: ; Active heading of the ship for: .

[0010] S22. Based on the track direction corrected in step S21, i.e., the active heading. , compared with the original ground heading The difference is the heading correction angle. This reflects the degree of environmental interference with the flight path direction. The original flight path was deviated westward due to environmental interference and needs to be corrected eastward; or The original flight path was deviated eastward due to environmental interference and needed to be corrected westward. The formula for the corrected position coordinates of a ship's trajectory is: ; in, , The corrected ship coordinates; , The ship's coordinates before correction; This refers to the update cycle of AIS data.

[0011] Preferably, step S3 specifically includes the following steps: S31. Based on AIS basic data and the corrected position coordinates of the ship's trajectory in step S22. Filter neighboring ships by distance The calculation formula is: ; in, This refers to the distance between this vessel and nearby vessels; , The coordinates of nearby vessels are set; the distance threshold for judging neighboring vessels is set to... , In this case, the nearby vessel is considered the vessel's neighbor. S32. Based on the neighbor ship filtering in step S31, output the neighbor ship set. Each element contains the neighboring ship's code, ship type, and real-time dynamic parameters; S33. Based on the ship's own hull and neighboring ships, construct a multi-ship interaction model and calculate the edge weights including ship type weights. The formula is: ; in, , This is an empirical coefficient; The weighting for ship type should be assigned according to the risk level corresponding to the ship type. The basic risk factor is calculated using the following formula: ; in, As a distance factor, , The attenuation coefficient is... ; The relative velocity factor, ; , The velocity component of the neighboring ship; , is the maximum relative speed of the ship; For heading cross factor, based on relative heading Assess the encounter situation and assign a risk coefficient. , For the neighboring ship's course; S34, Edge weights based on step S33 It is then normalized to achieve a unified measurement of multi-dimensional risk. The normalization formula is as follows: ; in, The initial edge weights, ; for Normalization processing; The maximum value of the initial weights, This represents the minimum value of the initial weights.

[0012] Preferably, step S4 specifically includes the following steps: S41. Based on the formula in step S2, obtain the track data of the ship and neighboring ships. Combining the principles of ship kinematics, calculate the real-time minimum encounter distance between the ship and each neighboring ship. and the time of arrival closest point , The calculation formula is: ; in, , These are the relative position vector and relative velocity vector of the ship and the neighboring ship, respectively. The calculation formula is: ; S42, Edge weights based on step S3 And based on step S41 , A single-ship comprehensive risk model and a multi-ship comprehensive risk model are constructed for each neighboring vessel relative to the vessel. The calculation process for the single-ship comprehensive risk model is as follows: Piecewise quantization Risk coefficient, formula is: ; in, for Risk factor; This is the preset minimum safe meeting distance; ,and hour, This refers to a collision between this vessel and a neighboring vessel. The distance between our vessel and neighboring vessels is safe at this time; Quantification Risk coefficient, formula is: ; in, For the risk factor of sailing time; The preset safe arrival time; ,and hour, This refers to a collision between this vessel and a neighboring vessel. The distance between our vessel and neighboring vessels is safe at this time; The formula for the comprehensive risk value of a single ship is: ; in, This is the comprehensive risk value for a single vessel, and , A large value indicates a high overall risk of collision with neighboring vessels. for The weighting coefficient is related to the waters where the ship is located; The calculation process for the multi-ship integrated risk model is as follows: A weighted model employing maximum risk as the primary factor and average risk as the secondary factor is adopted, and the formula is as follows: ; in, To mitigate the combined risks of multiple vessels, and , A large value indicates a high overall risk of collision with neighboring vessels. The maximum risk weighting coefficient highlights the dominant role of high-risk vessels in overall risk. for The highest single-ship risk value among neighboring vessels compared to this vessel.

[0013] Preferably, step S5 specifically includes the following steps: S51. Based on the comprehensive risk value of a single vessel in step S42, pre-screen the risk value of a single vessel and determine the risk level of a single vessel, and form a risk list of a single vessel to provide a basis for subsequent overall risk assessment. S52, Multi-ship integrated risk in step S4 As the initial level, combined with the single-ship risk list in step S51, the level is adjusted according to the principle of prioritizing local threats to determine the overall risk level.

[0014] Therefore, the beneficial effects of adopting the above-mentioned maritime traffic risk assessment method in this invention are as follows: 1. By supplementing short-term AIS packet loss data and combining it with environmental data from meteorological sensors, the system ensures continuous and complete basic data, completely eliminating misjudgments of motion status caused by data breaks. This avoids directional deviations in application-layer analysis and provides a "breakpoint-free" data source for all modules, including subsequent track correction, multi-ship interactive modeling, and risk assessment. This is the fundamental guarantee for the accurate operation of the entire system. It specifically addresses the problem that "AIS packet loss directly leads to incomplete ship trajectories and misjudgments of motion status, which in turn causes deviations in application-layer functions such as navigation monitoring and collision warning."

[0015] 2. By separating active navigation from environmental interference, the active trajectory is accurately corrected. This not only eliminates environmental interference and makes the TCPA calculation more accurate, but also truly reflects the ship's motion intention in complex operations, avoiding the omission of dynamic interactions between ships due to trajectory deviation. It also provides accurate trajectory data for multi-ship interaction modeling, which is a key prerequisite for subsequent quantification of multi-ship risks and capture of uncertainties in complex scenarios.

[0016] 3. By introducing ship type weights, the impact of maneuverability differences on risk is quantified, avoiding the one-sidedness of a unified model; at the same time, a multi-ship interaction model (not mutually independent) is constructed, which can accurately capture the coupling risks brought about by third-party intrusion, eliminating the impact of track deviation on risk assessment and avoiding the omission of risks from multiple ships; by transforming the characteristics of a single ship and the interaction relationship between multiple ships into quantifiable risk data, a differentiated and comprehensive core basis is provided for collision risk warning and collision avoidance decision-making.

[0017] 4. By constructing a multi-ship comprehensive assessment model, instead of being limited to pairwise analysis, it can fully cover the dynamic interactions between multiple ships; at the same time, it retains the single-ship assessment dimension, which can focus on high-risk ship pairs. This not only solves the uncertainty of complex scenarios, but also avoids the one-sidedness of multi-ship assessment, transforming scattered data sources into risk values ​​of single-ship + multi-ship dual dimensions, making risk assessment more systematic.

[0018] 5. Transforming multi-dimensional risk data into precise risk levels provides a final, controllable basis for navigation monitoring and collision avoidance command, directly improving the safety of navigation in complex waters.

[0019] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0020] Figure 1 This is a flowchart of the maritime traffic risk assessment method of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the embodiments of the present invention and are not intended to limit the embodiments of the present invention. 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 this application. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout.

[0022] It should be noted that the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, such as a process, method, system, product, or server that includes a series of steps or units, not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product, or device.

[0023] The following is in conjunction with the appendix Figure 1 The embodiments of the present invention will be described in detail below.

[0024] like Figure 1 As shown, the maritime traffic risk assessment method includes the following steps: S1. Collect and analyze the static, dynamic and maneuvering messages of the ship and neighboring ships, supplement environmental data through meteorological sensors, and calculate and supplement the lost data by combining the effective data before and after the lost data period when AIS data is lost for a short time, so as to provide continuous basic data input for the subsequent trajectory prediction module. In step S1, data samples of nearby vessels are collected via AIS. These data samples include static message data consisting of the vessel's basic dimensions, draft, and type; dynamic message data consisting of the vessel's real-time changing coordinates, heading, and speed; and maneuver message data consisting of the vessel's real-time rudder angle and propeller speed.

[0025] Environmental dynamic data should include at least the current wind speed, wind direction, water flow speed, and water flow direction in the water area; When supplementing environmental interference data, it is also necessary to achieve time alignment with AIS data. That is, extract the timestamp corresponding to each message from the AIS dynamic messages, and use the message timestamp as a reference to perform time interpolation processing on the collected environmental dynamic data so that the time nodes of the environmental dynamic data match the time nodes of the AIS dynamic data, so as to ensure that the ship status data and environmental interference data are in the same time dimension.

[0026] Based on short-term packet loss of AIS data in step S1, including packet loss in linear motion segments and packet loss in curved motion segments, it is necessary to first distinguish between straight-line and turning navigation by the change in heading during the packet loss period. The determination method is as follows: Let the heading angle of the preceding valid point be... The heading angle of the subsequent valid points is The difference in heading angle between the two points is And set a critical value for the difference in heading angle that distinguishes between straight lines and turns. , If so, it is determined to be a straight-line journey, or If so, it is determined to be a turning maneuver; For straight-line navigation scenarios, based on the ship's uniform linear motion, the position and speed at the time of packet loss are estimated according to the time ratio of valid data before and after packet loss. For turning navigation scenarios, based on the characteristic of the heading changing continuously over time, the heading at the time of packet loss is first estimated, and then the speed is decomposed into northward and eastward components. Combined with time, the position at the time of packet loss is calculated to complete the AIS short-term packet loss data.

[0027] In step S1, based on the data in the AIS system before and after packet loss, the navigation scenario is determined, and the packet loss data is calculated accordingly. When the ship is determined to be moving at a constant velocity in a straight line, data packet loss is compensated for by linear interpolation. The formula for calculating the position data completion is as follows: ; in, , The coordinates of the packet loss point during straight-line navigation; , The coordinates of the preceding valid points; , For the longitude and latitude of the subsequent valid points; The time percentage coefficient is calculated using the following formula: ; in, For packet loss time in straight-line navigation scenarios; For the preceding valid point time; For subsequent valid point times; The formula for calculating speed data completion is: ; in, The ground speed when packet loss occurs during straight-line navigation; The ground velocity of the preceding valid point; This represents the ground velocity of the subsequent valid points; When the scenario is determined to be a turning navigation scenario, the ship's course changes. Segmented course-position interpolation is used to complete the data packet loss. The course data completion calculation formula is as follows: ; in, The heading angle when packet loss occurs during a turning navigation scenario; For the rate of change of heading, ; For packet loss time in a turning navigation scenario; The formula for calculating speed data completion is: ; in, The ground speed when the packet is lost during a turn; decomposition Its eastward velocity component was obtained. Northbound velocity component ; The location at the moment of packet loss is calculated based on the average velocity components before and after packet loss. The formula for calculating the location data is as follows: ; in, , The coordinates of the point where the packet was lost during the turning navigation; , The coordinates of the preceding valid points; The average northbound speed before and after the packet loss during the turn; The average eastward speed before and after the packet loss during the turn.

[0028] S2. Based on the AIS message data and environmental data from step S1, the active speed of the ship and environmental interference are separated by the environmental interference separation model, and the dynamic trajectory correction calculation is performed to obtain the active trajectory, which provides ship trajectory data reflecting the actual motion for the subsequent multi-ship interaction modeling module. Step S2 specifically includes the following steps: S21. Based on the AIS dynamic data and environmental data throughout the entire time period in step S1, construct an environmental interference separation model to separate environmental interference from the ship's actual speed and obtain the ship's active speed. The formula is as follows: ; in, , The active velocity consists of the eastward and northward components; , These represent the eastward and northward components of the ship's speed. , These represent the eastward and northward components of the wind speed. , These represent the eastward and northward components of the water flow velocity; Ship active speed for: ; Active heading of the ship for: .

[0029] S22. Based on the track direction corrected in step S21, i.e., the active heading. , compared with the original ground heading The difference is the heading correction angle. This reflects the degree of environmental interference with the flight path direction. The original flight path was deviated westward due to environmental interference and needs to be corrected eastward; or The original flight path was deviated eastward due to environmental interference and needed to be corrected westward. The formula for the corrected position coordinates of a ship's trajectory is: ; in, , The corrected ship coordinates; , The ship's coordinates before correction; This refers to the update cycle of AIS data.

[0030] After completing the missing data, to ensure the accuracy of the basic data for track prediction, the accuracy of the AIS missing data completion data can be verified. The verification process uses the corrected track data output in step S2 as the verification benchmark. By comparing the simulated motion generated by the completed data with the actual motion reflected by the accurate track, the deviation between the two is quantified. The completed data with deviations exceeding the threshold is corrected a second time, and finally, continuous AIS data consistent with the actual motion of the ship is output. The specific process is as follows: Perform multi-dimensional deviation calculations, where the positional deviation is calculated as follows: ; in, This is for positional deviation; The average positional deviation is ; Maximum positional deviation is ; Speed ​​deviation is ; Average speed deviation is ; The heading deviation is ; The average heading deviation is ; Referencing the preset deviation threshold standards in Table 1, formulate the judgment logic: Table 1 Preset Deviation Threshold Standards

[0031] S3. Filter neighboring vessels through AIS, construct a multi-vessel interaction model based on the vessel trajectory data in step S2, and combine the basic vessel data provided in step S1 to calculate the edge weights containing vessel type weights in order to quantify the risk impact of neighboring vessels on the vessel and provide data support for vessel collision risk warning or collision avoidance decision-making. Step S3 specifically includes the following steps: S31. Based on AIS basic data and the corrected position coordinates of the ship's trajectory in step S22. Filter neighboring ships by distance The calculation formula is: ; in, This refers to the distance between this vessel and nearby vessels; , The coordinates of nearby vessels are set; the distance threshold for judging neighboring vessels is set to... , In this case, the nearby vessel is considered the vessel's neighbor. S32. Based on the neighbor ship filtering in step S31, output the neighbor ship set. Each element contains the neighboring ship's code, ship type, and real-time dynamic parameters; S33. Based on the ship's own hull and neighboring ships, construct a multi-ship interaction model and calculate the edge weights including ship type weights. The formula is: ; in, , This is an empirical coefficient; The weighting for ship type should be assigned according to the risk level corresponding to the ship type. The basic risk factor is calculated using the following formula: ; in, As a distance factor, , The attenuation coefficient is... ; The relative velocity factor, ; , The velocity component of the neighboring ship; , is the maximum relative speed of the ship; For heading cross factor, based on relative heading Assess the encounter situation and assign a risk coefficient. , For the neighboring ship's course; S34, Edge weights based on step S33 It is then normalized to achieve a unified measurement of multi-dimensional risk. The normalization formula is as follows: ; in, The initial edge weights, ; for Normalization processing; The maximum value of the initial weights, This represents the minimum value of the initial weights.

[0032] S4. Based on the continuous data supplemented in step S1, the accurate trajectory corrected in S2, and the multi-ship interaction model constructed in S3, construct comprehensive collision risk assessment models for single ships and multiple ships respectively, and obtain comprehensive collision risk values ​​for single ships and multiple ships to promote the connection from data processing to risk warning. Step S4 specifically includes the following steps: S41. Based on the track data of the ship and neighboring ships obtained in step S2, and combining the principles of ship kinematics, calculate the real-time minimum encounter distance between the ship and each neighboring ship. and the time of arrival closest point , The calculation formula is: ; in, , These are the relative position vector and relative velocity vector of the ship and the neighboring ship, respectively. The calculation formula is: ; S42, Edge weights based on step S3 And based on step S41 , A single-ship comprehensive risk model and a multi-ship comprehensive risk model are constructed for each neighboring vessel relative to the vessel. The calculation process for the single-ship comprehensive risk model is as follows: Piecewise quantization Risk coefficient, formula is: ; in, for Risk factor; This is the preset minimum safe meeting distance; ,and hour, This refers to a collision between this vessel and a neighboring vessel. The distance between our vessel and neighboring vessels is safe at this time; Quantification Risk coefficient, formula is: ; in, For the risk factor of sailing time; The preset safe arrival time; ,and hour, This refers to a collision between this vessel and a neighboring vessel. The distance between our vessel and neighboring vessels is safe at this time; The formula for the comprehensive risk value of a single ship is: ; in, This is the comprehensive risk value for a single vessel, and , A large value indicates a high overall risk of collision with neighboring vessels. for The weighting coefficient is related to the waters where the ship is located; The calculation process for the multi-ship integrated risk model is as follows: A weighted model employing maximum risk as the primary factor and average risk as the secondary factor is adopted, and the formula is as follows: ; in, To mitigate the combined risks of multiple vessels, and , A large value indicates a high overall risk of collision with neighboring vessels. The maximum risk weighting coefficient highlights the dominant role of high-risk vessels in overall risk. for The highest single-ship risk value among neighboring vessels compared to this vessel.

[0033] S5. Based on the correlation between the single-ship comprehensive collision risk value and the multi-ship comprehensive collision risk value in step S4, determine the risk level. With the overall risk of multiple ships as the basic framework and the risk of a single ship as a supplement to local anomalies, form a global-local linkage risk judgment model and output the risk level. Step S5 specifically includes the following steps: S51. Pre-screen individual vessel risk values ​​and determine individual vessel risk levels, and form an individual vessel risk list to provide a basis for subsequent overall risk assessment; S52, Multi-ship integrated risk in step S4 As the initial level, combined with the single-ship risk list in step S51, the level is adjusted according to the principle of prioritizing local threats to determine the overall risk level.

[0034] Referring to Table 2, based on the risk level determined in step S5, risk levels and core response strategies are categorized as follows: Table 2 Risk Level Table

[0035] Based on the risk level table in step S5, an initial collision avoidance decision that adapts to the ship's maneuverability and complies with collision avoidance rules is generated. Then, by simulating the multi-ship trajectories after the decision is executed, the risk of new collisions is predicted. Finally, an optimized decision with no secondary collision risk and which can be directly executed is output. In the multi-ship model, the risk proportion of each neighboring ship is calculated using the following formula: ; Here, L represents the risk contribution percentage of neighboring vessels. The larger the value, the greater the threat to the vessel. By calculating the risk contribution percentage of each neighboring vessel separately, the navigator can clearly focus on the neighboring vessel with the greatest risk, thereby reducing the risk of collision.

[0036] Experimental Example: Experimental Example of a Full-Process Model for Ship Collision Risk Assessment Experimental objectives: To verify the performance of the full-process dynamic optimization model (FDO model) in terms of AIS packet loss completion accuracy, environmental interference separation, multi-ship coupling risk assessment, and collision avoidance decision effectiveness; to compare with traditional collision risk assessment models and quantify the advantages of the FDO model in terms of risk misjudgment rate, missed judgment rate, and cascading risk avoidance rate; and to verify the adaptability of the FDO model to different types of ships and complex sea areas.

[0037] Data source: Selected sea area near the port, which is a high-traffic area containing merchant ships, fishing boats, and port operation vessels, with complex traffic patterns, including AIS data and environmental data; Comparison Model: Traditional Basic Model (TB Model): Calculates DCPA / TCPA based solely on raw AIS data, without handling packet loss, separating environmental interference, or considering differences in ship maneuverability; Traditional pairwise assessment model (TP model): Based on the logic of TB model, it calculates risk independently for each pair of ships in a multi-ship scenario, without considering the coupling effect of multiple ships or predicting cascading risks.

[0038] The experimental results are shown in Table 3-4: Table 3 Comparison of Data Accuracy Deviation

[0039] Table 4 Comparison of Risk Assessment Accuracy Indicators

[0040] In summary, the FDO model has a packet loss completion bias of 0.04 nautical miles, which is only 1 / 3 to 1 / 4 of that of the traditional model, and the trajectory prediction bias is reduced by more than 60%, verifying the effectiveness of S1.1 packet loss verification and S2 environmental interference separation. The FDO model has a risk misjudgment rate of 5.2%, which is much lower than that of the traditional model, and a false negative rate of 0. It can accurately capture high-risk events and solve the problems of "maneuverability neglect and false negatives due to multi-ship coupling" in the traditional method. Moreover, the FDO model can also generate collision avoidance decisions without cascading risks, improve the collision avoidance success rate, and make it suitable for complex sea areas and different ship types.

[0041] 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 method for assessing maritime traffic risks, characterized in that: Includes the following steps: S1. Collect and analyze the static, dynamic and maneuvering messages of the ship and neighboring ships, supplement environmental data through meteorological sensors, and calculate and supplement the lost data by combining the effective data before and after the lost data period when AIS data is lost for a short time, so as to provide continuous basic data input for the subsequent trajectory prediction module. S2. Based on the AIS message data and environmental data from step S1, the active speed of the ship and environmental interference are separated by the environmental interference separation model, and the dynamic trajectory correction calculation is performed to obtain the active trajectory, which provides ship trajectory data reflecting the actual motion for the subsequent multi-ship interaction modeling module. S3. Filter neighboring vessels through AIS, construct a multi-vessel interaction model based on the vessel trajectory data in step S2, and combine the basic vessel data provided in step S1 to calculate the edge weights containing vessel type weights in order to quantify the risk impact of neighboring vessels on the vessel and provide data support for vessel collision risk warning or collision avoidance decision-making. S4. Based on the continuous data supplemented in step S1, the track corrected in S2, and the multi-ship interaction model constructed in S3, construct comprehensive collision risk assessment models for single ships and multiple ships respectively, and obtain comprehensive collision risk values ​​for single ships and multiple ships to promote the connection from data processing to risk warning. S5. Based on the correlation between the single-ship comprehensive collision risk value and the multi-ship comprehensive collision risk value in step S4, determine the risk level. With the overall risk of multiple ships as the basic framework and the risk of a single ship as a supplement to local anomalies, form a global-local linkage risk judgment model and output the risk level.

2. The maritime traffic risk assessment method according to claim 1, characterized in that: In step S1, data samples of nearby vessels are collected via AIS. These data samples include static message data consisting of the vessel's basic dimensions, draft, and type; dynamic message data consisting of the vessel's real-time changing coordinates, heading, and speed; and maneuver message data consisting of the vessel's real-time rudder angle and propeller speed.

3. The maritime traffic risk assessment method according to claim 2, characterized in that: Environmental dynamic data should include at least the current wind speed, wind direction, water flow speed, and water flow direction in the water area; When supplementing environmental interference data, it is also necessary to achieve time alignment with AIS data. That is, extract the timestamp corresponding to each message from the AIS dynamic messages, and use the message timestamp as a reference to perform time interpolation processing on the collected environmental dynamic data so that the time nodes of the environmental dynamic data match the time nodes of the AIS dynamic data, so as to ensure that the ship status data and environmental interference data are in the same time dimension.

4. The maritime traffic risk assessment method according to claim 3, characterized in that: Based on short-term packet loss of AIS data in step S1, including packet loss in linear motion segments and packet loss in curved motion segments, it is necessary to first distinguish between straight-line and turning navigation by the change in heading during the packet loss period. The determination method is as follows: When a packet is lost, the heading angle of the preceding valid point at the time of packet loss is: The heading angle of the subsequent valid points is The difference in heading angle between the two points is And set a critical value for the difference in heading angle that distinguishes between straight lines and turns. , If so, it is determined to be a straight-line journey, or If so, it is determined to be a turning maneuver; For straight-line navigation scenarios, based on the ship's uniform linear motion, the position and speed at the time of packet loss are estimated according to the time ratio of valid data before and after packet loss. For turning navigation scenarios, based on the characteristic of the heading changing continuously over time, the heading at the time of packet loss is first estimated, and then the speed is decomposed into northward and eastward components. Combined with time, the position at the time of packet loss is calculated to complete the AIS short-term packet loss data.

5. The maritime traffic risk assessment method according to claim 4, characterized in that: In step S1, based on the data in the AIS system before and after packet loss, the navigation scenario is determined, and the packet loss data is calculated accordingly. When the ship is determined to be moving at a constant velocity in a straight line, data packet loss is compensated for by linear interpolation. The formula for calculating the position data completion is as follows: ； in, , The coordinates of the packet loss point during straight-line navigation; , The coordinates of the preceding valid points; , For the longitude and latitude of the subsequent valid points; The time percentage coefficient is calculated using the following formula: ； in, For packet loss time in straight-line navigation scenarios; For the preceding valid point time; For subsequent valid point times; The formula for calculating speed data completion is: ； in, The ground speed when packet loss occurs during straight-line navigation; The ground velocity of the preceding valid point; This represents the ground velocity of the subsequent valid points; When the scenario is determined to be a turning navigation scenario, the ship's course changes. Segmented course-position interpolation is used to complete the data packet loss. The course data completion calculation formula is as follows: ； in, The heading angle when packet loss occurs during a turning navigation scenario; For the rate of change of heading, ; For packet loss time in turning navigation scenarios; The formula for calculating speed data completion is: ； in, The ground speed when the packet is lost during a turn; decomposition Its eastward velocity component was obtained. Northbound velocity component ; The location at the moment of packet loss is calculated based on the average velocity components before and after packet loss. The formula for calculating the location data is as follows: ； in, , The coordinates of the point where the packet was lost during the turning navigation; , The coordinates of the preceding valid points; The average northbound speed before and after the packet loss during the turn; The average eastward speed before and after the packet loss during the turn.

6. The maritime traffic risk assessment method according to claim 5, characterized in that: Step S2 specifically includes the following steps: S21. Based on the AIS dynamic data and environmental data throughout the entire time period in step S1, construct an environmental interference separation model to separate environmental interference from the ship's actual speed and obtain the ship's active speed. The formula is as follows: ； in, , The active velocity consists of the eastward and northward components; , These represent the eastward and northward components of the ship's speed. , These represent the eastward and northward components of the wind speed; , These represent the eastward and northward components of the water flow velocity; Ship active speed for: ； Active heading of the ship for: ； S22. Based on the track direction corrected in step S21, i.e., the active heading. , compared with the original ground heading The difference is the heading correction angle. This reflects the degree of environmental interference with the flight path direction. The original flight path was deviated westward due to environmental interference and needs to be corrected eastward; or The original flight path was deviated eastward due to environmental interference and needed to be corrected westward. The formula for the corrected position coordinates of a ship's trajectory is: ； in, , The corrected ship coordinates; , The ship's coordinates before correction; This refers to the update cycle of AIS data.

7. The maritime traffic risk assessment method according to claim 6, characterized in that: Step S3 specifically includes the following steps: S31. Based on AIS basic data and the corrected position coordinates of the ship's trajectory in step S22. Filter neighboring ships by distance The calculation formula is: ； in, This refers to the distance between this vessel and nearby vessels; , The coordinates of nearby vessels are set; the distance threshold for judging neighboring vessels is set to... , In this case, the nearby vessel is considered the neighboring vessel of this vessel; S32. Based on the neighbor ship filtering in step S31, output the neighbor ship set. Each element contains the neighboring ship's code, ship type, and real-time dynamic parameters; S33. Based on the ship's own hull and neighboring ships, construct a multi-ship interaction model and calculate the edge weights including ship type weights. The formula is: ； in, , This is an empirical coefficient; The weighting for ship type should be assigned according to the risk level corresponding to the ship type. The basic risk factor is calculated using the following formula: ； in, As a distance factor, , The attenuation coefficient is... ; The relative velocity factor, ; , The velocity component of the neighboring ship; , is the maximum relative speed of the ship; For heading cross factor, based on relative heading Assess the encounter situation and assign a risk coefficient. , For the neighboring ship's course; S34, Edge weights based on step S33 It is then normalized to achieve a unified measurement of multi-dimensional risk. The normalization formula is as follows: ； in, The initial edge weights, ; for Normalization processing; The maximum value of the initial weights, This represents the minimum value of the initial weights.

8. The maritime traffic risk assessment method according to claim 7, characterized in that: Step S4 Specifically, the following steps are included: S41. Based on the track data of the ship and neighboring ships obtained in step S2, and combining the principles of ship kinematics, calculate the real-time minimum encounter distance between the ship and each neighboring ship. and the time of arrival at the closest point , The calculation formula is: ； in, , These are the relative position vector and relative velocity vector of the ship and the neighboring ship, respectively. The calculation formula is: ； S42, Edge weights based on step S3 And based on step S41 , A single-ship comprehensive risk model and a multi-ship comprehensive risk model are constructed for each neighboring vessel relative to the vessel. The calculation process for the single-ship comprehensive risk model is as follows: Piecewise quantization Risk coefficient, formula is: ； in, for Risk factor; This is the preset minimum safe meeting distance; ,and hour, This refers to a collision between this vessel and a neighboring vessel. The distance between our vessel and neighboring vessels is safe at this time; Quantification Risk coefficient, formula is: ； in, For the risk factor of sailing time; The preset safe arrival time; ,and hour, This refers to a collision between this vessel and a neighboring vessel. The distance between our vessel and neighboring vessels is safe at this time; The formula for the comprehensive risk value of a single ship is: ； in, This is the comprehensive risk value for a single vessel, and , A large value indicates a high overall risk of collision with neighboring vessels. for The weighting coefficient is related to the waters where the ship is located; The calculation process for the multi-ship integrated risk model is as follows: A weighted model employing maximum risk as the primary factor and average risk as the secondary factor is adopted, and the formula is as follows: ； in, To mitigate the combined risks of multiple vessels, and , A large value indicates a high overall risk of collision with neighboring vessels. The maximum risk weighting coefficient highlights the dominant role of high-risk vessels in overall risk. for The highest single-ship risk value among neighboring vessels compared to this vessel.

9. A maritime traffic risk assessment method according to claim 8, characterized in that: Step S5 specifically includes the following steps: S51. Based on the comprehensive risk value of a single vessel in step S42, pre-screen the risk value of a single vessel and determine the risk level of a single vessel, and form a risk list of a single vessel to provide a basis for subsequent overall risk assessment. S52, Multi-ship Integrated Risk Based on Step S42 As the initial risk level, combined with the single-ship risk list from step S51, the risk level is adjusted according to the principle of prioritizing local threats to determine the overall risk level.

Images