Ship climate route planning method and system based on graph theory

Abstract

The invention discloses a ship climate route planning method and system based on a graph theory, and the method comprises the following steps: constructing a navigation network of global ships based on global port information, geographic attributes, a traffic flow statistical method and a graph theory; cleaning the AIS trajectory data of each ship and extracting key node information to form a one-dimensional node sequence; according to the one-dimensional node sequence, a navigation network is constructed, and node attributes of node features related to climate conditions, weights of two node edges, maximum draught and width of ships allowed to pass through and the like are included; constructing an intermediate node sequence of the navigation network based on information such as node attributes; from a starting port to a destination port, an optimal node sequence obtained by taking maximum weight and meeting draught and ship width constraints as search conditions and combining node characteristics related to climate conditions is taken as a ship climate route. The method is simple and efficient, the climate route can be accurately and efficiently planned, and the ship characteristics and the climate of the route are fully considered.

Description

Technical Field

[0001] The present invention relates to the technical field of route planning, and particularly to a ship climate route planning method and system based on graph theory. Background Art

[0002] In the existing ship climate route design, crew members need to consult reference materials such as sailing directions, catalog of nautical charts, and world ocean routes before departure, and combine the draft, speed, hydrometeorology, navigation restricted areas, and ship routing systems of the ship to formulate a reasonable route. Since it is difficult to fully consider dynamic factors such as climate and ocean currents when manually drawing a route, some shipping companies build a historical route library by saving the historical routes of all ships under their company to assist in the design. When designing a route, crew members can select the route that best matches the current sailing task from the historical route library, and then check whether the selected route needs to be further corrected according to factors such as temporary notices in the nautical notice. From the result perspective, the historical route library integrates the navigation experience of various types of ships in different seas in history, improves the speed of crew members in route design, and makes full use of climate routes including regional and seasonal factors. However, from the data perspective, the construction of the historical route library still requires a large amount of data as a basis, and requires reasonable data cleaning strategies and attribute constraint strategies, and the update cycle of sailing directions is as long as three years, so the data timeliness is poor.

[0003] Currently, there are the following problems in the existing ship climate route design: 1. Many human interferences: The route is manually designed by crew members, and it is inevitable that there are human interference factors, and it is difficult to design a route that conforms to the characteristics of climate and ocean currents by following manual experience; 2. Poor data quality: The construction of the existing historical route library requires cleaning and removing abnormal trajectory points (such as special trajectory points such as ship maneuvers and drifting), and replacing abnormal trajectory points with reasonable trajectory points before it can be used as a general historical route for other ships, and the update cycle of the dependent sailing directions is too long, so the data timeliness is poor; 3. High complexity of the design process: During the design, it is necessary to cross-verify paper and electronic, text and graphic data, and there are many reference materials with different data dimensions, which pose a high threshold and great difficulty for crew members to design a route. Summary of the Invention

[0004] In order to solve the problems of many human interferences, poor data quality, high complexity of the design process, etc. existing in the existing ship climate route planning method, the present invention provides a ship climate route planning method based on graph theory, which can accurately and efficiently plan the climate route of a ship, fully consider the ship characteristics and the climate of the route, and provide safety guarantee for ship navigation. The present invention also relates to a ship climate route planning system based on graph theory.

[0005] The present invention is realized by the following technical solutions:

[0006] A method for ship climate route planning based on graph theory, comprising the following steps:

[0007] Steps for constructing a global ship navigation network: Based on global port information, geographical attributes, traffic flow statistics methods, and graph theory, extract the location information of all sea areas, and construct a navigation network for global ships. The navigation network includes nodes that divide several sea areas and cover key waypoints and important turning points, as well as edges representing the connection relationships between the nodes;

[0008] Steps for processing global ship AIS data: Based on the navigation network, clean the AIS trajectory data of each ship globally, and use the PostGIS geographic information engine to extract the key node information in the ship navigation trajectory from the cleaned AIS trajectory data of each ship to form a one-dimensional node sequence. The one-dimensional node sequence includes the nodes of the navigation network passed by the ship and the corresponding navigation characteristics;

[0009] Steps for constructing network node attributes: Construct the node attributes of the navigation network, including: counting the frequency, month, draft, and basic ship information of ships passing through each node in the navigation characteristics, calculating the conditional probability of a ship passing through a certain node and its next node, and using this probability as the weight of the edge between two nodes in the network; Using the kernel density estimation method, based on the draft and basic ship information when the ship passes through each node, statistically calculate the maximum draft and maximum width of the ships allowed to pass through each node; Based on the month when the ship passes through each node in the navigation characteristics, statistically calculate the ship passing month data of each node, and analyze the adaptive impact of climate conditions on the node navigation capacity to form node characteristics related to climate conditions;

[0010] Steps for constructing an intermediate node sequence: Based on the starting port, destination port, and each node attribute, construct an intermediate node sequence of the navigation network, and filter the massive graph data to within the order of thousands through hash mapping. The intermediate node sequence includes potential nodes from the starting port to the destination port;

[0011] Steps for ship climate route planning: Use a recursive algorithm to perform recursive search in the intermediate node sequence. Starting from the starting port, with the search conditions of maximizing the weight sum and satisfying the draft and ship width constraints, and combining the node characteristics related to climate conditions, search to the destination port to obtain the optimal node sequence as the ship climate route.

[0012] Preferably, in the steps for constructing a global ship navigation network, the steps for constructing the navigation network of global ships include:

[0013] Based on global port information and geographical attributes, extract the location information of all straits, canals, and oceans;

[0014] Based on the extracted location information of all straits, canals and oceans and the global hot route map, determine the large nodes that divide the oceans or waters, and the large nodes pass through all the hot routes passing through the oceans or waters;

[0015] Based on the large nodes, small nodes describing key waypoints and important turning points are selected using traffic flow statistics methods and graph theory, and the small nodes are subsets of the large nodes;

[0016] Based on the large nodes and the small nodes, a global ship navigation network is constructed.

[0017] Preferably, in the global ship AIS data processing step, the step of cleaning the AIS track data of each ship in the world based on the navigation network, and extracting key node information in the ship navigation track from each cleaned AIS track data using the PostGIS geographic information engine to form a one-dimensional node sequence includes:

[0018] The AIS trajectory data of all ships in the world are cleaned by deduplication, invalid point filtering, speed abnormal point filtering, and drift point filtering to obtain the first AIS trajectory data;

[0019] According to the AIS drift rate, the AIS loss rate and the ship information validity index requirements, an AIS track that meets the requirements is selected from the first AIS track data to obtain the second AIS track data;

[0020] The second AIS trajectory data are grouped based on route similarity, and routes with navigation frequencies lower than a preset threshold in each group are eliminated to obtain third AIS trajectory data;

[0021] By using the PostGIS geographic information engine technology in the PostgreSQL database, the trajectory points that do not belong to the small nodes of the navigation network in the third AIS trajectory data after each cleaning process are eliminated, and then the key node information in the ship's navigation trajectory is extracted to obtain the corresponding one-dimensional node sequence.

[0022] Preferably, in the step of constructing the network node attributes, the frequency, month, draft and basic information of the ships passing through each node in the navigation characteristics are counted, and the conditional probability of the ship passing through a node and its next node is calculated using the Bayesian formula.

[0023] Preferably, in the step of constructing the network node attributes, when calculating the conditional probability of a ship passing through a certain node and its next node, the ship type factor is also combined. The ship types are divided into different categories. For different types of ships, the frequencies, months, and drafts of passing through each node and its next node are respectively counted. Combining the node characteristics related to the climate conditions corresponding to different types of ships, the conditional probabilities of different types of ships passing through a certain node and its next node are respectively calculated, and a node attribute table exclusive to the ship type is generated and stored in the node attributes. When planning the route, the corresponding conditional probability is selected as the weight of the edge according to the actual ship type.

[0024] Preferably, in the step of constructing the intermediate node sequence, the construction step of the intermediate node sequence of the navigation network includes:

[0025] Determine the starting port and the destination port;

[0026] In the navigation network, obtain the large node sequence passed by the route with the shortest voyage as the goal;

[0027] According to the large node sequence, list the potential small node sequence pairs between adjacent large nodes;

[0028] Using the hash algorithm, based on all the small node sequence pairs, with the starting port, destination port, and month as the hash keys, and the edge weight constraint, climate condition constraint, and draft constraint as the hash values, map the hash keys to a storage space with a fixed length through the hash function for constraint filtering, and filter the massive graph data to within the order of thousands to construct the intermediate node sequence of the navigation network.

[0029] Preferably, after the ship climate route planning step, there is also a ship climate route visualization step: Visualize the determined ship climate route through the geographic information system so that the crew can clearly see the planning of the ship climate route.

[0030] A ship climate route planning system based on graph theory, including a global ship navigation network construction module, a global ship AIS data processing module, a network node attribute construction module, an intermediate node sequence construction module, and a ship climate route planning module connected in sequence, where,

[0031] The navigation network construction module is used to extract the location information of all sea areas based on the global port information, geographical attributes, traffic flow statistics methods, and graph theory, and construct the navigation network of global ships. The navigation network includes nodes that divide several sea areas and cover key waypoints and important turning points, and edges representing the connection relationships between the nodes;

[0032] The ship AIS data processing module is used to clean the AIS trajectory data of ships around the world based on the navigation network, and use the PostGIS geographic information engine to extract the key node information in the ship navigation trajectory from the cleaned AIS trajectory data, forming a one-dimensional node sequence. The one-dimensional node sequence includes the nodes of the navigation network passed by the ship and the corresponding navigation characteristics;

[0033] The network node attribute construction module is used to construct the node attributes of the navigation network, including: counting the frequency, month, draft and basic ship information of ships passing through each node in the navigation characteristics, calculating the conditional probability of a ship passing through a certain node and its next node, and using this probability as the weight of the edge between two nodes in the network; using the kernel density estimation method, based on the draft and basic ship information when the ship passes through each node, statistically calculate the maximum draft and maximum width of the ships allowed to pass through each node; based on the month when the ship passes through each node in the navigation characteristics, statistically calculate the ship passing month data of each node, and analyze the adaptive impact of climate conditions on the node navigation capacity, forming node characteristics related to climate conditions;

[0034] The intermediate node sequence construction module is used to construct the intermediate node sequence of the navigation network based on the starting port, destination port and each node attribute, and filter the massive graph data to within the order of thousands through hash mapping. The intermediate node sequence includes the potential nodes from the starting port to the destination port;

[0035] The ship climate route planning module is used to perform recursive search in the intermediate node sequence using a recursive algorithm. Starting from the starting port, with the maximization of the weight sum and meeting the draft and ship width constraints as the search conditions, combined with the node characteristics related to climate conditions, search to the destination port to obtain the optimal node sequence as the ship climate route.

[0036] Preferably, in the global ship navigation network construction module, constructing the navigation network of global ships includes:

[0037] Based on the global port information and geographical attributes, extract the location information of all straits, canals and oceans;

[0038] Based on the extracted location information of all straits, canals and oceans and the global popular route map, determine the large nodes that divide the ocean or water area. The large nodes pass through all the popular routes passing through this ocean or water area;

[0039] Based on the large nodes, use the traffic flow statistics method and graph theory to select small nodes that describe the key waypoints and important turning points. The small nodes are subsets of the large nodes;

[0040] Based on the large nodes and the small nodes, construct the navigation network of global ships.

[0041] Preferably, it further includes a ship climate route visualization module, which is connected to the ship climate route planning module and is used to visualize the determined ship climate route through a geographic information system after the ship climate route is planned, so that the crew can clearly see the planning of the ship climate route.

[0042] The beneficial effects of the present invention are as follows:

[0043] The present invention provides a ship climate route planning method based on graph theory. This method constructs a global ship navigation network, which includes nodes that divide the sea areas and cover key waypoints and important turning points, as well as edges representing the connection relationships between the nodes. The nodes in this navigation network maintain a certain spacing, which can reduce the computational complexity. More nodes are set in busy route areas, and the nodes are mainly set at key waypoints and turning points, which can reflect the main characteristics of the global ship routes, form the topological structure of the global navigation path, and provide a standardized basis for subsequent data analysis. Using graph theory, the sea areas (large nodes) and detailed routes (small nodes) are divided to achieve dynamic segmentation optimization, ensuring that the network can not only reflect the macroscopic sea area distribution but also capture local navigation details (such as turning points), converting the complex ocean space into a computable graph structure and providing a clear framework for subsequent path search. Nodes are set in areas such as straits and canals that are prone to congestion or turning to ensure that the route planning covers the core paths of ship navigation and enables the network to naturally adapt to the actual ship traffic distribution, realizing traffic flow integration and enhancing the practicality of path planning; Based on the navigation network, the global ship AIS trajectory data is cleaned and extracted to form a one-dimensional node sequence. The one-dimensional node sequence includes the nodes of the navigation network passed by the ship and the corresponding navigation characteristics. The one-dimensional node sequence obtained after processing the ship AIS trajectory data records which nodes the ship has passed through in the global ship navigation network in sequence, facilitating the association between the ship AIS trajectory and the navigation network, and greatly reducing the noise of the original AIS trajectory data, refining highly reliable navigation data, thereby improving the data quality. Using the PostGIS engine to map the continuous trajectory into a discrete node sequence, realizing data dimensionality reduction, reducing computational redundancy, simplifying complex navigation paths, facilitating statistical analysis and algorithm processing, and by extracting the historical laws (such as high-frequency paths) of the ship passing through the nodes, converting the crew's experience into quantifiable data, providing a basis for weight calculation; Then, the node attributes of the navigation network are constructed. Based on the analysis of the months when ships pass through each node, the adaptability impact of climate conditions on the navigability of the nodes is analyzed to form node characteristics related to climate conditions, and the attribute information of each node in the navigation network and the weight value of the edge between two nodes are determined. The attribute information includes the month when each ship passes through the node, the draft and ship type corresponding to each ship, and the frequency of a ship passing through a certain node and the next node. It can be understood as the conditional probability that the next node appears when the current node appears in the ship's node sequence data. Using this conditional probability as the weight value of the edge between two nodes in the navigation network can reflect the potential navigation route of the ship. Binding climate data (such as ocean currents, monsoons, etc.) to the nodes realizes dynamic climate integration, enabling the path planning to respond to climate changes (such as downstream gain). Combining the node characteristics related to climate conditions, scientific weight calculation is carried out based on the conditional probability (historical transition frequency) to ensure that the path selection not only conforms to historical experience but also adapts to the climate environment;From the draft and ship type corresponding to each ship in the node attribute information, determine the maximum draft and maximum width of the ships allowed to pass through each node. In this way, the maximum draft and the maximum width of the ships allowed to pass through, which are statistically obtained from the historical route information, are used as safety thresholds to provide physical constraints, enabling the judgment of the safety of the target ship passing through each node and reducing navigation risks. Based on the starting port, destination port, and the attributes of each node, construct the intermediate node sequence of the navigation network. The intermediate node sequence includes the potential nodes from the starting port to the destination port. This intermediate node sequence can filter the graph data in the order of tens of millions to within the order of thousands, greatly reducing the computational complexity and the difficulty of data processing, achieving millisecond-level response, improving the efficiency of climate navigation planning, and filtering out the nodes with low navigation safety at the current time through the month in the node attributes, thus improving the safety of the route. Starting from the starting port and ending at the destination port, search the intermediate node sequence with the maximum weight value and meeting the draft constraint as the search conditions to obtain the optimal node sequence as the ship's climate route. In this way, with the maximum weight and meeting the draft constraint as the search conditions, the climate route searched in the intermediate node sequence effectively ensures the safety and accuracy of the planned climate route. The recursive algorithm combines the node characteristics related to the climate conditions to ensure that the route adapts to the current ocean currents and other climates. With the maximum weight as the goal, while meeting the draft and width constraints, taking into account both economy and safety, achieving a multi-objective balance. This route conforms to the historical crew experience rules and ship safety constraints, and also provides a benchmark and experience for ship meteorological route planning. Crew members can adjust the specific navigation path according to the meteorological dynamics based on the climate route. By constructing a navigation network, processing ship AIS data, determining the attributes of the navigation network, statistically obtaining the maximum draft and width of the ships allowed to pass through each node of the navigation network, constructing the intermediate node sequence of the navigation network, and searching the intermediate node sequence, combined with efficient algorithm designs such as graph theory, PostGIS geographic information engine technology, climate data fusion, kernel density estimation method, hash function, recursive algorithm, etc., to obtain the optimal node sequence as the ship's climate route and other methods, it is possible to accurately, real-time, safely, economically, and efficiently plan the climate route of the target ship, solve the core problems of traditional reliance on manual experience, poor data quality, and lagging response, fully consider the ship characteristics and the climate of the route, provide safety guarantees for ship navigation, and provide innovative technical support for intelligent shipping.;

[0044] Through the technical solution of the present invention, it is possible to dynamically adjust the positions of network nodes; achieve millisecond-level real-time route planning under tens of billions of AIS data; and superimpose network node attributes in any dimension, including safety, area, crew experience, and climate factors. These attributes include the safe draft of ships passing through nodes, navigation restricted areas, the types of ships most frequently passing through nodes, and the impact of ocean currents on speed when passing through nodes. The present invention simplifies the key parameters required for route design, such as ship type, ship draft, start and end ports of the route, estimated departure time, and estimated arrival time, and designs specific climate routes based on these key parameters to avoid human interference. At the same time, based on the global ship navigation network, it efficiently counts the historical navigation data of ships, abstracts the navigation experience of crew members into climate routes, automatically cleans the dynamic data of ship navigation, monitors and extracts the navigation trajectories of each ship on the ocean, improves the data quality while reducing the dimension of the data; integrates all the data required for traditional route design, including climate and ocean current data, ship data, nautical chart data, and navigation notice information, no longer relying on various forms of original data such as nautical charts, messages, and weather faxes, and solidifies the route planning process, reducing the learning cost and execution difficulty of route design tasks.

[0045] Therefore, the ship climate route planning method based on graph theory of the present invention has the following advantages: accuracy: by constructing a global navigation network and combining global port information, it can accurately and real-time plan the optimal navigation path; efficiency: by using the shortest voyage objective to screen large nodes and combining a recursive algorithm to search for the optimal climate route, the planning efficiency is greatly improved; safety: according to the maximum draft and ship information that each node can accommodate, the safety of ship navigation is effectively ensured; practicality: it provides a fast, accurate, and convenient route planning method, reducing human intervention and improving the planning quality; economy: by optimizing the navigation path, fuel consumption is saved and operating costs are reduced; environmental protection: the optimized navigation path can effectively reduce carbon emissions, which is beneficial to environmental protection.

[0046] Due to the relatively vague spatial constraints of maritime ship navigation, and from a geographical perspective, the ocean is mainly separated by island groups or continents, which are often regarded as key nodes in the navigation network. Therefore, by combining global port information and geographical attributes, using traffic flow statistical methods and graph theory, large nodes are set for all straits, canals, and narrow channels. And based on the east, central, west or north, central, south, combined with all popular shipping lanes, large nodes are set to divide each ocean into three regions. The large nodes ensure all popular voyages passing through the regions where they are located. Based on the large nodes, small nodes are set according to the key waypoints and important turning points of the popular voyages, which can reflect the main characteristics of the global ship routes. On the global ship navigation network, a large node sequence is obtained using the shortest voyage objective. The large nodes, as a tool for filtering redundant data, provide a macroscopic navigation path. While the small nodes are used to recursively search for climate routes and provide a more refined navigation path. After obtaining the large node sequence, according to the subset of small nodes in the large node sequence, an intermediate table (i.e., the intermediate node sequence) is constructed to describe the potential small node sequence. This intermediate table filters a large amount of graph data to a more manageable order of magnitude.

[0047] According to indicators such as the AIS drift rate, AIS loss rate, and the effectiveness of ship information in each global AIS trajectory data, high-quality AIS trajectories that meet the requirements are selected from the global AIS trajectory data. Then, the selected AIS trajectories are subjected to cleaning processes including duplicate removal, invalid point filtering, speed anomaly point filtering, drift point filtering, etc., ensuring the reliability and accuracy of the selected AIS trajectories. Using the PostGIS geographic information engine technology in the PostgreSQL database, the trajectory points that do not belong to the small nodes of the navigation network are removed from each high-quality AIS trajectory after the cleaning process, so that the processed one-dimensional node sequence only records which nodes the ship has passed through in the global ship navigation network in sequence, facilitating the association between the ship AIS trajectory and the navigation network and the node data statistics.

[0048] Based on the situation of each one-dimensional node sequence passing through each node in the navigation network, the attribute information of each small node is determined, including the ship type, month, draft, etc. of the ships passing through the small node. Based on the month, it can be analyzed which months of the year are safe to pass through the small node. Based on the longitude and latitude, the location of the small node can be judged. Based on the draft and ship type of each ship, the maximum draft and maximum width allowed to pass through the small node can be statistically obtained. The frequency of a ship passing through a certain node and its next node can be understood as the conditional probability of the appearance of the latter node when the current node appears in the node sequence data of the ship passing through. And taking this conditional probability as the weight value of the edge between two nodes in the navigation network can reflect the potential navigation route of the ship, and then the preferred shipping route of the ship can be selected.

[0049] The present invention generates a potential node sequence based on the starting port, destination port, and month, narrowing the recursive search scope and improving the algorithm efficiency. In the navigation network, a large node sequence passed by the route is obtained with the shortest voyage as the goal. Then, based on the large node sequence, potential small node sequence pairs between adjacent large nodes are listed. Using the hash algorithm, an intermediate node sequence of the navigation network is constructed with all the small node sequence pairs. In this way, determining the large node sequence can filter redundant data, improve the efficiency of route planning, filter graph data in the tens of millions to within the thousands level, greatly reducing the difficulty of route planning. The design of the hash key (starting port - destination port - month) and the hash value (weight constraint, climate constraint, draft constraint) has dynamic adaptability, supports fast matching of multi-dimensional conditions, and the hash algorithm can keep the data unique and ensure efficient querying.

[0050] The present invention uses the kernel density estimation method to statistically obtain the maximum draft and maximum width data of ships allowed to pass through each node. Through the probability density function estimated by the kernel density estimation method and combined with the specified allowable values of the ship, the draft value and ship type width value corresponding to the specified quantile can be calculated as the lower limit of the safety threshold, that is, the requirements for the maximum draft and maximum width data of the node. This method is simple and efficient and can accurately calculate the maximum draft and maximum width data of the node. By statistically recording the draft and width of ships when passing through each node, the risk of grounding or collision caused by the ship being too wide or having too large a draft can be avoided, which is of crucial significance for the safety of route planning.

[0051] The present invention can provide visual display, presenting the planned climate route in a visual form in the geographic information system, enabling the crew to clearly see the route planning and the real-time status of navigation, providing real-time, accurate, and effective reference information for the crew to assist in decision-making.

[0052] The present invention also relates to a ship climate route planning system based on graph theory. This system corresponds to the above-mentioned ship climate route planning method based on graph theory and can be understood as a system that implements the above-mentioned ship climate route planning method based on graph theory. It includes a global ship navigation network construction module, a global ship AIS data processing module, a network node attribute construction module, an intermediate node sequence construction module, and a ship climate route planning module. Each module works collaboratively. Through the synergistic effects of hierarchical network construction, data cleaning, weight generation, hash filtering, and recursive search, it has high real-time performance and dynamic adaptability, data-driven high-precision planning (the deep integration of historical AIS data and climate data enables path selection to inherit the experience of crew members and scientifically quantify the impact of climate, improving planning accuracy), dual improvements in safety and economy, high-efficiency calculation and scalability (hash mapping and intermediate table technology reduce the computational complexity from exponential level to linear level, supporting millisecond-level responses for networks with thousands of nodes and being applicable to global ship real-time scheduling), and reduces reliance on manual labor. Automated data cleaning, weight calculation, and path search lower the operation threshold for crew members, reduce human errors, and improve shipping management efficiency. The present invention breaks through the limitations of traditional route planning in terms of safety and efficiency, forming an intelligent, adaptive, and all-element integrated ship climate route planning solution, which is applicable to the ship navigation needs in any sea area of the world. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flowchart of the ship climate route planning method based on graph theory of the present invention.

[0054] Figure 2 It is an example diagram of nodes in the North Indian Ocean of the present invention.

[0055] Figure 3 It is an example diagram of nodes in the North Pacific Ocean of the present invention.

[0056] Figure 4 It is an example diagram of nodes in the North Atlantic Ocean of the present invention.

[0057] Figure 5 It is an example diagram of nodes along the coast of China of the present invention.

[0058] Figure 6 It is an example diagram of the climate route from Yangshan, China to Los Angeles, USA in January of the present invention.

[0059] Figure 7 It is an example diagram of the climate route from Yangshan, China to Los Angeles, USA in April of the present invention.

[0060] Figure 8 It is an example diagram of the climate route from Yangshan, China to Los Angeles, USA in July of the present invention.

[0061] Figure 9This is an example diagram of the climate route from Yangshan, China to Los Angeles, USA in October for the present invention.

[0062] Figure 10 This is the structural diagram of the ship climate route planning system based on graph theory for the present invention. Detailed implementation manners

[0063] To understand the content of the present invention more clearly, it will be described in detail in conjunction with the drawings and embodiments.

[0064] The present invention discloses a ship climate route planning method based on graph theory. Aiming at reducing manual interference, improving data quality, simplifying the design process, etc., from the perspective of simplifying the key parameters required for route design and avoiding human interference, it realizes accurate and efficient planning of the ship's climate route, providing safety guarantee for ship navigation. This method can be executed by a processor or a processing-capable electronic device, such as Figure 1 As shown, first construct a global ship navigation network, and then based on the navigation network, clean the global ship AIS trajectory data, extract the key node information in each cleaned AIS trajectory data and process it into a one-dimensional node sequence respectively. Then, according to the one-dimensional node sequence, determine the node features related to climate conditions, the weights of the edges between two nodes in the network, and count the maximum draft and maximum width of the ships allowed to pass through each node. Next, construct an intermediate node sequence (also called an intermediate table) of the navigation network. Finally, starting from the starting port and ending at the destination port, with the maximum weight value and meeting the draft constraint as the search conditions, combined with the node features related to climate conditions, search the intermediate node sequence to obtain the optimal node sequence as the ship climate route. The present invention combines global port information, geographical attributes, traffic flow statistics methods and graph theory to abstract the location information of all straits, canals and oceans, and constructs a global ship navigation network. This navigation network defines more than 2,000 nodes to cover all potential key waypoints and important turning points; uses the cleaned AIS trajectory data to statistically analyze the attribute information of each node in the navigation network, fully considering the ship characteristics and the climate of the route; constructs an intermediate node sequence to filter the graph data of tens of millions of levels to within thousands of levels, and based on the intermediate node sequence, with the maximum weight value and meeting the draft and ship width constraints as the search conditions, combined with the node features related to climate conditions, search to obtain the optimal node sequence as the ship climate route, which can accurately and efficiently plan the ship's climate route. Specifically, this method includes the following steps:

[0065] I. Steps for constructing the global ship navigation network: Based on global port information, geographical attributes, traffic flow statistics methods and graph theory, extract the location information of all sea areas and construct a global ship navigation network. The navigation network includes several nodes that divide sea areas and cover key waypoints and important turning points, and edges representing the connection relationships between the nodes. The specific steps of this step are as follows:

[0066] Extract the location information of all straits, canals, oceans, etc. based on global port information and geographical attributes;

[0067] Based on the location information of all straits, canals, oceans, etc. extracted and the global popular shipping routes map, determine the large nodes that divide the ocean or water area, and the large nodes pass through all the popular shipping routes passing through this ocean or water area;

[0068] Based on the large nodes, use the traffic flow statistics method and graph theory to select small nodes that describe key route points and important turning points, and the small nodes are subsets of the corresponding large nodes;

[0069] Based on the large nodes and the small nodes, construct the navigation network of global ships.

[0070] Since the spatial constraints of maritime ship navigation are relatively vague, and from a geographical perspective, the ocean is mainly separated by island groups or continents, which are often regarded as key nodes in the navigation network. Therefore, the global ship navigation network constructed in the embodiments of the present invention classifies and summarizes the nodes into large nodes and small nodes. Define large nodes to divide the ocean or water area, and define small nodes to describe key route points and important turning points. Small nodes are subsets of large nodes, and each large node includes at least one small node. From a geographical perspective, large nodes (such as canals, straits, etc.) are key nodes connecting various water areas; from a meteorological perspective, ships at small nodes will encounter ocean currents and climates in different directions and magnitudes. Nodes need to meet the following conditions: 1. All nodes must be distributed in areas where ships may navigate and cannot appear in non-navigable places, such as on land or in shallow water areas; 2. All nodes should maintain a certain spacing and cannot be too concentrated or too sparse. Being too concentrated will lead to too high a computational complexity, and being too sparse will result in the loss of the fineness of route planning; 3. Nodes should be able to reflect the main characteristics of the route, such as turning points, traffic busy areas, etc., which can be achieved by setting more nodes in these places.

[0071] Embodiments of the present invention combine global port information and geographical attributes, utilize traffic flow statistical methods and graph theory, extract the location information of all straits, canals, oceans, etc., combine the popular shipping routes of all ships, set large nodes for all straits, canals, narrow channels, etc., and divide each ocean into three regions according to the east, middle, west or north, middle, south, and set large nodes. Based on the large nodes, more than 2,000 small nodes are defined to cover all potential key navigation points and important turning points. The method for constructing the global ship navigation network refers to the traffic flow statistical method: for example, magnetic coils are set on the road, and when a vehicle passes above the magnetic coil, the traffic statistical counter will detect the current change caused by the magnetic coil and count it as a vehicle passing. Similarly, according to the maritime geographical attribute information, gate lines are set for all straits, canals, oceans, etc. as large nodes of the navigation network, and then according to the AIS trajectory data of global ships on the popular shipping routes, it is judged whether the ship passes through this node. If the ship passes through this node, the node counter is incremented, and finally the node with more counts is used as the small node on the large node. Therefore, the nodes set in this way (including large nodes and small nodes) can reflect the main characteristics of the global ship routes. Some node data is shown in Table 1. In Table 1, each node number corresponds to a small node, the node code represents the large node to which the small node belongs, and the node direction of the small node represents the course of most ships passing through this small node.

[0072] Table 1 Example of Node Data and Fields

[0073]

[0074]

[0075] Exemplarily, such as Figure 2 shown Figure 2 shows an example of nodes in the North Indian Ocean, including large nodes such as the Somalia large node, the Indian Ocean (West) large node, the Lakshadweep large node, the Palk Strait large node, the south of Matara large node, the Indian Ocean (East) large node, the Malacca Strait large node, etc. Among them, the Somalia large node includes multiple small nodes such as Somalia_01 and Somalia_02, and the Palk Strait large node only includes one small node, that is, the Palk Strait itself. Similarly, such as Figure 3 shown Figure 3 shows an example of nodes in the North Pacific Ocean, including large nodes such as the Pacific Ocean (East) large node, the Pacific Ocean (West) large node, and the Attu Island large node; such as Figure 4 shown Figure 4 shows an example of nodes in the North Atlantic Ocean, including large nodes such as the St. Paul Island large node, the North Atlantic Ocean (West) large node, and the North Atlantic Ocean (East) large node; such as Figure 5 shown Figure 5Examples of nodes along China’s coast are shown, including the lower reaches of the Yangtze River, the Korea (Tsushima) Strait, and the Kagoshima node.

[0076] 2. Global Ship AIS Data Processing Step: Based on the navigation network, clean the AIS track data of each ship in the world, and use the PostGIS geographic information engine to extract the key node information in the ship's navigation track from the cleaned AIS track data to form a one-dimensional node sequence. The one-dimensional node sequence includes the nodes of the navigation network that the ship passes through and the corresponding navigation characteristics. This step is specifically as follows:

[0077] The AIS trajectory data of all ships in the world are cleaned by deduplication, invalid point filtering, speed abnormal point filtering, drift point filtering, etc. to obtain the first AIS trajectory data;

[0078] According to the AIS drift rate, the AIS loss rate and the ship information validity index requirements, an AIS track that meets the requirements is selected from the first AIS track data to obtain the second AIS track data;

[0079] The second AIS trajectory data are grouped based on route similarity, and routes with a navigation frequency lower than a preset threshold (such as 20, 30 or 40, etc.) in each group are eliminated to obtain third AIS trajectory data;

[0080] By using the PostGIS geographic information engine technology in the PostgreSQL database, the trajectory points that do not belong to the small nodes of the navigation network in the third AIS trajectory data after each cleaning process are eliminated, and then the key node information in the ship's navigation trajectory is extracted to obtain the corresponding one-dimensional node sequence.

[0081] Furthermore, the spatial functions of PostGIS (such as ST_Intersects) are used to determine whether the ship trajectory passes through the node.

[0082] Furthermore, the AIS drift rate indicator is calculated according to the following formula:

[0083] R=N1 / (N1+T1)

[0084] e=100-R*r

[0085] Among them, R represents the drift rate, N1 represents the number of cleaned trajectory points, T1 represents the number of valid samples (trajectory points) retained after data cleaning, r represents the proportional coefficient, which is set according to the importance of the AIS drift rate indicator, and e represents the final score; the smaller the AIS drift rate indicator score, the higher the AIS trajectory quality corresponding to this indicator.

[0086] Furthermore, the AIS loss rate indicator is calculated according to the following formula:

[0087] Formula for scoring based on frequency:

[0088] R1 = N1 / T1

[0089] e1 = 100 - R1 * r

[0090] Wherein, R1 represents the loss rate, N1 represents the number of abnormal trajectory points with abnormal time intervals in the AIS trajectory, T1 represents the number of valid samples (trajectory points) retained after data cleaning, r represents the proportionality coefficient set according to the importance of the AIS loss rate index, and e1 represents the final score based on frequency scoring;

[0091] Formula for scoring based on cumulative quantity:

[0092] R2 = V1 / S1

[0093] e2 = 100 - R2 * r

[0094] Wherein, R2 represents the loss rate, V1 represents the total abnormal time of abnormal trajectory points with abnormal time intervals in the AIS, S1 represents the total time interval of valid samples (trajectory points) retained after data cleaning, r represents the proportionality coefficient set according to the importance of the AIS loss rate index, and e2 represents the final score based on cumulative quantity scoring;

[0095] Total score calculation formula:

[0096] e = e1 * ω1 + e2 * ω2

[0097] Wherein, e1 represents the final score based on frequency scoring, e2 represents the final score based on cumulative quantity scoring, ω1 and ω2 represent weights, and ω1 + ω2 = 1;

[0098] The above-mentioned trajectory points with abnormal time intervals refer to abnormal trajectory points with time intervals exceeding 6 hours. The smaller the score of the AIS loss rate index, the higher the quality of the AIS trajectory corresponding to this index.

[0099] Furthermore, the method for judging the effectiveness index of ship information:

[0100] Judgment of valid International Maritime Organization (IMO) number: The IMO consists of 7 Arabic numerals. The first 6 digits are sequential numbers, and the 7th digit is the check digit. If the units digit of the sum after multiplying the first 6 digits by the coefficients 7, 6, 5, 4, 3, and 2 respectively is equal to the check digit, it indicates that the IMO number is valid;

[0101] Judgment of valid MMSI number: The MMSI consists of 12 Arabic numerals. The first three digits are the valid country code. Take the first three digits of the MMSI to judge whether the country code is valid. If it is valid, it indicates that the MMSI number is valid;

[0102] Valid Call Sign Number Judgment: Query through the IMO database or the International Telecommunication Union (ITU) ship identification code database to determine whether the Call Sign number is normal;

[0103] Ship Size Judgment: A ship size with a length of 10 - 500 meters and a width of 1 - 125 meters is a valid ship size;

[0104] If all of the above four judgment criteria are met, the ship information validity index meets the requirements.

[0105] III. Steps for Constructing Network Node Attributes: Construct the node attributes of the navigation network, including: counting the frequency, month, draft, and basic ship information of ships passing through each node in the navigation characteristics, calculating the conditional probability of a ship passing through a certain node and its next node, and using this probability as the weight of the edge between two nodes in the network; using the kernel density estimation method, based on the draft of ships passing through each node and the basic ship information, statistically calculate the maximum draft and maximum width of ships allowed to pass through each node; based on the month in which ships pass through each node in the navigation characteristics, statistically calculate the ship passing month data of each node, analyze the adaptability impact of climate conditions on the node's navigation ability, and form node characteristics related to climate conditions.

[0106] Embodiments of the present invention can calculate the passing situation of small nodes in global AIS trajectory data. During the calculation process, information such as the month, direction, draft of the ship, and the ship type when the ship passes through the node is recorded. Based on the ship passing months of each small node, the adaptability of climate conditions (such as ocean currents, monsoons, water temperature, etc.) to navigation is analyzed, and then the climate window period suitable for passing through the small node is determined, forming node characteristics related to climate conditions. Finally, all the statistically obtained data can be stored in the node attributes in json format. In addition to recording the attributes of small nodes, it is also necessary to calculate the frequency of a ship passing through a certain small node and its next small node, which is understood as the conditional probability of the next small node appearing when the current small node appears in the sequence data of ship passing through small nodes, and use this conditional probability as the weight value of the edge between these two small nodes in the navigation network.

[0107] Preferably, embodiments of the present invention can use Bayes' formula to calculate the conditional probability of a ship passing through a certain node and its next node.

[0108] Preferably, in the step of constructing the network node attributes, when calculating the conditional probability of a ship passing through a certain node and its next node, the ship type factor can also be combined. The ship types are divided into different categories. For different types of ships, the frequencies, months, and drafts of passing through each node and its next node are respectively counted. Combining the node characteristics related to the climate conditions corresponding to different types of ships, the conditional probabilities of different types of ships passing through a certain node and its next node are respectively calculated, and a node attribute table exclusive to the ship type is generated and stored in the node attributes. When planning the route, the corresponding conditional probability can be selected as the weight of the edge according to the actual ship type.

[0109] In the embodiment of the present invention, the kernel density estimation method can be used to statistically calculate the maximum draft and maximum width data of ships allowed to pass through each small node. By the probability density function estimated by the kernel density estimation method, combined with the specified allowable values of the ship, the draft value and ship type width value corresponding to the specified quantile can be calculated as the lower limit of the safety threshold, that is, the requirements for the maximum draft and maximum width data of the node. This method is simple and efficient, and can accurately calculate the maximum draft and maximum width data of the node, avoiding the risk of stranding or collision caused by the ship being too wide or having too large a draft.

[0110] IV. Intermediate node sequence construction step: Based on the starting port, destination port and each node attribute, construct the intermediate node sequence of the navigation network, and filter the massive graph data to within the order of thousands through hash mapping. The intermediate node sequence includes potential nodes from the starting port to the destination port. This step is specifically as follows:

[0111] Determine the starting port and the destination port;

[0112] In the navigation network, obtain the sequence of large nodes passed by the route with the shortest voyage as the goal;

[0113] According to the sequence of large nodes, list the pairs of potential small node sequences between adjacent large nodes;

[0114] Using the hash algorithm, based on all pairs of small node sequences, with the starting port, destination port and month as the hash keys, and the edge weight constraint, climate condition constraint, and draft constraint as the hash values, map the hash keys to a storage space with a fixed length through the hash function for constraint filtering, and filter the massive graph data to within the order of thousands to construct the intermediate node sequence of the navigation network.

[0115] Embodiments of the present invention can filter out a large amount of redundant data using large nodes, improving the efficiency of route planning, reducing the intermediate node sequence finally constructed from tens of millions of graph data to within the order of thousands, and greatly reducing the difficulty of route planning. The design of the hash key (departure port - destination port - month) and hash value (weight constraint, climate constraint, draft constraint) has dynamic adaptability, supports fast matching of multi-dimensional conditions, and the hash algorithm can keep the data unique and ensure efficient querying. Part of the intermediate node sequence is shown in Table 2. Each row of data includes information from the departure port (source port in Table 2) or small node (node in Table 2) to the destination port or small node. Based on this information, several feasible routes from the departure port to the destination port can be deduced.

[0116] Table 2 Partial Example of Intermediate Node Sequence

[0117]

[0118] V. Steps for Ship Climate Route Planning: Use a recursive algorithm to perform recursive search in the intermediate node sequence. Starting from the departure port, with the maximum weight and meeting draft and ship width constraints as search conditions, combined with node features related to climate conditions, search until the destination port to obtain the optimal node sequence as the ship climate route.

[0119] Embodiments of the present invention can use a recursive algorithm to perform recursive search in the intermediate node sequence (including depth-first search or breadth-first search). According to parameters such as ship type, ship draft, route departure and destination ports, estimated departure time, and estimated arrival time, starting from the departure port, with the maximum weight (selecting the corresponding conditional probability as the edge weight according to the actual ship type) and meeting draft and ship width constraints as search conditions, combined with node features related to climate conditions (judging whether the ship is suitable to sail through this node in the current month), until the destination port is searched, and the obtained node sequence is the ship climate route.

[0120] The route planning method of embodiments of the present invention conforms to historical crew experience rules and ship safety constraints, and also provides a benchmark and experience for ship meteorological route planning. Due to the uncertainty of ocean meteorology, it is necessary to monitor the position and meteorological conditions of the ship in real time, and optimize and adjust the route in real time to cope with sudden meteorological changes and other unforeseen situations. Therefore, after obtaining the ship's climate route, the crew can dynamically adjust the specific sailing path according to the meteorology, thereby improving safety.

[0121] VI. Steps for Visualizing the Ship Climate Route, which are preferred steps of this embodiment. The determined ship climate route can be visualized through a geographic information system, enabling the crew to clearly see the planned route of the ship climate route and the real-time status of navigation, providing real-time, accurate, and effective reference information for the crew to assist them in making decisions.

[0122] Exemplarily, as Figure 6 shown, it is the climate route in January from Yangshan, Shanghai ( Figure 6 on the left side in the figure) to Los Angeles ( Figure 6 on the right side in the figure). This climate route successively passes through large nodes such as Kagoshima, the large node in the (East) Pacific Ocean, the large node in the (West) Pacific Ocean, the National Park large node, etc.; as Figure 7 shown, it is the climate route in April from Yangshan, Shanghai to Los Angeles. This climate route successively passes through large nodes such as Zhongtong Island, Qingyi Strait, the large node in the (East) Pacific Ocean, the large node in the (West) Pacific Ocean, the National Park large node, etc.; as Figure 8 shown, it is the climate route in July from Yangshan, Shanghai to Los Angeles. This climate route successively passes through large nodes such as Zhongtong Island, Qingyi Strait, the large node in the (East) Pacific Ocean, Attu Island, the Fox Islands, the large node in the (West) Pacific Ocean, etc.; as Figure 9 shown, it is the climate route in October from Yangshan, Shanghai to Los Angeles. This climate route successively passes through large nodes such as Kagoshima, the large node in the (East) Pacific Ocean, the large node in the (West) Pacific Ocean, etc. It can be seen from this that the climate, ocean currents and other factors of the ocean are different in different months, resulting in different plans for the climate navigation of ships.

[0123] In the embodiment of the present invention, by constructing a navigation network, processing ship AIS data, determining the attributes of the navigation network, and statistically calculating the maximum draft and width of ships allowed to pass through each node of the navigation network, constructing an intermediate node sequence of the navigation network, and searching for the intermediate node sequence, combined with efficient algorithm designs such as graph theory, PostGIS geographic information engine technology, climate data fusion, kernel density estimation method, hash function, recursive algorithm, etc., an optimal node sequence is obtained and used as the climate route of the ship. In this way, the climate route of the target ship can be accurately, real - time, safely, economically and efficiently planned, solving the core problems such as traditional reliance on manual experience, poor data quality, and response lag. It fully considers the ship characteristics and the climate of the route, provides safety guarantee for ship navigation, and provides innovative technical support for intelligent shipping. The present invention simplifies the key parameters required for route design, such as ship type, ship draft, starting and ending ports of the route, estimated departure time, and estimated arrival time, and designs a specific climate route according to these key parameters to avoid human interference.

[0124] Based on the same inventive concept, one or more embodiments of this specification also provide a ship climate route planning system based on graph theory. Since the principle of the problem solved by the ship climate route planning system based on graph theory is similar to that of the aforementioned ship climate route planning method based on graph theory, the implementation of the ship climate route planning system based on graph theory can refer to the implementation of the aforementioned ship climate route planning method based on graph theory, and the repeated parts will not be elaborated again.

[0125] Figure 10 The structure diagram of a ship climate route planning system based on graph theory provided for one or more embodiments of this specification. As Figure 10 shown, the ship climate route planning system based on graph theory includes a global ship navigation network construction module 101, a global ship AIS data processing module 102, a network node attribute construction module 103, an intermediate node sequence construction module 104, a ship climate route planning module 105, and a ship climate route visualization module 106 that are connected in sequence. Among them,

[0126] The global ship navigation network construction module 101 is used to extract the location information of all sea areas based on global port information, geographical attributes, traffic flow statistics methods, and graph theory, and construct a navigation network for global ships. The navigation network includes nodes that divide several sea areas and cover key waypoints and important turning points, as well as edges representing the connection relationships between the nodes.

[0127] The global ship AIS data processing module 102 is used to clean the AIS trajectory data of each ship globally based on the navigation network, and use the PostGIS geographic information engine to extract the key node information in the ship navigation trajectory from the cleaned AIS trajectory data of each ship to form a one-dimensional node sequence. The one-dimensional node sequence includes the nodes of the navigation network passed by the ship and the corresponding navigation characteristics.

[0128] The network node attribute construction module 103 is used to construct the node attributes of the navigation network, including: counting the frequency, month, draft, and basic ship information of the ship passing through each node in the navigation characteristics, calculating the conditional probability of the ship passing through a certain node and its next node, and using this probability as the weight of the edge between two nodes in the network; using the kernel density estimation method to statistically calculate the maximum draft and maximum width of the ships allowed to pass through each node based on the draft and basic ship information when the ship passes through each node; based on the month when the ship passes through each node in the navigation characteristics, statistically calculate the ship passing month data of each node, and analyze the adaptability impact of climate conditions on the node navigation ability to form node characteristics related to climate conditions.

[0129] The intermediate node sequence construction module 104 is used to construct the intermediate node sequence of the navigation network based on the starting port, destination port, and each node attribute, and filter the massive graph data to within the order of thousands through hash mapping. The intermediate node sequence includes potential nodes from the starting port to the destination port.

[0130] The ship climate route planning module 105 is used to perform recursive search in the intermediate node sequence using a recursive algorithm, starting from the starting port, with the maximum weight sum and meeting the draft and ship width constraints as the search conditions, and combining the node characteristics related to climate conditions, search to the destination port to obtain the optimal node sequence as the ship climate route.

[0131] The ship climate route visualization module 106 is used to visualize the determined ship climate route through a geographic information system after the ship climate route planning, so that the crew can clearly see the planning of the ship climate route.

[0132] Further, in the global ship navigation network construction module 101, constructing the global ship navigation network includes:

[0133] Extracting the location information of all straits, canals and oceans based on the global port information and geographical attributes;

[0134] Based on the location information of all the extracted straits, canals and oceans and the global popular route map, determining large nodes that divide the ocean or water area, and the large nodes pass through all the popular routes passing through the ocean or water area;

[0135] Based on the large nodes, using the traffic flow statistics method and graph theory, selecting small nodes that describe key waypoints and important turning points, and the small nodes are subsets of the large nodes;

[0136] Based on the large nodes and the small nodes, constructing the global ship navigation network.

[0137] The ship climate route planning system based on graph theory proposed by the present invention. In the navigation network constructed by the global ship navigation network construction module of the system, the nodes maintain a certain spacing, which can reduce the computational complexity. More nodes are set in busy shipping areas, and the nodes are mainly set at key waypoints and turning points, which can reflect the main characteristics of the global ship routes. The global ship AIS data processing module selects high-quality AIS trajectories that meet the requirements from the global AIS trajectory data according to indicators such as the AIS drift rate, AIS loss rate, and the validity of ship information in each global AIS trajectory data. Then, the selected AIS trajectories are subjected to cleaning processing including duplicate removal, invalid point filtering, speed anomaly point filtering, drift point filtering, etc. Compared with traditional methods, it does not rely on sailing directions, ensuring data timeliness, the reliability and accuracy of the selected AIS trajectories, and improving the quality of ship AIS trajectory data. The intermediate node sequence construction module obtains the large node sequence passed by the route with the shortest voyage as the goal in the navigation network, and then lists the potential small node sequence pairs between adjacent large nodes according to the large node sequence. Using the hash algorithm, all small node sequence pairs are used to construct the intermediate node sequence of the navigation network. Compared with traditional methods, the graph data in the tens of millions is filtered to within the thousands level, greatly reducing the route planning difficulty, achieving millisecond-level response, and improving the efficiency of route planning. The ship climate route planning module uses a recursive algorithm to perform recursive search in the lightweight intermediate node sequence. Compared with traditional methods, it is more efficient. Each module of the system automatically completes tasks in the processor without manual intervention, and integrates all the data required for traditional route design, including climate and ocean current data, ship data, nautical chart data, navigational notice information, etc. There is no longer a need to rely on raw data in the form of various types of nautical charts, messages, meteorological facsimiles, etc., and the route planning process is solidified, reducing the learning cost and execution difficulty of route design tasks.

[0138] It should be noted that the above specific embodiments can enable those skilled in the art to understand the present invention more comprehensively, but do not limit the present invention in any way. Therefore, although this specification has described the present invention in detail with reference to the drawings and embodiments, those skilled in the art should understand that the present invention can still be modified or equivalently replaced. In short, all technical solutions and their improvements that do not depart from the spirit and scope of the present invention should be covered by the protection scope of the patent of the present invention.

Claims

1. A ship climate route planning method based on graph theory, characterized in that Including the following steps: Steps for constructing the global ship navigation network: Based on global port information, geographical attributes, traffic flow statistics methods, and graph theory, extract the location information of all sea areas, and construct the navigation network of global ships. The navigation network includes nodes that divide sea areas and cover key route points and important turning points, as well as edges representing the connection relationships between nodes; Steps for processing global ship AIS data: Based on the navigation network, clean the AIS trajectory data of each ship globally, and use the PostGIS geographic information engine to extract the key node information in the ship navigation trajectory from the cleaned AIS trajectory data of each ship to form a one-dimensional node sequence. The one-dimensional node sequence includes the nodes of the navigation network passed by the ship and the corresponding navigation characteristics; Steps for constructing network node attributes: Construct the node attributes of the navigation network, including: counting the frequency, month, draft, and basic ship information of ships passing through each node in the navigation characteristics, calculating the conditional probability of a ship passing through a certain node and its next node, and using this probability as the weight of the edge between two nodes in the network; Using the kernel density estimation method, based on the draft and basic ship information of ships passing through each node, count the maximum draft and maximum width of ships allowed to pass through each node; Based on the month when ships pass through each node in the navigation characteristics, count the ship passing month data of each node, and analyze the adaptability impact of climate conditions on the node navigation ability to form node characteristics related to climate conditions; Steps for constructing the intermediate node sequence: Based on the starting port, destination port, and each node attribute, construct the intermediate node sequence of the navigation network, and filter the massive graph data to within the order of thousands through hash mapping. The intermediate node sequence includes potential nodes from the starting port to the destination port; Steps for planning the ship climate route: Use a recursive algorithm to perform recursive search in the intermediate node sequence. Starting from the starting port, with the maximum weight sum and meeting the draft and ship width constraints as the search conditions, combined with the node characteristics related to climate conditions, search to the destination port to obtain the optimal node sequence as the ship climate route.

2. The method according to claim 1, wherein In the steps for constructing the global ship navigation network, the steps for constructing the navigation network of global ships include: Based on global port information and geographical attributes, extract the location information of all straits, canals, and oceans; Based on the extracted location information of all straits, canals, and oceans and the global popular route map, determine the large nodes that divide the ocean or water area. The large nodes pass through all popular routes passing through the ocean or water area; Based on the large nodes, use traffic flow statistics methods and graph theory to select small nodes that describe key route points and important turning points. The small nodes are subsets of the large nodes; Based on the large nodes and the small nodes, construct the navigation network of global ships.

3. The method according to claim 2, wherein In the steps for processing global ship AIS data, the steps of cleaning the AIS trajectory data of each ship globally based on the navigation network and using the PostGIS geographic information engine to extract the key node information in the ship navigation trajectory from the cleaned AIS trajectory data of each ship to form a one-dimensional node sequence include: The AIS trajectory data of all ships in the world are cleaned by deduplication, invalid point filtering, speed abnormal point filtering, and drift point filtering to obtain the first AIS trajectory data; According to the AIS drift rate, the AIS loss rate and the ship information validity index requirements, an AIS track that meets the requirements is selected from the first AIS track data to obtain the second AIS track data; The second AIS trajectory data are grouped based on route similarity, and routes with navigation frequencies lower than a preset threshold in each group are eliminated to obtain third AIS trajectory data; By using the PostGIS geographic information engine technology in the PostgreSQL database, the trajectory points that do not belong to the small nodes of the navigation network in the third AIS trajectory data after each cleaning process are eliminated, and then the key node information in the ship's navigation trajectory is extracted to obtain the corresponding one-dimensional node sequence.

4. The method according to any one of claims 1 to 3, characterized in that, In the step of constructing the network node attributes, the frequency, month, draft and basic information of the ships passing through each node in the navigation characteristics are counted, and the conditional probability of the ship passing through a node and its next node is calculated using the Bayesian formula.

5. The method according to claim 4, wherein In the network node attribute construction step, when calculating the conditional probability of a ship passing through a node and its next node, the ship type factor is also combined to classify the ship types into different categories. For different types of ships, the frequency, month, and draft of passing through each node and its next node are respectively counted. In combination with the node characteristics related to the climatic conditions corresponding to different types of ships, the conditional probabilities of different types of ships passing through a node and its next node are respectively calculated, and a node attribute table exclusive to the ship type is generated and stored in the node attribute. When planning the route, the corresponding conditional probability is selected as the weight of the edge according to the actual ship type.

6. The method according to claim 2, wherein In the intermediate node sequence construction step, the intermediate node sequence construction step of the navigation network includes: Determine the port of origin and port of destination; In the navigation network, a sequence of large nodes that a route passes through is obtained with the shortest voyage as the goal; According to the large node sequence, potential small node sequence pairs between adjacent large nodes are listed; Using a hash algorithm, based on all small node sequence pairs, with the starting port, destination port and month as hash keys, edge weight constraints, climate condition constraints and draft constraints as hash values, the hash keys are mapped to a fixed-length storage space through a hash function, and constraint filtering is performed to filter the massive graph data to within a thousand items, thus constructing the intermediate node sequence of the navigation network.

7. The method according to claim 1, characterized in that, After the ship climate route planning step, a ship climate route visualization step is also included: the determined ship climate route is visualized through a geographic information system so that the crew can clearly see the planning of the ship climate route.

8. A ship climate route planning system based on graph theory, characterized in that, It includes a global ship navigation network construction module, a global ship AIS data processing module, a network node attribute construction module, an intermediate node sequence construction module and a ship climate route planning module, which are connected in sequence. The navigation network construction module is used to extract the location information of all sea areas based on global port information, geographical attributes, traffic flow statistics methods and graph theory, and construct a navigation network for global ships. The navigation network includes nodes that divide sea areas and cover key waypoints and important turning points, as well as edges representing the connection relationships between nodes. The ship AIS data processing module is used to clean the AIS trajectory data of each ship globally based on the navigation network, and use the PostGIS geographic information engine to extract the key node information in the ship navigation trajectory from the cleaned AIS trajectory data of each ship to form a one-dimensional node sequence. The one-dimensional node sequence includes the nodes of the navigation network passed by the ship and the corresponding navigation characteristics. The network node attribute construction module is used to construct the node attributes of the navigation network, including: counting the frequency, month, draft of ships passing through each node in the navigation characteristics and the basic information of the ships, calculating the conditional probability of a ship passing through a certain node and its next node, and using this probability as the weight of the edge between two nodes in the network; using the kernel density estimation method, based on the draft of the ship when passing through each node and the basic information of the ship, to count the maximum draft and maximum width of the ships allowed to pass through each node; based on the month when the ship passes through each node in the navigation characteristics, to count the ship passing month data of each node, and analyze the adaptability impact of climate conditions on the node navigation ability, to form node characteristics related to climate conditions. The intermediate node sequence construction module is used to construct the intermediate node sequence of the navigation network based on the starting port, destination port and each node attribute, and filter the massive graph data to within the order of thousands through hash mapping. The intermediate node sequence includes potential nodes from the starting port to the destination port. The ship climate route planning module is used to perform recursive search in the intermediate node sequence using a recursive algorithm, starting from the starting port, with the maximization of the weight sum and meeting the draft and ship width constraints as the search conditions, and combining the node characteristics related to climate conditions, search to the destination port to obtain the optimal node sequence as the ship climate route.

9. The system according to claim 8, wherein In the global ship navigation network construction module, constructing the navigation network for global ships includes: extracting the location information of all straits, canals and oceans based on global port information and geographical attributes; determining large nodes that divide the ocean or water area based on the extracted location information of all straits, canals and oceans and the global popular route map. The large nodes pass through all popular routes passing through the ocean or water area; selecting small nodes describing key waypoints and important turning points based on the large nodes using traffic flow statistics methods and graph theory. The small nodes are subsets of the large nodes; constructing the navigation network for global ships based on the large nodes and the small nodes.

10. The system according to claim 8 or 9, characterized in that, It also includes a ship climate route visualization module. The ship climate route visualization module is connected to the ship climate route planning module and is used to visualize the determined ship climate route through a geographic information system after the ship climate route is planned, so that the crew can clearly see the planning of the ship climate route.

Images