A method for predicting and warning ship encounters in an inland river hub section

Abstract

The present invention discloses a method for predicting and warning ship encounters in an inland river hub section, comprising: extracting, filtering and fusion processing of historical route data: mining historical tracks: using the data after fusion processing in step 1 to obtain a historical route set; track matching: acquiring real-time dynamic data of ships and waterway operation-related data, and finding a route track that best matches the current real-time ship track from the historical route set; calculating the ship's navigation track for a period of time in the future; track prediction optimization: caching the predicted navigation track, and using an iterative algorithm to continuously optimize the estimated arrival time of subsequent waypoints; encounter recognition: performing spatial segmentation based on the predicted track, and when the segmented area encounters the predicted track of other ships, estimating the area, ship and corresponding time of encounter, and performing encounter and warning prompts; combining video monitoring with multi-source data, the early warning of key areas is more reliable, and the navigation safety of inland river hub ships and the navigation efficiency of waterways are improved.

Description

Technical Field

[0001] The invention belongs to the field of inland river ship traffic, and specifically relates to a method for predicting and warning ship encounters in an inland river hub section. Background Art

[0002] Most of the research on ship encounter prediction is based on the analysis of ship navigation data in open waters. By analyzing the ship navigation data in such waters, it can be found that the encounter situation of ships in such waters is relatively complex. Due to factors such as curved waterways, turbulent water flow, different water depths, and improper navigation mark settings, ships often collide and run aground during navigation. There is little research on the encounter situation of ships in complex inland waterways, especially in inland hub sections.

[0003] In short, the prediction and management of ship encounters in complex waterways in inland river hubs require in-depth research and analysis, and a series of measures and technical means should be taken to improve the safety of waterways in inland river hubs and the navigation safety of ships. Summary of the invention

[0004] In view of the problems or defects in the prior art, the present invention provides a method for predicting and warning ship encounters in an inland river hub section, so as to improve the safety of the waterway in the inland river hub section and the navigation safety of ships.

[0005] In order to achieve the above-mentioned object, the present invention relates to a method for predicting and warning ship encounters in an inland river hub section, comprising the following steps:

[0006] Step 1: Extract, filter and integrate historical route data:

[0007] Step 2: Historical trajectory mining: Using the data fused and processed in step 1, the ship's navigation trajectory is divided from the business node, and different trajectory segments are formed based on the divided data. When the trajectories that meet the set distance condition are aggregated into a certain trajectory, a historical route set is obtained;

[0008] Step 3: Trajectory matching

[0009] Obtain the real-time dynamic data of the ship and the data related to the waterway operation, and find a route track that best matches the current real-time ship track from the historical route set in step 2; calculate the ship's navigation track for a period of time in the future through the best matching route track and combined with the variable factors that need to be considered in the hub waterway, and asynchronously (manually) return the inference result to the front end;

[0010] Step 4: Trajectory prediction optimization

[0011] Cache the predicted navigation trajectory, wait for the ship to actually sail to the predicted position, then compare it with the actual position, and use an iterative algorithm to continuously optimize the estimated arrival time of subsequent waypoints;

[0012] Step 5: Identify the meeting

[0013] Based on the predicted trajectory, space segmentation is performed. When the segmented area encounters the predicted trajectory of other ships, the encounter area, ship and corresponding time are estimated, and encounter and warning prompts are given according to the ship's driving safety factor J, navigation area, alarm triggering conditions;

[0014] Furthermore, the method for extracting, filtering and fusing historical route data in step 1 is specifically as follows: extracting route data using data from multiple devices including historical AIS, Beidou positioning, and radar; filtering out dirty data with abnormal positioning using a filtering algorithm; fusing the ship positioning information identified by AIS, radar, and Beidou through fusion positioning. The multi-data fusion framework based on AIS includes an AIS multi-target tracker, a radar multi-target tracker, and a Beidou multi-target tracker. The three work independently of each other to form their own tracks, and then all tracks are extrapolated to the same time point through time calibration, and the tracks are calibrated using a data fusion algorithm.

[0015] Furthermore, the filtering algorithm adopts a time smoothing algorithm, that is, the offset angle and distance of the surface horizontal position coordinates of two adjacent time points are 3 times the average distance of the first three points.

[0016] Furthermore, the variable factors to be considered for the hub channel in step 3 include:

[0017] Changes in route during flood and dry seasons, such as lateral displacement of the route under the influence of flow.

[0018] Switching of routes for launching and launching in a specific area due to traffic control reasons.

[0019] Different lock routes are selected according to the scheduling plan.

[0020] Furthermore, the method of using an iterative algorithm to continuously optimize the estimated arrival time of subsequent waypoints in step 4 is: when the ship travels to the vicinity of the predicted point range (reference 20M*20M represents within this range), input the current speed.

[0021] By comparing the data, the weighted average speed of the historical data of this point in the big data planning route is obtained.

[0022] Calculate by the formula XIGMA=V / V(HIS)

[0023] ΔTn=T(HIS) / XIGMA

[0024] Re-adjust the time when the ship arrives at the subsequent coordinate sequence points of the route δT1, δT2, δT3.....δTn.

[0025] In the formula, XIGMA represents the speed contrast coefficient;

[0026] V(HIS) represents the average speed of historical big data weighted to the current coordinate area point;

[0027] V represents the actual speed of the ship arriving at the current coordinate area point;

[0028] ΔTn represents the actual calculation time used to adjust the current speed and heading to reach the subsequent coordinate point.

[0029] Output the optimized time δT1, δT2, δT3, .....δTn for the ship to reach the subsequent target point.

[0030] When the ship arrives at the next coordinate point, the formula is repeated to iteratively correct the optimized estimated arrival time under the route.

[0031] Furthermore, the ship driving safety factor J in step 5 is determined according to the following table:

[0032] Ship Type Safety Factor Ordinary passenger ship 1 General bulk carrier 1 Old small low ship 1.5 Dangerous goods ships 2 Container Ship 1.5 .

[0033] Furthermore, the alarm triggering conditions in step 5 include: general prompt: a ship encounter prompt calculated within T<120s.

[0034] Warning reminder: Design the warning distances for different areas according to the severity and urgency of the encounter.

[0035] According to navigation experience, sail in the same direction for 400 meters.

[0036] The two ships sail towards each other for 800 meters, and one of the ships deviates from the historical direction vector by 5°. 1200 meters is the alarm threshold, and the encounter at future time t is calculated and compared.

[0037] For ferry crossings, anchorages and piers, radar, AIS, Beidou data combined with the Eagle Eye camera in the area are used. When the ship is less than 1000M away from this area and the ship supervision facilities find abnormal movement of the ship in the area, an alarm is triggered immediately; the alarm distance for the main ship is L = L*J (J is the ship safety factor)

[0038] Furthermore, the alarm triggering conditions in step 5 also include: In particular, for ships docking in key areas such as piers, the Eagle Eye camera algorithm generates abnormal positioning and immediately pushes it to the main ship in the sailing direction. For anchored ships, abnormal positioning is also required to generate alarm signals for the main ships within the distance range;

[0039] In particular, when the ship's lateral deviation from the planned route is >30M and the oncoming ship is <1000m, an encounter alarm is issued;

[0040] In particular, for the range of the channel where overtaking is prohibited, an alarm is generated when L < the set safety distance;

[0041] In particular, when leaving the channel and entering the ship lock power generation area, an alarm is generated when the direction vector changes abnormally;

[0042] Specifically, when a ship changes from a parallel route to a single route, such as when it sails out from under the ship lift and the Three Gorges ship lock, or when it sails out from under the No. 2 and No. 3 locks, an alarm is issued when the two headings of ships are about to intersect in the route area.

[0043] Compared with the prior art, the present invention has the following effects:

[0044] (1) The inland river hub section ship encounter prediction and early warning method of the present invention realizes high-accuracy trajectory prediction of ships in inland river hub sections, provides a basis for early warning of ship collision risks, combines video monitoring with multi-source data, and makes early warnings in key areas more reliable, thereby improving the navigation safety of inland river hub ships.

[0045] (2) The inland river hub section ship encounter prediction and warning method of the present invention can effectively warn ships to make navigation smoother and reduce navigation congestion and delays caused by temporary emergency avoidance; differentiate warnings based on ship types and navigation areas can be used to avoid a decrease in navigation efficiency caused by excessive warnings; and warning information can help traffic management departments to dynamically dispatch waterways, thereby improving waterway navigation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A schematic diagram of a flow chart of a preferred embodiment of the present invention; DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0048] See also Figure 1 The present invention relates to a method for predicting and warning ship encounters in an inland river hub section, comprising the following steps:

[0049] Step 1: Extract, filter and fuse historical route data: The specific method is: extract route data using data from multiple devices including historical AIS, Beidou positioning, and radar; use filtering algorithms to filter out dirty data with abnormal positioning; fuse the ship positioning information identified by AIS, radar, and Beidou through fusion positioning. The fusion of multi-source data has the following advantages:

[0050] 1. Improve the comprehensiveness of data; AIS, Beidou and radar each have their own advantages. AIS can obtain dynamic information of ships, Beidou can perform high-precision positioning, and radar can detect ships in the blind area of ​​AIS signals. Multi-source data fusion can complement each other's shortcomings.

[0051] 2. Improve data accuracy. A single data source may have errors, such as AIS failure, signal loss due to blocking, and Beidou errors due to weather. Multi-source data fusion can reduce errors through mutual verification and correction.

[0052] 3. Improve the real-time performance of data. Different systems have different update rates. Multi-source fusion can utilize the data with better real-time performance to improve the overall data real-time performance.

[0053] 4. Improve the continuity of the trajectory. A single system may have breakpoints due to various factors, and multi-source data fusion can make the trajectory more continuous and complete.

[0054] 5. Provide redundancy effect. When a single system fails, data from other systems can be used to compensate.

[0055] Improved system reliability.

[0056] Comprehensively reflect the ship's motion status. Multi-source data can reflect the ship's position, speed, trajectory, heading and other motion status from different aspects, and can more accurately predict the trajectory;

[0057] Step 2: Historical trajectory mining: Using the data fused and processed in step 1, the ship's navigation trajectory is divided from the business node, and different trajectory segments are formed based on the divided data (the channel is divided into different segments S1, S2, S3 divided from the business node as a parallel segment, S4, S5, S6 as a parallel segment, SN-2, SN-1, SN as a parallel segment. By importing these variable units, different route serial segments such as S1, S4, ..., SN are automatically formed and selected). When the trajectories that meet the set distance condition are aggregated into a certain trajectory, a historical route set is obtained;

[0058] Step 3: Trajectory matching:

[0059] Obtain the real-time dynamic data of the ship and the data related to the waterway operation, and find a route track that best matches the current real-time ship track from the historical route set in step 2; calculate the ship's navigation track for a period of time in the future through the best matching route track and combined with the variable factors that need to be considered in the hub waterway, and asynchronously (manually) return the inference result to the front end;

[0060] Step 4: Trajectory prediction optimization:

[0061] Cache the predicted navigation trajectory, wait for the ship to actually sail to the predicted position, then compare it with the actual position, and use an iterative algorithm to continuously optimize the estimated arrival time of subsequent waypoints;

[0062] Step 5: Meeting Identification:

[0063] Based on the predicted trajectory, spatial segmentation is performed. When the segmented area encounters the predicted trajectory of other ships, the encounter area, ship and corresponding time are estimated, and encounter and warning prompts are given according to the ship's driving safety factor J, navigation area, alarm triggering conditions.

[0064] The AIS-based multi-data fusion framework includes AIS multi-target tracker, radar multi-target tracker and Beidou multi-target tracker. The three work independently to form their own tracks, and then all tracks are extrapolated to the same time point through time calibration, and the tracks are calibrated using data fusion algorithm.

[0065] The filtering algorithm adopts a time smoothing algorithm, that is, the offset angle and distance of the surface horizontal position coordinates of two adjacent time points are three times the average distance of the first three points.

[0066] The variables to be considered in the hub channel in step 3 include:

[0067] Route changes during the flood season and dry season, such as lateral displacement of the route under the influence of flow. (Set route selection under different hydrological conditions)

[0068] Switching of routes for water transportation and water transport in a specific area due to traffic control (taking the route switching from Laohudong to Dashaba section between the two dams of Gezhouba in the Three Gorges from 18:00 on June 1st to 18:00 on September 30th every year, setting the route selection under different traffic management conditions)

[0069] Different lock routes are selected according to the scheduling plan (for example, the selection of Gezhouba Lock No. 1 / 2 / 3, the selection of Three Gorges Lock and Ship Lift; route selection under different scheduling conditions).

[0070] The method of using an iterative algorithm to continuously optimize the estimated arrival time of subsequent waypoints in step 4 is: when the ship travels to the vicinity of the predicted point range (reference 20M*20M represents within this range), input the current speed.

[0071] By comparing the data, the weighted average speed of the historical data of this point in the big data planning route is obtained.

[0072] Calculate by the formula XIGMA=V / V(HIS)

[0073] ΔTn=T(HIS) / XIGMA

[0074] Re-adjust the time δT1, δT2, δT3... of the ship arriving at the coordinate sequence points behind the route.

[0075] ΔTn.

[0076] In the formula, XIGMA represents the speed contrast coefficient;

[0077] V(HIS) represents the average speed of historical big data weighted to the current coordinate area point;

[0078] V represents the actual speed of the ship arriving at the current coordinate area point;

[0079] ΔTn represents the actual calculation time used to adjust the current speed and heading to reach the subsequent coordinate point.

[0080] Output the optimized time δT1, δT2, δT3, .....δTn for the ship to reach the subsequent target point.

[0081] When the ship arrives at the next coordinate point, the formula is repeated to iteratively correct the optimized estimated arrival time under the route.

[0082] In step 5, the ship's driving safety factor J is determined according to the following table:

[0083] According to dangerous goods ships, passenger ferries, cargo ships, old, small and low-speed vessels, etc., the safety factor of ordinary cargo ships is 1. According to relevant management regulations, the larger the value, the greater the safety factor that needs to be paid attention to, and the more urgent the supervision needs to be strengthened. (For example, old, small and low-speed vessels are 1.5, dangerous goods are 2, etc., combined with actual traffic control regulations and experience.

[0084] Ship Type Safety Factor Ordinary passenger ship 1 General bulk carrier 1 Old small low ship 1.5 Dangerous goods ships 2 Container Ship 1.5 .

[0085] The alarm triggering conditions in step 5 include: General prompt: the ship will encounter a prompt within T<120s.

[0086] Warning reminder: Design the warning distances for different areas according to the severity and urgency of the encounter.

[0087] According to navigation experience, sail in the same direction for 400 meters.

[0088] The two ships sail towards each other for 800 meters, and one of the ships deviates from the historical direction vector by 5°. 1200 meters is the alarm threshold, and the encounter at future time t is calculated and compared.

[0089] For ferry crossings, anchorages and berthing areas, radar, AIS, Beidou data combined with the Eagle Eye camera in the area are used. When the ship is less than 1000M away from this area and the ship supervision facilities detect abnormal movement of ships in the area, an alarm is triggered immediately; the alarm distance for the main ship is L = L*J (J is the ship safety factor)

[0090] The alarm triggering conditions in step 5 also include: In particular, for ships docking in key areas such as piers, the Eagle Eye camera algorithm generates abnormal positioning and immediately pushes it to the main ship in the sailing direction. For anchored ships, abnormal positioning is also required to generate alarm signals for the main ships within the distance range;

[0091] In particular, when the ship's lateral deviation from the planned route is >30M and the oncoming ship is <1000m, an encounter alarm is issued;

[0092] In particular, for the range of the channel where overtaking is prohibited, an alarm is generated when L < the set safety distance;

[0093] In particular, when leaving the channel and entering the ship lock power generation area, an alarm is generated when the direction vector changes abnormally;

[0094] Specifically, when a ship changes from a parallel route to a single route, such as when it sails out from under the ship lift and the Three Gorges ship lock, or when it sails out from under the No. 2 and No. 3 locks, an alarm is issued when the two headings of ships are about to intersect in the route area.

[0095] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting and warning ship encounters in an inland river hub section, characterized in that: The steps include: Step 1: Extract, filter and integrate historical route data: Step 2: Historical trajectory mining: Using the data fused and processed in step 1, the ship's navigation trajectory is divided from the business node, and different trajectory segments are formed based on the divided data. When the trajectories that meet the set distance condition are aggregated into a certain trajectory, a historical route set is obtained; Step 3: Trajectory matching: Obtain the real-time dynamic data of the ship and the data related to the waterway operation, and find a route track that best matches the current real-time ship track from the historical route set in step 2; calculate the ship's navigation track for a period of time in the future through the best matching route track and combined with the variable factors that need to be considered in the hub waterway, and return the inferred result to the front end; Step 4: Trajectory prediction optimization: Cache the predicted navigation trajectory, wait for the ship to actually sail to the predicted position, then compare it with the actual position, and use an iterative algorithm to continuously optimize the estimated arrival time of subsequent waypoints; Step 5: Meeting Identification: Based on the predicted trajectory, the space is segmented. When the segmented area encounters the predicted trajectory of other ships, the area, ship and corresponding time of encounter are estimated, and the encounter and warning prompts are given according to the ship's driving safety factor J, navigation area, and alarm triggering conditions; the alarm triggering conditions include: General prompt: calculate the ships within the encounter time T<120s and give the encounter prompt; Warning reminder: Design the warning distance of different areas according to the severity and urgency of the encounter; According to navigation experience, sail 400 meters in the same direction; Sailing towards each other for 800 meters, and one of the ships deviates from the historical direction vector by 5°, 1200 meters is the alarm threshold, and the encounter at the future time t is calculated and compared; For ferry crossings, anchorages and berthing areas, radar, AIS, Beidou data combined with the Eagle Eye camera in the area are used. When the ship is less than 1000M away from the area and the ship supervision facilities detect abnormal movement of the ship in the area, an alarm is triggered immediately; the alarm distance for the main ship is L=L*J, where J is the ship safety factor; The method of using an iterative algorithm to continuously optimize the estimated arrival time of subsequent waypoints in step 4 is: when the ship travels to the vicinity of the predicted point range, the weighted average speed of the historical data of the point of the planned route is compared by data; Calculated by the formula: XIGMA = V / V(HIS) ΔTn=T(HIS) / XIGMA Re-adjust the time δT1, δT2, δT3...δTn for the ship to arrive at the next coordinate sequence point on the route: XIGMA in the formula represents the speed comparison coefficient; V(HIS) represents the weighted average speed of the historical big data to arrive at the current coordinate area point; "T(HIS)" is the weighted average travel time of the historical big data planned route to this point; V represents the actual speed of the ship arriving at the current coordinate area point; ΔTn represents the actual calculation time used to adjust to the subsequent coordinate point with the current speed and heading; output the optimized time δT1', δT2', δT3'...δTn' for the ship to arrive at the subsequent target point; when the ship arrives at the next coordinate point, repeat this formula to iteratively correct the optimized estimated arrival time under the route.

2. The method for predicting and warning ship encounters in an inland river hub section according to claim 1 is characterized in that: The method for extracting, filtering and fusing historical route data in step 1 is specifically as follows: extracting route data using data from multiple devices including historical AIS, Beidou positioning, and radar; filtering out dirty data with abnormal positioning using a filtering algorithm; The ship positioning information identified by AIS, radar and Beidou is fused through fusion positioning. The AIS-based multi-data fusion framework includes AIS multi-target tracker, radar multi-target tracker and Beidou multi-target tracker. The three work independently of each other and form their own tracks. Then, all tracks are extrapolated to the same time point through time calibration, and the tracks are calibrated using data fusion algorithm.

3. The method for predicting and warning of ship encounters in inland river hub sections according to claim 2 is characterized in that: The filtering algorithm adopts a time smoothing algorithm, that is, the offset angle and distance of the surface horizontal position coordinates of two adjacent time points are three times the average distance of the first three points.

4. The method for predicting and warning of ship encounters in inland river hub sections according to claim 1 is characterized in that: The variables to be considered in the hub channel in step 3 include: Change of route during flood season and dry season; Switching of routes for launching and disembarking in a specific area due to traffic control reasons; Different lock routes are selected according to the scheduling plan.

5. The method for predicting and warning of ship encounters in inland river hub sections according to claim 1 is characterized in that: In step 5, the ship's driving safety factor J is determined according to the following table: 。 6. The method for predicting and warning ship encounters in an inland river hub section according to claim 1 is characterized in that: The alarm triggering conditions in step 5 also include: when a ship is docked and sideways in the key area of ​​the pier, the Eagle Eye camera algorithm generates abnormal positioning and immediately pushes it to the main ship in the sailing direction; when the anchored ship is anchored, abnormal positioning is also required to generate an alarm signal trigger for the main ship sailing within the distance range; When the ship's lateral deviation from the planned route is greater than 30 meters and the oncoming ship is less than 1000 meters, an encounter alarm is issued; For the range of the channel where overtaking is prohibited, an alarm is generated when the warning distance L is less than the set safety distance; When leaving the channel and entering the ship lock power generation area, an alarm is generated when the direction vector changes abnormally; When a ship changes from a parallel route to a single-track route, such as when sailing out from under the ship lift and the Three Gorges Ship Lock, or when sailing out from under the No. 2 and No. 3 locks, an alarm is issued when the two headings of ships are about to intersect in the route area.

Images