Inland river bridge group water area ship navigation collision avoidance method and system

Abstract

The invention belongs to the technical field of navigation, and discloses an inland river bridge group water area ship navigation collision avoidance method and system, and the method comprises the steps: carrying out the environment modeling based on an electronic river map; designing a bridge group collision avoidance path planning algorithm; and designing a bridge group navigation mark auxiliary yaw early warning system. An inland river bridge group water area ship navigation collision avoidance system is established, and environment information is obtained based on an electronic river map and processed into an environment model capable of being directly utilized. A ship-bridge group is used as a research object, a ship-bridge group collision-free path is established based on a voronoi method, factor constraints are considered, and a ship bridge group operation collision avoidance strategy is researched. The navigation track of the ship is predicted, navigation mark acousto-optic early warning prompt is supplemented for the ship yawing or having the collision risk, and the bridge group water area ship yawing early warning function is achieved.

Description

Technical Field

[0001] The invention belongs to the field of navigation technology, and in particular relates to a method and system for avoiding collision of ships navigating in inland bridge group waters. Background Art

[0002] With the rapid development of my country's economy, there are more and more inland bridges, and the number and tonnage of ships are increasing, which increases the risk of ship-bridge collision. For ship navigation, bridges and piers are regarded as obstacles. The construction of bridge piers changes the direction of water flow in the channel, which in turn induces the evolution of the riverbed, which greatly increases the difficulty of ship driving personnel to operate, and it will bring potential collision risks to ship navigation. According to statistics, accidents of ships colliding with bridge piers occur frequently in the three major river systems of the Yangtze River, Pearl River and Heilongjiang River in China. In the winding Yangtze River, taking the Wuhan section as an example, more than ten cross-river bridges span the 18km waters, forming a bridge group area. The bridge group waters have the characteristics of small bridge spacing, different bridge axis directions, staggered navigation holes, and more complex flow patterns in the bridge group waters. Ship navigation in the bridge group waters is more complicated than navigation in the single bridge waters. The waters connected by multiple bridges in a short distance have complex water flow environments and changeable navigation conditions. The decision-making of ship-bridge avoidance is restricted before and after, and it is necessary to adjust the ship position multiple times in a short time and take ship avoidance measures. At present, most inland river bridges have implemented active anti-collision warning systems or passive anti-collision devices. For waters with bridge groups, if the problem is solved that ships can obtain safe and reliable navigation paths in waters with bridge groups, avoid collisions with bridge piers and navigate safely, ships can avoid the risk of collision, which is more meaningful for the safe navigation and supervision of ships in bridge groups. Summary of the invention

[0003] In view of the problems existing in the prior art, the present invention provides a method and system for ship navigation collision avoidance in inland bridge group waters.

[0004] The present invention is achieved by providing a method for avoiding collisions between ships in inland bridge group waters, the method comprising:

[0005] S1: Environmental modeling based on electronic river map;

[0006] S2: Design of collision avoidance path planning algorithm for bridge groups;

[0007] S3: Design of bridge group navigation aid deviation warning system.

[0008] Further, the S1 specifically includes:

[0009] (1) Extraction and storage of the environment: converting the information in the electronic river map into an environmental model that can be directly used. The representation and storage of the environmental model is the basis for solving the problem of model establishment;

[0010] (2) Determination of dangerous obstacle areas. For high-speed ships, dangerous obstacle areas include obstacles such as land, islands, reefs, and buoys. In the required corresponding layer, the index file (*.shx) can read the storage location of the required target in the master file, and use the obtained index value to read the specific data of the required target in the master file;

[0011] (3) Extraction and storage of environmental information

[0012] ① Extraction of regional information. A region is composed of a closed edge set or a closed connection point. In order to better extract information data representing a region, the vector records constituting the boundary of such a region must be viewed in a certain order. The order must be met. First, the internal boundary is encoded, and then the external boundary is encoded. Before encoding the internal boundary, the external boundary must be completely encoded.

[0013] The direction of the region's encoding extraction:

[0014] When encoding a region, its direction also needs to be specified. The outer boundary of the region must be encoded in a clockwise direction, and the boundary inside the region needs to be encoded in a counterclockwise direction.

[0015] ② Extraction of isobath and shoreline information data

[0016] Perform data extraction operations in a given electronic river map data file; in the data file, first read the first logical record, and correctly identify the total number of fields, area identifiers, control field length, and other information in the header area, directory area, control field area, and description field area. After parsing the control field area, you can get the attributes of the class; after parsing the description field area, you can get the control structure of the field. After executing this program, jump to the spatial record, traverse the shoreline and isobath boundary points, determine whether the boundary is divided by the data boundary or is blocked, store the information in an array, and finally output;

[0017] ③Data information storage and display process

[0018] Store the filtered and extracted information, and then display the required content more conveniently and quickly;

[0019] (4) Electronic river map object processing

[0020] Intercept the electronic river map of a certain water area. In the vicinity of actual islands and land, as the distance of the water area close to the land decreases, the water depth also gradually decreases. In order to enable ships to safely navigate in a certain water depth, the 10m isobath is taken as the collision avoidance limit for ships in the electronic river map. Water areas and land areas within a water depth of 10m are considered dangerous areas. Therefore, in the process of ship collision avoidance, collision avoidance treatment is performed on the 10m isobath to safely avoid collisions in shallow water areas and land areas.

[0021] Extract relevant graphic information such as islands, land, isobaths, buoys, etc. from the electronic river map, and accurately find the longitude and latitude coordinates of any point in the map;

[0022] For the coordinates of dangerous objects such as buoys and reefs (x i ,y i ), i is the center point of the dangerous object, and there is Then the buoy and reef can be described as π2r 2 , r is the radius of dangerous objects with particle characteristics such as buoys and reefs;

[0023] When avoiding collisions with dangerous objects such as land and islands, it is only necessary to avoid collisions with the 10m isobaths along the perimeter of the object. The isobaths are composed of countless points. In the process of processing them, the coordinates (x m ,y m ), m=1,2,3......n, and there is Then the 10m isobath can be described as a line consisting of (x m ,y m ) as the center, with r m =A 10m isobath composed of circles with a radius of m, so that the ship can regard m circles as static obstacles, and at the same time:

[0024]

[0025] l is the length of the 10m isobath, and when the equal sign is taken, m circles are tangent. Using this method, graphic symbols can be converted into mathematical symbols so that ships can directly read and calculate.

[0026] (5) Determination of narrow water obstacle areas

[0027] When a water area is detected to be a narrow area and there are multiple islands, shoals and other areas in the water area, the obstacle area is first calibrated, and then the islands and shoals are rasterized with the help of planning methods to determine the scope of the dangerous area.

[0028] (6) Obstacle Detection Method

[0029] In the electronic river map, 100m along its heading direction is set as the detection base, and the square dynamic window of axa is used as the detection area. When it is beyond 100m, a takes a larger value, and when it is within 100m, a takes a smaller value. And as the distance to the obstacle area shortens, the value of a gradually decreases to adapt to the changes in the environment. When it is sailing on the water surface, when the dynamic window detects dangerous objects such as 10m isobaths, islands, reefs or buoys, the distribution and scale of the surrounding obstacles are detected. Obstacles such as land and buoys in open waters are determined. The circular mark processing environment model is used to determine the value of the mark scale r; when it is determined that obstacles are densely distributed in non-open waters, rasterization environment processing is performed to determine the obstacle area for collision avoidance.

[0030] Further, the S2 specifically includes:

[0031] According to the layered electronic river map, a grid map with a certain accuracy is established; on the established grid map, the dynamic particle objects and the longitude and latitude coordinates of the pier positions are used to construct the Voronoi diagram to form a path set, and the common part of the navigable waters and the path set is selected to establish an approximate navigable path. Its main disadvantage is that there are many turning points in the path, which is not conducive to the maneuvering and navigation of the ship; therefore, it is necessary to optimize the approximate navigable path and construct the optimal safe and time-saving navigable path; in the process of path optimization, the ship's maneuvering motion characteristics and related rule constraints and external environmental interference are considered, and the approximate navigable path with multiple turning points is optimized, and the optimal safe and time-saving navigation path is selected.

[0032] Further, the S3 specifically includes:

[0033] (1) Ship track prediction

[0034] The ship track prediction method based on recurrent neural network is adopted, which mainly includes two parts: data preprocessing and neural network prediction. In the data preprocessing part, a data subset similar to the predicted target motion characteristics is extracted from the AIS data based on the symmetric segmented path distance to eliminate the influence of a large amount of data redundancy and noise; in the neural network prediction part, considering the characteristics of the neural network model such as nonlinearity, adaptability, self-learning ability and arbitrary function approximation ability, a recurrent neural network model with gated recurrent unit GRU as the core is constructed to achieve accurate and efficient prediction of ship position;

[0035] (2) Simulation experiment

[0036] The measured ship AIS data is used as the original data, and the noise error generated by the simulation software is added to simulate the historical stock AIS trajectory of the ship. The ship trajectory prediction algorithm based on the recurrent neural network is verified in the integrated development environment Matlab2022b, and the actual value, observation value and prediction value are finally obtained;

[0037] (3) Design of navigation aid warning system

[0038] The system is connected to the navigation mark management system. After predicting that the ship is at risk of deviating from the planned route based on the track, the auxiliary warning module of the navigation mark management system starts the warning function. The warning intensity can be determined by itself based on the distance data between the navigation mark and the target ship. If the ship is relatively close to the navigation mark at this time, the voice alarm combined with VHF communication and the flash alarm based on the matrix light will be started at the same time; when the distance is far, only the VHF voice alarm will be started to avoid the sudden flashing of the navigation mark matrix light causing visual impact on other normal sailing ships around.

[0039] Furthermore, the data preprocessing algorithm based on SSPD first converts the latitude and longitude coordinates of the data into plane coordinates through the Gauss-Krüger projection method, and then uses the linear interpolation method to supplement the track data. Then, based on the SSPD method, a large number of historical tracks and the target tracks to be measured are measured for similarity, and the distance between each track is calculated. Finally, the tracks are screened according to the distance to finally obtain a subset of relevant track data that meets the screening conditions.

[0040] Furthermore, the GRU recurrent neural network prediction model structure includes 1 input node, 5 hidden layers (GRU layer, Dropout layer and Dense layer) and 1 output node. The length of the input node receiving sequence X is n and the dimension is 2 (i.e., the longitude and latitude coordinates of the ship). Through hidden layer training, the best prediction sequence Y is obtained at the output node, including the longitude and latitude coordinates of the predicted position.

[0041] Furthermore, the voice alarm of the VHF communication specifically includes: making a warning instruction according to the bridge group area where the ship is located, and the navigation mark near the waters where the target ship is located broadcasts the warning information by sound, prompting the ship to deviate from the planned path or have a risk of collision with the bridge group;

[0042] The flash alarm of the matrix light specifically includes: the navigation mark management system makes a warning instruction according to the warning area where the ship with deviation risk is located, and sends the flash alarm instruction to the warning navigation mark through the wireless transmission module. After the navigation mark data receiving module receives the warning instruction, it immediately activates the matrix light to flash the dangerous ship.

[0043] Another object of the present invention is to provide a system for avoiding collisions between ships in inland river bridge group waters for implementing the method for avoiding collisions between ships in inland river bridge group waters, the system comprising:

[0044] Modeling module, based on electronic Jiangtu environment modeling;

[0045] Algorithm design module, used for bridge group collision avoidance path planning algorithm design;

[0046] System design module, used for the design of bridge group navigation aid yaw warning system.

[0047] Another object of the present invention is to provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for avoiding collisions between ships navigating in waters of inland bridge groups.

[0048] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the method for avoiding collisions between ships navigating in waters of inland bridge groups.

[0049] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0050] First, in view of the technical problems existing in the above-mentioned prior art, some creative technical effects are brought about after solving the problems. The specific description is as follows:

[0051] 1. In the waters of inland bridge groups, a collision avoidance path for ships from the first bridge to the last bridge of the bridge group is generated, and a real-time steering and speed control strategy for sailing along a safe and reliable planned path in the waters of the bridge group is proposed.

[0052] 2. Considering the ship motion characteristics, inland ship collision avoidance rules, lane separation rules and bridge group navigation clearance restrictions, a bridge group water area collision avoidance path planning algorithm based on the Voronoi diagram method is proposed. The path is locally optimized for the changeable inland river environment to form a navigable collision avoidance path with the lowest risk.

[0053] 3. Design a ship navigation collision avoidance system in bridge group waters, supplemented by navigation aid warning function, to realize the ship navigation collision avoidance function and deviation warning in bridge group waters.

[0054] The present invention mainly studies the collision avoidance method for ships passing through bridge group waters, and designs a corresponding system to realize the collision avoidance function for ships sailing through bridge group waters. First, the ship position information is obtained based on the acquired GPS data; the bridge group information and ship information are obtained based on the electronic river map and AIS information, and an environmental model is established. Secondly, the ship-bridge group waters collision avoidance method is designed, the collision avoidance trajectory is planned, and the maneuvering strategy is proposed. Finally, a ship navigation collision avoidance system in bridge group waters is realized based on the ship-bridge group trajectory planning and the ship navigation bridge group maneuvering strategy analysis.

[0055] Second, the present invention establishes a collision avoidance system for ships in inland bridge group waters, obtains environmental information based on electronic river maps, and processes it into an environmental model that can be directly used. Taking the ship-bridge group as the research object, based on the Thiessen polygon method (Voronoi method), a collision-free path for the ship-bridge group is established, and the constraints of factors are considered to study the collision avoidance strategy for the ship crossing the bridge group. The navigation trajectory of the ship is predicted, and the sound and light warning prompts of navigation marks are provided for ships that are yawed or have a collision risk, so as to realize the ship yaw warning function in the bridge group waters.

[0056] Third, the expected benefits and commercial value of the technical solution of the present invention after transformation are:

[0057] As the number of inland river bridges gradually increases and their density increases, and there are also complex and changing river environments, it becomes more difficult for ships to safely navigate in bridge-group waters. In the past, bridge-group navigation was basically based on the experience of the crew, and required careful driving and multiple steering. Accidents such as collisions with bridge piers often occurred without attention or due to negligence. Therefore, the traditional bridge-group navigation method can no longer meet the safe operation of ships, and it is necessary to adopt intelligent and efficient methods to innovate and improve the collision avoidance maneuvering strategy in the bridge area. The research of this project can form auxiliary measures for ships to safely navigate in bridge-group waters and reduce the probability of safety accidents such as collisions.

[0058] With the continuous improvement of economic development and urban transportation, the number of bridges in inland waters has also increased. In short-distance dense bridge waters, safety issues such as ship safe navigation and bridge avoidance have become increasingly prominent. This project focuses on reducing the risk of ship-bridge collisions, which will reduce the labor cost of inland shipping and the cost of collision accidents.

[0059] Third, the present invention aims at the collision avoidance problem of ships in inland bridge group waters, especially in complex hydrological and geographical environments, such as navigation safety problems in bridge-dense areas, and proposes a ship navigation collision avoidance method based on electronic river maps. This method integrates environmental modeling, intelligent path planning and real-time warning system, solves the following existing technical problems, and has achieved significant technical progress:

[0060] 1. Solve the problem of insufficient environmental information: The present invention extracts and stores detailed environmental information of inland river bridge waters, including land, islands, bridges, waterways, etc., through electronic river map environmental modeling, providing rich and accurate environmental data for ship navigation, solving the problem of traditional navigation relying on simplified or outdated information.

[0061] 2. Optimized path planning algorithm: The present invention designs a collision avoidance path planning algorithm tailored to the characteristics of bridge group waters. It not only takes into account the dynamic characteristics and safe navigation distance of the ship, but also integrates multi-ship collision avoidance strategies, realizing efficient and safe path planning in complex waters, and significantly improving navigation efficiency and safety.

[0062] 3. Real-time warning and automatic adjustment: Combining navigation path planning with real-time position monitoring, the navigation aided deviation warning system of the present invention can evaluate the risk of ship deviation in real time and issue warnings, and even automatically adjust the course when necessary, greatly reducing the risk of human operational errors and improving the level of navigation automation and intelligence.

[0063] 4. Monitoring and data analysis: The monitoring and control module of the present invention not only realizes the real-time monitoring and path adjustment of ship navigation, but also can collect navigation data for subsequent analysis, providing data support for ship navigation safety management and system optimization.

[0064] The present invention provides a comprehensive, intelligent and efficient collision avoidance solution for ships sailing in inland bridge group waters, significantly improving the safety and efficiency of inland navigation. It has important practical value and application prospects for ensuring the safety of ship navigation and promoting the development of inland waterway transportation. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is a mind map of a method for avoiding collisions between ships in inland bridge group waters provided by an embodiment of the present invention;

[0066] Figure 2 It is the region code extraction direction provided by the embodiment of the present invention;

[0067] Figure 3 It is a storage and display process of data information provided by an embodiment of the present invention;

[0068] Figure 4 is a division diagram of the A and B areas provided by an embodiment of the present invention;

[0069] Figure 5 It is the navigation obstacle identification of the electronic river map provided by the embodiment of the present invention;

[0070] Figure 6 It is the navigation condition of a ship in over-expansion narrow waters provided by an embodiment of the present invention;

[0071] Figure 7 is the turning circle at different rudder angles provided by the embodiment of the present invention;

[0072] Figure 8 is a schematic diagram of an optimization process provided by an embodiment of the present invention;

[0073] Fig. 9 It is the optimal safe navigable path in (a) and (b) provided in the embodiment of the present invention; (a) optimizes the approximate navigable path to A; (b) optimizes the approximate navigable path to B;

[0074] Fig.10 It is a GRU recurrent neural network structure diagram provided by an embodiment of the present invention;

[0075] Fig.11 This is a ship track prediction process provided by an embodiment of the present invention;

[0076] Fig.12 The real track, observed track and predicted track of the target track provided by the embodiment of the present invention;

[0077] Fig.13 It is a diagram of the matrix light warning structure provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0078] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0079] 1. Explanatory Examples In order to enable those skilled in the art to fully understand how to implement the present invention, this section provides an illustrative example that expands and describes the technical solution of the claims.

[0080] like Figure 1 As shown, an embodiment of the present invention provides a method for avoiding collisions between ships in inland bridge group waters, the method comprising:

[0081] S1: Environmental modeling based on electronic river map;

[0082] S2: Design of collision avoidance path planning algorithm for bridge groups;

[0083] S3: Design of bridge group navigation aid deviation warning system.

[0084] The S1 specifically includes:

[0085] (1) Extraction and storage of the environment: converting the information in the electronic river map into an environmental model that can be directly used. The representation and storage of the environmental model is the basis for solving the problem of model establishment;

[0086] (2) Determination of dangerous obstacle areas. For high-speed ships, dangerous obstacle areas include obstacles such as land, islands, reefs, and buoys. In the required corresponding layer, the index file (*.shx) can read the storage location of the required target in the master file, and use the obtained index value to read the specific data of the required target in the master file;

[0087] (3) Extraction and storage of environmental information

[0088] ① Extraction of regional information. A region is composed of a closed edge set or a closed connection point. In order to better extract information data representing a region, the vector records constituting the boundary of such a region must be viewed in a certain order. The order must be met. First, the internal boundary is encoded, and then the external boundary is encoded. Before encoding the internal boundary, the external boundary must be completely encoded.

[0089] The direction of the region's encoding extraction:

[0090] like Figure 2 As shown, when encoding a region, its direction also needs to be specified. The outer boundary of the region must be encoded in a clockwise direction, and the boundary inside the region needs to be encoded in a counterclockwise direction.

[0091] ② Extraction of isobath and shoreline information data

[0092] Perform data extraction operations in a given electronic river map data file; in the data file, first read the first logical record, and correctly identify the total number of fields, area identifiers, control field length, and other information in the header area, directory area, control field area, and description field area. After parsing the control field area, you can get the attributes of the class; after parsing the description field area, you can get the control structure of the field. After executing this program, jump to the spatial record, traverse the shoreline and isobath boundary points, determine whether the boundary is divided by the data boundary or is blocked, store the information in an array, and finally output;

[0093] ③Data information storage and display process

[0094] like Figure 3 As shown, the filtered and extracted information is stored, and then the required content is displayed more conveniently and quickly;

[0095] (4) Electronic river map object processing

[0096] like Figure 4 As shown in the figure, an electronic river map of a certain water area is intercepted. In the vicinity of actual islands and land, as the distance of the water area close to the land decreases, the water depth also gradually decreases. In order to enable ships to safely navigate in a certain water depth, the 10m isobath is taken as the collision avoidance limit for ships in the electronic river map. Water areas and land areas within a water depth of 10m are considered dangerous areas. Therefore, in the process of ship collision avoidance, collision avoidance treatment is performed on the 10m isobath to safely avoid collisions in shallow water areas and land areas.

[0097] Extract relevant graphic information such as islands, land, isobaths, buoys, etc. from the electronic river map, and accurately find the longitude and latitude coordinates of any point in the map;

[0098] For the coordinates of dangerous objects such as buoys and reefs (xi ,y i ), i is the center point of the dangerous object, and there is Then the buoy and reef can be described as π2r 2 , r is the radius of dangerous objects with particle characteristics such as buoys and reefs; Figure 5 As shown, all points with light buoys in the electronic river map are detected and marked.

[0099] When avoiding collisions with dangerous objects such as land and islands, it is only necessary to avoid collisions with the 10m isobaths along the perimeter of the object. The isobaths are composed of countless points. In the process of processing them, the coordinates (x m ,y m ), m=1,2,3......n, and there is Then the 10m isobath can be described as a line consisting of (x m ,y m ) as the center, with r m =A 10m isobath composed of circles with a radius of m, so that the ship can regard m circles as static obstacles, and at the same time:

[0100]

[0101] l is the length of the 10m isobath, and when the equal sign is taken, m circles are tangent. Using this method, graphic symbols can be converted into mathematical symbols so that ships can directly read and calculate.

[0102] (5) Determination of narrow water obstacle areas

[0103] For the case of two islands in area A as shown in the figure, if the r of all dangerous objects is set to a fixed value during the navigation of the ship, then for obstacles such as small islands, buoys, reefs, etc. in a narrow waterway with a width of d, the obstacle will expand excessively and the waterway will become blocked and unable to pass (such as Figure 5 shown).

[0104] When a water area is detected to be a narrow area and there are multiple islands, shoals and other areas in the water area, the obstacle area is first calibrated, and then the islands and shoals are rasterized with the help of planning methods to determine the scope of the dangerous area.

[0105] (6) Obstacle Detection Method

[0106] In the electronic river map, 100m along its heading direction is set as the detection base, and the square dynamic window of axa is used as the detection area. When it is beyond 100m, a takes a larger value, and when it is within 100m, a takes a smaller value. And as the distance to the obstacle area shortens, the value of a gradually decreases to adapt to the changes in the environment. When it is sailing on the water surface, when the dynamic window detects dangerous objects such as 10m isobaths, islands, reefs or buoys, the distribution and scale of the surrounding obstacles are detected. Obstacles such as land and buoys in open waters are determined. The circular mark processing environment model is used to determine the value of the mark scale r; when it is determined that obstacles are densely distributed in non-open waters, rasterization environment processing is performed to determine the obstacle area for collision avoidance.

[0107] The S2 specifically includes:

[0108] According to the layered electronic river map, a grid map with a certain accuracy is established; on the established grid map, the dynamic particle objects and the longitude and latitude coordinates of the pier positions are used to construct the Voronoi diagram to form a path set, and the common part of the navigable waters and the path set is selected to establish an approximate navigable path. Its main disadvantage is that there are many turning points in the path, which is not conducive to the maneuvering and navigation of the ship; therefore, it is necessary to optimize the approximate navigable path and construct the optimal safe and time-saving navigable path; in the process of path optimization, the ship's maneuvering motion characteristics and related rule constraints and external environmental interference are considered, and the approximate navigable path with multiple turning points is optimized, and the optimal safe and time-saving navigation path is selected.

[0109] Algorithm design:

[0110] In the layer of the electronic river map that extracts the islands and shoals that hinder navigation, a grid with an accuracy of 150m is established, and the separated layer is rasterized to determine the navigable waters; a Voronoi diagram is established based on the coordinates of particle obstacles such as ships and navigation marks. The risk traversal algorithm is used to select approximate navigable paths. When a ship sails along each navigable path, each path has a certain path risk, which mainly includes the risk of collision of the ship and the risk of navigation overtime. The collision risk refers to the degree of danger of collision between the ship and the obstacle. The overtime risk refers to the risk that the ship's navigation time exceeds the scheduled time to complete the task.

[0111] The collision risk of a ship's navigation path can be calculated using the following formula:

[0112] J threat =(α·DCPA i ) 2 +(β·TCPA i ) 2 (i=1,2,......,n)=

[0113] Among them: J threatis the ship collision risk, DCPA is the minimum encounter distance, TCPA is the minimum encounter time; α and β are weighted values, which are experimental data obtained through statistics.

[0114] Optimization and selection of approximate navigable paths:

[0115] Take the laboratory ship as an example for analysis. Since the planned approximate navigable path has many broken line turns, which is not conducive to the ship's maneuvering and navigation, it is necessary to optimize the broken line turning points. In the optimization process, the motion characteristics of the ship should be considered. The determined approximate navigation path is optimized according to the motion characteristics.

[0116] Table 1 Experimental ship scale parameters

[0117] Ship Type Ships Ship length L(m) 10.08 Ship width B(m) 3.08 Ship speed (kn) 20 K′ 2.32 T′ 2.15

[0118] The simulation experiment was carried out in MATLAB to obtain the position coordinates of the ship during navigation. In the absence of external environmental interference, the turning trajectory of the ship under different rudder angles is as follows Figure 7 shown.

[0119] The optimization process of the ship's navigable path is as follows: Figure 8 , 9 As shown, in a selected curved path, the planned navigation path of the ship is shown as OA-AB-BC. When the ship is sailing in the OA section, a certain time (point M) is selected to output a certain rudder angle (δ1) to turn, and the rudder angle is maintained. The ship performs a turning motion under the action of the rudder, and the turning track line of the ship starting to turn at point M and controlled by the rudder angle 61 can be simulated. Then, the position of point M, the rudder angle δ1 and the sailing speed are changed respectively, and the turning track line of the ship under different turning points, rudder angles δ1 and speeds can be obtained. When the ship is sailing in the BC section, at a certain position point in the feasible turning track line of the first turn that meets the limited parameters, the opposite rudder angle is used to make a second turn to make the ship sail to the planned route BC.

[0120] The S3 specifically includes:

[0121] (1) Ship track prediction

[0122] The ship track prediction method based on recurrent neural network is adopted, which mainly includes two parts: data preprocessing and neural network prediction. In the data preprocessing part, a data subset similar to the predicted target motion characteristics is extracted from the AIS data based on the symmetric segmented path distance to eliminate the influence of a large amount of data redundancy and noise; in the neural network prediction part, considering the characteristics of the neural network model such as nonlinearity, adaptability, self-learning ability and arbitrary function approximation ability, a recurrent neural network model with gated recurrent unit GRU as the core is constructed to achieve accurate and efficient prediction of ship position;

[0123] (2) Simulation experiment

[0124] The measured ship AIS data is used as the original data, and the noise error generated by the simulation software is added to simulate the historical stock AIS trajectory of the ship. The ship trajectory prediction algorithm based on the recurrent neural network is verified in the integrated development environment Matlab2022b, and the actual value, observation value and prediction value are finally obtained;

[0125] (3) Design of navigation aid warning system

[0126] The system is connected to the navigation mark management system. After predicting that the ship is at risk of deviating from the planned route based on the track, the auxiliary warning module of the navigation mark management system starts the warning function. The warning intensity can be determined by itself based on the distance data between the navigation mark and the target ship. If the ship is relatively close to the navigation mark at this time, the voice alarm combined with VHF communication and the flash alarm based on the matrix light will be started at the same time; when the distance is far, only the VHF voice alarm will be started to avoid the sudden flashing of the navigation mark matrix light causing visual impact on other normal sailing ships around.

[0127] The data preprocessing algorithm based on SSPD first converts the latitude and longitude coordinates of the data into plane coordinates through the Gauss-Krüger projection method, and then uses the linear interpolation method to supplement the track data. Then, based on the SSPD method, a large number of historical tracks and the target tracks to be measured are measured for similarity, and the distance between each track is calculated. Finally, the track data are filtered according to the distance to finally obtain a subset of relevant track data that meets the filtering conditions.

[0128] The GRU recurrent neural network prediction model structure is as follows Fig.10 As shown in the figure, it includes 1 input node, 5 hidden layers (GRU layer, Dropout layer and Dense layer) and 1 output node. The length of the input node receiving sequence X is n and the dimension is 2 (i.e. the longitude and latitude coordinates of the ship). Through hidden layer training, the best prediction sequence Y is obtained at the output node, including the longitude and latitude coordinates of the predicted position. The prediction process is as follows Fig.11 shown.

[0129] Simulation experiment:

[0130] The measured ship AIS data is used as the original data, and the noise error generated by the simulation software is added to simulate the historical AIS trajectory of the ship. The ship trajectory prediction algorithm based on the recurrent neural network is verified in the integrated development environment Matlab2022b. The actual value, observed value and predicted value are finally obtained as follows Fig.12As shown in the figure, the actual track of the ship is almost consistent with the observed track and the predicted track, and the prediction effect is good, which is basically not affected by the large-angle maneuver of the ship. Within the threshold range, the difference in the time interval of the point track has little effect on the overall track. It can be seen that the ship track prediction algorithm based on the recurrent neural network has a good prediction effect on the measured AIS data track. In terms of time, the program

[0131] The voice alarm of the VHF communication specifically includes: issuing warning instructions according to the bridge group area where the ship is located, and the navigation mark near the waters where the target ship is located broadcasts warning information through sound, prompting the ship to deviate from the planned path or have a risk of collision with the bridge group;

[0132] like Fig.13 As shown, the flash alarm of the matrix light specifically includes: the navigation mark management system makes a warning instruction according to the warning area where the ship with deviation risk is located, and sends the flash alarm instruction to the warning navigation mark through the wireless transmission module. After the navigation mark data receiving module receives the warning instruction, it immediately activates the matrix light to flash the dangerous ship.

[0133] The embodiment of the present invention provides a system for avoiding collisions between ships in inland river bridge group waters for implementing the method for avoiding collisions between ships in inland river bridge group waters, the system comprising:

[0134] Modeling module, based on electronic Jiangtu environment modeling;

[0135] Algorithm design module, used for bridge group collision avoidance path planning algorithm design;

[0136] System design module, used for the design of bridge group navigation aid yaw warning system.

[0137] An embodiment of the present invention provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for avoiding collisions between ships navigating in waters of inland bridge groups.

[0138] An embodiment of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps of the method for avoiding collisions between ships navigating in waters of an inland bridge group.

[0139] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0140] The specific implementation scheme involves integrating the above systems and related algorithms to form a complete set of ship navigation collision avoidance system in inland bridge waters. The following are the detailed implementation steps combined with the specific algorithm:

[0141] 1. Implementation of environment modeling module:

[0142] - Use Geographic Information System (GIS) technology to analyze electronic river maps and extract environmental features, including the location and geometry of obstacles such as land, islands, bridges, reefs, buoys, and river width, water depth and other data.

[0143] -Convert obstacle and water depth information into grid information and establish an environmental grid model, in which the grid points where obstacles are located are set to an impassable state and areas with shallow water depths are set to a high-risk state.

[0144] - Receive hydrological and meteorological data in real time and dynamically update environmental models to reflect changes in water levels and other environmental variables.

[0145] 2. Bridge group collision avoidance path planning algorithm design:

[0146] - Apply the A* algorithm for path planning. Based on the environmental grid model, define a heuristic function, comprehensively consider the path length, collision avoidance safety distance and navigation cost, and give priority to safe and efficient paths.

[0147] - Consider ship dynamics to ensure the planned path is feasible in terms of turning radius and ship speed.

[0148] -Add multi-vessel collision avoidance strategy, obtain other vessel information through VHF communication or AIS system, and plan navigation paths to avoid mutual interference.

[0149] 3. Design of navigation aid deviation warning system:

[0150] - Combine the ship's position information (such as GPS data) with the planned path to calculate the ship's deviation in real time.

[0151] - Set the yaw threshold. When the real-time calculated yaw exceeds the threshold, an alarm will be issued through sound or visual signal.

[0152] -Design warning signals of different urgency for different levels of deviation, such as a warning tone for minor deviation and an alarm and suggestion or automatic adjustment of course for severe deviation.

[0153] 4. Monitoring and control module:

[0154] -Integrated environmental modeling, path planning and early warning systems, real-time display of ship and environmental status, and provision of an operating interface for crew members to monitor and control navigation.

[0155] - Realize the interface with the ship navigation system, support manual and automatic mode switching, and automatically adjust the ship's heading according to the early warning system and path planning results in automatic mode.

[0156] -Collect navigation data, including navigation path, time, collision avoidance events, etc., for subsequent analysis and system optimization.

[0157] Through the above implementation scheme, a ship navigation collision avoidance system in inland bridge waters can be built, which integrates environmental perception, intelligent path planning, real-time warning and ship control, and improves the safety and efficiency of inland navigation.

[0158] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for avoiding collisions between ships in inland bridge group waters, characterized in that: The method includes: S1: Environmental modeling based on electronic river map; S2: Design of collision avoidance path planning algorithm for bridge groups; S3: Design of bridge group navigation aid deviation warning system.

2. The method for avoiding collision of ships in inland bridge group waters as claimed in claim 1, characterized in that: The S1 specifically includes: (1) Extraction and storage of the environment: converting the information in the electronic river map into an environmental model that can be directly used. The representation and storage of the environmental model is the basis for solving the problem of model establishment; (2) Determination of dangerous obstacle areas. For high-speed ships, dangerous obstacle areas include land, islands, reefs, and buoy obstacles. In the required corresponding layer, the index file (*.shx) can read the storage location of the required target in the master file, and use the obtained index value to read the specific data of the required target in the master file; (3) Extraction and storage of environmental information ① Extraction of regional information. A region is composed of a closed edge set or a closed connection point. In order to better extract information data representing a region, the vector records constituting the boundary of such a region must be viewed in a certain order. The order must be met. First, the internal boundary is encoded, and then the external boundary is encoded. Before encoding the internal boundary, the external boundary must be completely encoded. When encoding a region, its direction also needs to be specified. The outer boundary of the region must be encoded in a clockwise direction, and the boundary inside the region needs to be encoded in a counterclockwise direction. ② Extraction of isobath and shoreline information data Perform data extraction operations in a given electronic river map data file; in the data file, first read the first logical record, and correctly identify the total number of fields, area identifiers, control field length, and other information in the header area, directory area, control field area, and description field area. After parsing the control field area, you can get the attributes of the class; after parsing the description field area, you can get the control structure of the field. After executing this program, jump to the spatial record, traverse the shoreline and isobath boundary points, determine whether the boundary is divided by the data boundary or is blocked, store the information in an array, and finally output; ③Data information storage and display process Store the filtered and extracted information, and then display the required content more conveniently and quickly; (4) Electronic river map object processing Take the electronic river map of a certain water area, and take the 10m isobath in the electronic river map as the collision avoidance boundary for ships. Waters and land areas within a water depth of 10m are considered dangerous areas. Therefore, in the process of ship collision avoidance, collision avoidance treatment is performed on the 10m isobath to safely avoid collisions in shallow water areas and land areas. Extract relevant islands, land, isobaths, and buoy graphic information from the electronic river map, where the longitude and latitude coordinates of any point in the graphic can be accurately found; For the coordinates of dangerous objects such as buoys and reefs (x i ,y i ), i is the center point of the dangerous object, and there is Then the buoy and reef can be described as π2r 2 , r is the radius of dangerous objects with particle characteristics such as buoys and reefs; When avoiding collisions with dangerous objects such as land and islands, it is only necessary to avoid collisions with the 10m isobaths along the perimeter of the object. The isobaths are composed of countless points. In the process of processing them, the coordinates (x m ,y m ), m=1,2,3......n, and there is Then the 10m isobath can be described as a line consisting of (x m ,y m ) as the center, with r m =A 10m isobath composed of circles with a radius of m, so that the ship can regard m circles as static obstacles, and at the same time: l is the length of the 10m isobath, and when the equal sign is taken, m circles are tangent. Using this method, graphic symbols can be converted into mathematical symbols so that ships can directly read and calculate. (5) Determination of narrow water obstacle areas When a water area is detected to be a narrow area and there are multiple islands, shoals and other areas in the water area, the obstacle area is first calibrated, and then the islands and shoals are rasterized with the help of planning methods to determine the scope of the dangerous area. (6) Obstacle Detection Method In the electronic river map, 100m along its heading direction is set as the detection base, and the square dynamic window of axa is used as the detection area. When it is beyond 100m, a takes a larger value, and when it is within 100m, a takes a smaller value. And as the distance to the obstacle area shortens, the value of a gradually decreases to adapt to environmental changes. When it is sailing on the water surface, when the dynamic window detects the 10m isobath, island, reef or buoy dangerous objects, the distribution and scale of the surrounding obstacles are detected. The land and buoy obstacles in the open waters are determined by using the circular marking processing environment model and determining the value of the marking scale r; when it is determined that the obstacles in the non-open waters are densely distributed, the grid environment processing is carried out to determine the obstacle area for collision avoidance.

3. The method for avoiding collision of ships in inland bridge group waters as claimed in claim 1, characterized in that: The S2 specifically includes: According to the layered electronic river map, a grid map with a certain accuracy is established; on the established grid map, the latitude and longitude coordinates of the dynamic particle objects and the pier positions are used to construct the Voronoi diagram to form a path set, and the common part of the navigable waters and the path set is selected to establish an approximate navigable path. Its main disadvantage is that there are many turning points in the path, which is not conducive to the manipulation and navigation of the ship; therefore, it is necessary to optimize the approximate navigable path and construct the optimal safe and time-saving navigable path; in the process of path optimization, the ship's maneuvering motion characteristics and related rule constraints and external environmental interference are considered, the approximate navigable path with multiple turning points is optimized, and the optimal safe and time-saving navigation path is selected.

4. The method for avoiding collision of ships in inland bridge group waters as claimed in claim 1, characterized in that: The S3 specifically includes: The ship track prediction method based on recurrent neural network is adopted, which mainly includes two parts: data preprocessing and neural network prediction. In the data preprocessing part, a data subset similar to the predicted target motion characteristics is extracted from the AIS data based on the symmetric segmented path distance to eliminate the influence of a large amount of data redundancy and noise; in the neural network prediction part, considering the characteristics of the neural network model such as nonlinearity, adaptability, self-learning ability and arbitrary function approximation ability, a recurrent neural network model with gated recurrent unit GRU as the core is constructed to achieve accurate and efficient prediction of ship position; The measured ship AIS data is used as the original data, and the noise error generated by the simulation software is added to simulate the historical stock AIS trajectory of the ship. The ship trajectory prediction algorithm based on the recurrent neural network is verified in the integrated development environment Matlab2022b, and the actual value, observation value and prediction value are finally obtained; The system is connected to the navigation mark management system. After predicting that the ship is at risk of deviating from the planned route based on the track, the auxiliary warning module of the navigation mark management system starts the warning function. The warning intensity can be determined by itself based on the distance data between the navigation mark and the target ship. If the ship is relatively close to the navigation mark at this time, the voice alarm combined with VHF communication and the flash alarm based on the matrix light will be started at the same time; when the distance is far, only the VHF voice alarm will be started to avoid the sudden flashing of the navigation mark matrix light causing visual impact on other normal sailing ships around.

5. The method for avoiding collision of ships in inland bridge group waters as claimed in claim 4, characterized in that: The data preprocessing algorithm based on SSPD first converts the latitude and longitude coordinates of the data into plane coordinates through the Gauss-Krüger projection method, and then uses the linear interpolation method to supplement the track data. Then, based on the SSPD method, a large number of historical tracks and the target tracks to be measured are measured for similarity, and the distance between each track is calculated. Finally, the track data are filtered according to the distance to finally obtain a subset of relevant track data that meets the filtering conditions.

6. The method for avoiding collisions between ships in inland bridge group waters as claimed in claim 4, characterized in that: The GRU recurrent neural network prediction model structure includes 1 input node, 5 hidden layers and 1 output node. The length of the input node receiving sequence X is n and the dimension is 2. Through hidden layer training, the best prediction sequence Y is obtained at the output node, including the latitude and longitude coordinates of the prediction position.

7. The method for avoiding collisions between ships in inland bridge group waters as claimed in claim 4, characterized in that: The voice alarm of the VHF communication specifically includes: issuing warning instructions according to the bridge group area where the ship is located, and the navigation mark near the waters where the target ship is located broadcasts warning information through sound, prompting the ship to deviate from the planned path or have a risk of collision with the bridge group; The flash alarm of the matrix light specifically includes: the navigation mark management system makes a warning instruction according to the warning area where the ship with deviation risk is located, and sends the flash alarm instruction to the warning navigation mark through the wireless transmission module. After the navigation mark data receiving module receives the warning instruction, it immediately activates the matrix light to flash the dangerous ship.

8. A system for avoiding collisions between ships in inland river bridge group waters that implements the method for avoiding collisions between ships in inland river bridge group waters as claimed in any one of claims 1 to 7, characterized in that: The system includes: a) Environmental modeling module, used to extract and store navigation environment information based on electronic river charts, including obstacles such as land, islands, reefs, buoys, and their location and size information; b) Path planning module, which designs a bridge group collision avoidance path planning algorithm based on the environmental model and the current position, heading and speed of the ship, and generates a safe navigation path; c) The navigation mark auxiliary deviation warning module provides deviation warning information based on the real-time position and planned path of the ship to assist the ship in collision avoidance operations; d) Monitoring and control module, which integrates the functions of the above modules, monitors the ship status and navigation environment in real time, controls the ship to execute the collision avoidance path, and makes adjustments to cope with the dynamically changing environment.

9. The ship navigation collision avoidance system in the inland bridge group waters as claimed in claim 8, characterized in that: The environment modeling module is further capable of: a) Automatically identify and classify different types of obstacles, such as land, islands, reefs, etc., by analyzing the vector data in the electronic river map, and record their locations and geometric features; b) Based on the water depth information in the electronic river map, define the water area where ships can safely navigate, and mark areas with water depths less than a certain value as dangerous areas; c) Update the environmental model in real time to reflect environmental changes, such as tidal changes, river water level changes, etc., to ensure the accuracy of navigation path planning.

10. The ship navigation collision avoidance system in the inland bridge group waters as claimed in claim 8, characterized in that: The path planning module is further capable of: a) Combine the dynamic information of the ship and the environmental model, and use advanced algorithms such as genetic algorithm, A* algorithm or Dijkstra algorithm to generate the optimal navigation path; b) Consider the ship’s navigation speed, turning radius and other navigation characteristics to ensure the feasibility and safety of the planned path; c) Consider the information of other ships in navigation path planning to achieve collaborative collision avoidance among multiple ships; The navigation aid deviation warning module can further: a) Dynamically provide deviation warning information based on the real-time position difference between the ship and the planned path; b) Automatically trigger an audible or visual alarm when the ship is about to deviate from the safe navigation path, prompting the crew to adjust the course in time; c) It can interface with the ship’s automatic navigation system to achieve automatic collision avoidance control and automatically adjust the ship’s course to avoid collision in an emergency.

Images