A ship collision avoidance identification and path planning method and system based on ship-shore cooperation

Abstract

This invention discloses a ship collision avoidance identification and path planning method and system based on ship-shore cooperation. It acquires multi-source data such as ship navigation status, positioning, and panoramic video by constructing a ship-shore cooperative data network. The video data is used to calibrate positioning information, and combined with a ship maneuvering model, future navigation trajectories are predicted. Based on the predicted trajectories, the probability of encounter between ships, the nearest encounter distance, and the time to collision are calculated to assess collision risk. When the risk exceeds a threshold, the optimal collision avoidance path, including heading, speed, and turning points, is planned based on the A* algorithm, integrating the predicted trajectory, calibrated positioning, risk value, and safety constraints. This invention achieves accurate collision risk perception and automatic generation of safe collision avoidance paths.

Description

Technical Field

[0001] This invention relates to the field of ship collision avoidance and management technology, and in particular to a ship collision avoidance identification and path planning method and system based on ship-shore cooperation. Background Technology

[0002] With the development of smart shipping, improving ship navigation safety has become a key focus for the industry. Traditional ship collision avoidance mainly relies on crew experience for lookout and VHF voice communication, which suffers from problems such as delayed judgment and susceptibility to errors in complex navigation sections and adverse weather conditions. In existing technologies, navigation aids based on VTS (Vessel Traffic Services) monitor electronic charts through shore-based operators and issue voice commands to ships. Although this achieves some ship-shore collaboration, its decision-making relies on human experience, resulting in insufficient efficiency and accuracy. Furthermore, it lacks intelligent path planning algorithms, making it a one-way, passive collision avoidance mode.

[0003] In recent years, although some studies have attempted to achieve partial automation by combining data such as AIS, the following core pain points are generally present: First, the problem of information silos is serious. Ship-shore data cannot achieve two-way real-time interaction and fusion, the ship cannot obtain the overall navigation situation, and the shore cannot grasp the real-time execution status of the ship. Second, the level of intelligence in collision avoidance decision-making is low, relying mostly on simple rules or manual judgment, lacking quantitative risk assessment and high-precision trajectory prediction based on multi-source data (such as AIS, Beidou, and visual) fusion calibration, resulting in inaccurate collision risk identification. Third, path planning is disconnected from maritime rules. Existing algorithms are mostly focused on geometric obstacle avoidance and fail to deeply embed international rules such as the 1972 International Convention for Preventing Collisions at Sea (COLREGs) into the core of the algorithm, so the compliance of the planned path cannot be guaranteed. Fourth, there is a lack of closed-loop optimization mechanism. The planning, execution, and feedback links are separated, and the system cannot iterate itself based on actual navigation feedback.

[0004] Therefore, there is an urgent need for a collaborative collision avoidance method and system that can deeply integrate data from ships, shore, and the cloud to achieve intelligent risk perception, automatic generation of compliance paths, and closed-loop optimization of execution feedback, so as to fundamentally improve navigation safety and efficiency in complex waters. Summary of the Invention

[0005] To address at least one of the aforementioned technical problems, this invention proposes a ship collision avoidance identification and path planning method and system based on ship-shore cooperation.

[0006] The first aspect of this invention provides a ship collision avoidance identification and path planning method based on ship-shore cooperation, comprising: A ship-shore collaborative data network is constructed based on the shipborne terminal system of the demonstration vessel, the shore-based control center and the cloud platform. Multi-source heterogeneous data of ships navigating in the target waters are acquired based on the ship-shore collaborative data network. The multi-source heterogeneous data includes navigation status data of the ships, first positioning data and panoramic video data of the demonstration vessel. Based on the panoramic video data of the demonstration ship, the first positioning data of the vessel navigating in the target waters is calibrated, and the second positioning data of the calibrated data is output. Based on the second positioning data and the navigation status data of the vessel, the navigation trajectory of the vessel within a preset time period is predicted in combination with the vessel maneuvering characteristic model, and the predicted navigation trajectory data is obtained. Based on the predicted navigation trajectory data, the probability of encounter between vessels navigating in the target waters, the nearest encounter distance, and the time to reach the nearest encounter point are calculated to obtain encounter data. Based on the encounter data, the collision risk between vessels is determined. If the collision risk is greater than a preset value, the predicted flight trajectory data, the second positioning data, the collision risk, and the preset safe encounter distance constraints are analyzed based on the A* algorithm to determine the collision avoidance path, which includes the heading, speed, and turning point sequence.

[0007] In this solution, a ship-shore collaborative data network is constructed based on the shipborne terminal system of the demonstration vessel, the shore-based control center, and the cloud platform. This network is used to acquire multi-source heterogeneous data of vessels navigating in the target waters. This multi-source heterogeneous data includes vessel navigation status data, first positioning data, and panoramic video data from the demonstration vessel. Specifically: The demonstration vessel is equipped with a shipborne terminal system that includes radar, satellite positioning modules and a panoramic camera at the bow. The shipborne terminal system of the demonstration vessel establishes a communication connection with the shipborne terminals of vessels navigating in the target waters through the AIS system and VDES system, forming a preliminary inter-ship collaborative network. The preliminary ship-to-ship collaborative network is connected to the shore-based control center in the target waters via shore-based AIS base stations and VHF communication networks to build a ship-to-shore collaborative network. The data collected by the ship-to-shore collaborative network is uploaded to the cloud platform to form a ship-to-ship, shore-to-ship, and cloud-to-cloud collaborative data network. Analyze the data transmission quality of vessels navigating in the target waters and the data reception quality of the demonstration vessel to determine the data collection points of the demonstration vessel for vessels navigating in the target waters. The demonstration vessel, based on the shipborne terminal system at the data collection point, acquires the ship navigation status data and first positioning data transmitted back by AIS and VDES from the ships navigating in the target waters. The ship navigation status data includes the ship's position, heading, speed, and heading, which characterize the ship's dynamics, as well as the ship's engine operating status, such as the speed, power, and fuel consumption. The first positioning data consists of the ship's AIS positioning data and satellite positioning data. The demonstration vessel's onboard terminal system uses panoramic cameras to acquire panoramic video data of the vessel's navigation, which is then aggregated to form a multi-source heterogeneous dataset.

[0008] In this scheme, the analysis of data transmission quality of vessels navigating in the target waters and data reception quality of demonstration vessels, to determine the data collection points of the demonstration vessels for vessels navigating in the target waters, specifically involves: The signal strength, packet loss rate, and transmission delay of ships navigating in the target waters when transmitting data back to the demonstration ship, as well as the signal reception strength and signal-to-noise ratio of the demonstration ship when receiving data, are obtained. The data transmission quality score between each ship navigating in the target waters and the demonstration ship is calculated in a comprehensive manner. Based on the real-time location information of the demonstration vessel, the historical navigation trajectory of the demonstration vessel in the target waters is obtained, the historical navigation trajectory is divided according to a preset grid, the average quality score of the navigation vessel data received by the demonstration vessel in each grid is calculated, and a data reception quality distribution map is formed. The K-means clustering algorithm is introduced. k grids are randomly selected from the data reception quality distribution map as initial cluster centers. Then, the Euclidean distance from the quality score feature vector of each grid in the distribution map to each initial cluster center is calculated. Each grid is assigned to the cluster containing the nearest cluster center, resulting in k data quality clusters. Select several clusters with the highest average grid quality score within each cluster from k clusters, and determine their corresponding spatial regions as candidate regions for high-quality data collection. Based on the boundary coordinates of the candidate areas for high-quality data acquisition, and combined with the real-time distribution density of ships, coordinate points are calculated within each candidate area to enable the demonstration ship to maintain high-quality communication connections with the maximum number of ships. These coordinate points are then determined as the optimal locations for the demonstration ship to collect data from ships in the target waters.

[0009] In this scheme, the first positioning data of the vessel navigating in the target waters is calibrated based on the panoramic video data of the demonstration vessel, and the calibrated second positioning data is output. Based on the second positioning data and the vessel's navigation status data, the vessel's navigation trajectory within a preset time period is predicted using a vessel maneuvering characteristic model, resulting in predicted navigation trajectory data. Specifically: Compare the AIS positioning data in the first positioning data with the satellite positioning data, and identify the navigation vessels whose positioning deviation is greater than a preset threshold as vessels to be corrected for positioning data. Based on the panoramic video data of the demonstration ship, the navigation targets in the video are identified by machine vision recognition algorithm, and the measured visual distance between the ship to be located is corrected by calculating the data to be located based on the panoramic video data. Based on the current satellite positioning data of the demonstration vessel, and according to the measured visual distance, the first positioning data of the vessel to be positioned is corrected, and the calibrated second positioning data is output. If the positioning deviation of the first positioning data is not greater than a preset threshold, the average value of the AIS positioning data and the satellite positioning data is output as the second positioning data. A ship maneuvering characteristic model is constructed. The second positioning data, ship heading, speed and bow direction are used as the initial state input of the model. Based on the ship's tonnage, draft, ship type parameters and current speed, maneuvering characteristic parameters including ship turning performance, stroke and rudder response time are calculated. Acquire real-time navigation environment data, including water flow velocity, flow direction, wind speed, and wind direction, as external disturbance vector input, and calculate the impact of external disturbance on the ship's motion state; A discrete-time Markov chain model is introduced to predict the navigation trajectory of the ship within a preset time period based on the second positioning data, the navigation status data of the ship, and the influence of the ship's operating status, thus obtaining the predicted navigation trajectory data.

[0010] In this scheme, a discrete-time Markov chain model is introduced to predict the navigation trajectory of the ship within a preset time period based on the second positioning data, the navigation status data of the ship, and the influence of the ship's operating status, thereby obtaining predicted navigation trajectory data. Specifically: The navigation state of the vessel in the target waters is discretized into a state space, where each state is determined by a combination of the vessel's position, heading, and speed. Based on the second positioning data and the vessel's navigation state data, the initial state of the vessel at the current moment is determined. Based on the navigation state data sequence of historical ships, the probability of transitioning from the current state to the next possible state in the state space is statistically analyzed, and a one-step state transition probability matrix is ​​constructed. Using the one-step state transition probability matrix as the model parameter of the Markov chain, and taking the current state of the sailing ship as the initial state, the probability distribution of the sailing ship's state in multiple discrete time steps in the future is calculated through the one-step state transition probability matrix. Based on the influence of the real-time navigation environment factors on the ship's motion state, the probability distribution of each state transition is corrected to obtain the corrected state probability distribution. From the corrected state probability distribution, the state with the highest probability of occurrence at each future time step is selected as the predicted state of the navigating vessel at that time step. Connect the predicted states at all predicted time steps from the current moment to the end of a preset future time period to form a continuous state sequence consisting of the states with the highest probability. Extract the state components representing the ship's position from this sequence to generate the predicted navigation trajectory data of the ship in the preset future time period.

[0011] In this scheme, the step of calculating the encounter probability, nearest encounter distance, and time to reach the nearest encounter point between vessels navigating in the target waters based on the predicted navigation trajectory data to obtain encounter data, and determining the collision risk between vessels based on the encounter data, specifically involves: Based on the predicted navigation trajectory data, calculate the potential intersection points of the predicted trajectories of any two ships within a preset time period in the future, and for each intersection point, calculate the time difference between the arrival of the two ships at the intersection point. If the time difference is not less than the preset safe meeting time window threshold, it is determined that there is a meeting risk between the two ships at the meeting point. The meeting point is marked as a valid meeting point. For each valid meeting point, the probability distribution of the predicted navigation trajectory of the two ships at the meeting point is calculated, and the probability of the two ships actually meeting at the meeting point is calculated based on the probability distribution to obtain the meeting probability. From the effective meeting points, find the minimum geometric distance between the two ships' trajectories as the nearest meeting distance, and calculate the time required for the two ships to reach the nearest meeting point in their current state, thus obtaining the time to reach the nearest meeting point. The encounter probability, the nearest encounter distance, and the time to reach the nearest encounter point are used as encounter data, and the collision risk level between ships is determined based on the encounter data.

[0012] In this scheme, if the collision risk exceeds a preset value, the predicted flight trajectory data, second positioning data, collision risk, and pre-set safe encounter distance constraints are analyzed based on the A* algorithm to determine a collision avoidance path. The collision avoidance path includes a heading, speed, and a sequence of turning points, specifically: Vessels with a collision risk greater than the preset value are designated as collision avoidance planning vessels. Based on the second positioning data and predicted navigation trajectory data, the electronic nautical chart model of the target waters is discretized into grid nodes. The grid position in front of the collision avoidance planning vessel is taken as the starting node, and the grid position in the preset collision avoidance waypoint set is taken as the target node. Construct a path search cost function for the A* algorithm. The cost function includes the actual travel cost from the starting node to the current node and the expected travel cost from the current node to the target node estimated by the heuristic function. The pre-set safe encounter distance constraint is transformed into a penalty term in the cost function. In the grid model, the grid occupied by the predicted future navigation trajectory data of the collision avoidance planning vessel and surrounding vessels is marked as a dynamic obstacle area. Based on the A* algorithm, starting from the starting node, the total cost of adjacent grid nodes is iteratively evaluated. The node with the minimum total cost is added to the optimal path set, and the current node is updated to generate a collision avoidance path composed of a sequence of grid nodes. Based on the sequence of grid nodes traversed by the initial collision avoidance path, calculate the heading changes between adjacent nodes and the required speed adjustments, and output a collision avoidance path containing the heading, speed, and turning point sequence.

[0013] A second aspect of the present invention also provides a ship collision avoidance identification and path planning system based on ship-shore cooperation. The system includes: a memory and a processor. The memory includes a program for a ship collision avoidance identification and path planning method based on ship-shore cooperation. When the program for the ship collision avoidance identification and path planning method based on ship-shore cooperation is executed by the processor, it implements the steps of the ship collision avoidance identification and path planning method based on ship-shore cooperation as described in any of the above claims.

[0014] This invention discloses a ship collision avoidance identification and path planning method and system based on ship-shore cooperation. It acquires multi-source data such as ship navigation status, positioning, and panoramic video by constructing a ship-shore cooperative data network. The video data is used to calibrate positioning information, and combined with a ship maneuvering model, future navigation trajectories are predicted. Based on the predicted trajectories, the probability of encounter between ships, the nearest encounter distance, and the time to collision are calculated to assess collision risk. When the risk exceeds a threshold, the optimal collision avoidance path, including heading, speed, and turning points, is planned based on the A* algorithm, integrating the predicted trajectory, calibrated positioning, risk value, and safety constraints. This invention achieves accurate collision risk perception and automatic generation of safe collision avoidance paths. Attached Figure Description

[0015] Figure 1 A flowchart of a ship collision avoidance identification and path planning method based on ship-shore cooperation according to the present invention is shown; Figure 2 This invention illustrates a flowchart of the process for determining data collection points for navigation vessels from a demonstration vessel in a target waterway. Figure 3 A flowchart illustrating the present invention for determining the collision risk between navigating vessels is shown; Figure 4 The diagram shows a block diagram of a ship collision avoidance identification and path planning system based on ship-shore cooperation according to the present invention. Detailed Implementation

[0016] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0017] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0018] Figure 1 The flowchart of a ship collision avoidance identification and path planning method based on ship-shore cooperation according to the present invention is shown.

[0019] like Figure 1 As shown, the first aspect of the present invention provides a ship collision avoidance identification and path planning method based on ship-shore cooperation, comprising: S102, Based on the shipborne terminal system of the demonstration vessel, the shore-based control center and the cloud platform, a ship-shore collaborative data network is constructed. Based on the ship-shore collaborative data network, multi-source heterogeneous data of ships navigating in the target waters are obtained. The multi-source heterogeneous data includes navigation status data of the ships, first positioning data and panoramic video data of the demonstration vessel. S104, perform data calibration on the first positioning data of the vessel navigating in the target waters based on the panoramic video data of the demonstration vessel, output the second positioning data of the data calibration, and predict the navigating trajectory of the vessel within a preset time period based on the second positioning data and the navigating status data of the vessel, combined with the vessel maneuvering characteristic model, to obtain the predicted navigating trajectory data. S106, calculate the probability of encounter between vessels navigating in the target waters, the nearest encounter distance and the time to reach the nearest encounter point based on the predicted navigation trajectory data, obtain encounter data, and determine the collision risk between vessels based on the encounter data. S108, if the collision risk is greater than the preset value, the predicted flight trajectory data, the second positioning data, the collision risk and the preset safe encounter distance constraints are analyzed based on the A* algorithm to determine the collision avoidance path, which includes the heading, speed and turning point sequence.

[0020] It should be noted that by constructing a ship-shore collaborative data network, integrating multi-source heterogeneous data such as the ship's own status, AIS / satellite positioning, and ship-side vision, and calibrating the primary positioning data such as AIS using ship-side video data, and combining this with ship maneuvering models to predict future trajectories, the accuracy of ship position and future dynamic predictions is greatly improved, enabling earlier and more accurate identification of potential collision risks. Finally, when the collision risk is determined to exceed a preset threshold, the method uses the A* algorithm to automatically generate a compliant, safe, and globally optimal collision avoidance path, taking into account safe encounter distance, maritime rule constraints, and dynamic obstacles (predicted trajectories of surrounding vessels). This path specifies the course, speed, and turning point sequence, providing clear and actionable operational guidance to the crew, achieving a closed loop from risk identification to intelligent decision-making, and effectively improving navigation safety in complex waters.

[0021] According to an embodiment of the present invention, a ship-shore collaborative data network is constructed based on the shipborne terminal system of the demonstration vessel, the shore-based control center, and the cloud platform. Multi-source heterogeneous data of vessels navigating in the target waters is acquired based on this ship-shore collaborative data network. The multi-source heterogeneous data includes navigation status data of the navigating vessels, first positioning data, and panoramic video data from the demonstration vessel. Specifically: The demonstration vessel is equipped with a shipborne terminal system that includes radar, satellite positioning modules and a panoramic camera at the bow. The shipborne terminal system of the demonstration vessel establishes a communication connection with the shipborne terminals of vessels navigating in the target waters through the AIS system and VDES system, forming a preliminary inter-ship collaborative network. The preliminary ship-to-ship collaborative network is connected to the shore-based control center in the target waters via shore-based AIS base stations and VHF communication networks to build a ship-to-shore collaborative network. The data collected by the ship-to-shore collaborative network is uploaded to the cloud platform to form a ship-to-ship, shore-to-ship, and cloud-to-cloud collaborative data network. Analyze the data transmission quality of vessels navigating in the target waters and the data reception quality of the demonstration vessel to determine the data collection points of the demonstration vessel for vessels navigating in the target waters. The demonstration vessel, based on the shipborne terminal system at the data collection point, acquires the ship navigation status data and first positioning data transmitted back by AIS and VDES from the ships navigating in the target waters. The ship navigation status data includes the ship's position, heading, speed, and heading, which characterize the ship's dynamics, as well as the ship's engine operating status, such as the speed, power, and fuel consumption. The first positioning data consists of the ship's AIS positioning data and satellite positioning data. The demonstration vessel's onboard terminal system uses panoramic cameras to acquire panoramic video data of the vessel's navigation, which is then aggregated to form a multi-source heterogeneous dataset.

[0022] It should be noted that in traditional ship collision avoidance, "information silos" often form between ships and shore-based control centers due to single communication links and unstable data transmission quality. This results in ships being unable to obtain a global navigation situation, and shore-based decision-making being limited by the lack of real-time, accurate navigation status data from the ship. A multi-layered, highly reliable ship-shore collaborative data network has been constructed. Its technical advantage lies in the fact that this network integrates three layers of communication links: ship-to-ship AIS / VDES (Very High Frequency Data Exchange System), shore-to-ship (shore-based AIS / VHF), and cloud-based, achieving real-time two-way data interconnection among ships, shore, and the cloud. In particular, by analyzing data transmission quality to determine the optimal data collection points, and by leveraging multi-sensor fusion, including panoramic cameras on the demonstration ship, it can stably and comprehensively collect multi-source heterogeneous data, including ship navigation status, engine operation status, AIS / satellite positioning, and panoramic video from the ship. The demonstration ship is the vessel collecting data, and the navigating vessel is the vessel operating in the target waters.

[0023] Figure 2 The flowchart illustrates the present invention for determining the data collection points of a demonstration vessel for navigation vessels in a target waterway.

[0024] According to an embodiment of the present invention, the analysis of data transmission quality of vessels navigating in the target waters and data reception quality of demonstration vessels to determine the data collection points of the demonstration vessels on the navigating vessels in the target waters specifically includes: The signal strength, packet loss rate, and transmission delay of ships navigating in the target waters when transmitting data back to the demonstration ship, as well as the signal reception strength and signal-to-noise ratio of the demonstration ship when receiving data, are obtained. The data transmission quality score between each ship navigating in the target waters and the demonstration ship is calculated in a comprehensive manner. Based on the real-time location information of the demonstration vessel, the historical navigation trajectory of the demonstration vessel in the target waters is obtained, the historical navigation trajectory is divided according to a preset grid, the average quality score of the navigation vessel data received by the demonstration vessel in each grid is calculated, and a data reception quality distribution map is formed. The K-means clustering algorithm is introduced. k grids are randomly selected from the data reception quality distribution map as initial cluster centers. Then, the Euclidean distance from the quality score feature vector of each grid in the distribution map to each initial cluster center is calculated. Each grid is assigned to the cluster containing the nearest cluster center, resulting in k data quality clusters. Select several clusters with the highest average grid quality score within each cluster from k clusters, and determine their corresponding spatial regions as candidate regions for high-quality data collection. Based on the boundary coordinates of the candidate areas for high-quality data acquisition, and combined with the real-time distribution density of ships, coordinate points are calculated within each candidate area to enable the demonstration ship to maintain high-quality communication connections with the maximum number of ships. These coordinate points are then determined as the optimal locations for the demonstration ship to collect data from ships in the target waters.

[0025] It should be noted that in complex real-world aquatic environments, the communication link quality between the demonstration vessel and other vessels is not uniform across all locations due to a combination of factors, including geographical environment, weather conditions, equipment performance, and vessel distribution. Significant regional differences exist. Simply selecting data collection locations based on preset routes or experience may result in the demonstration vessel being in areas with poor communication signals and unstable data quality for extended periods, affecting the integrity and reliability of data transmitted from systems such as AIS / VDES. Therefore, by introducing the K-means clustering algorithm, high-quality communication areas can be automatically and efficiently identified from the complex data reception quality distribution map of the target waters, providing a scientific basis for determining the optimal data collection points. The algorithm first divides the waters into multiple grid cells based on historical navigation trajectories and communication quality scores, and calculates the average data quality for each cell. Then, it groups these grids with similar quality characteristics into different clusters through clustering. Finally, it selects several clusters with the highest average quality scores from all clusters, representing the spatial regions with the most stable communication signals. By combining the boundary coordinates of these "high-quality candidate areas" with the real-time distribution density of ships currently in transit, a coordinate point that maximizes coverage of surrounding ships and ensures optimal communication link quality can be accurately calculated within each area.

[0026] According to an embodiment of the present invention, the first positioning data of a vessel navigating in the target waters is calibrated based on panoramic video data from the demonstration vessel, and the calibrated second positioning data is output. Based on the second positioning data and the vessel's navigation status data, and combined with a vessel maneuvering characteristic model, the vessel's navigation trajectory within a preset time period is predicted to obtain predicted navigation trajectory data. Specifically: Compare the AIS positioning data in the first positioning data with the satellite positioning data, and identify the navigation vessels whose positioning deviation is greater than a preset threshold as vessels to be corrected for positioning data. Based on the panoramic video data of the demonstration ship, the navigation targets in the video are identified by machine vision recognition algorithm, and the measured visual distance between the ship to be located is corrected by calculating the data to be located based on the panoramic video data. Based on the current satellite positioning data of the demonstration vessel, and according to the measured visual distance, the first positioning data of the vessel to be positioned is corrected, and the calibrated second positioning data is output. If the positioning deviation of the first positioning data is not greater than a preset threshold, the average value of the AIS positioning data and the satellite positioning data is output as the second positioning data. A ship maneuvering characteristic model is constructed. The second positioning data, ship heading, speed and bow direction are used as the initial state input of the model. Based on the ship's tonnage, draft, ship type parameters and current speed, maneuvering characteristic parameters including ship turning performance, stroke and rudder response time are calculated. Acquire real-time navigation environment data, including water flow velocity, flow direction, wind speed, and wind direction, as external disturbance vector input, and calculate the impact of external disturbance on the ship's motion state; A discrete-time Markov chain model is introduced to predict the navigation trajectory of the ship within a preset time period based on the second positioning data, the navigation status data of the ship, and the influence of the ship's operating status, thus obtaining the predicted navigation trajectory data.

[0027] It should be noted that the influence on the ship's motion state includes the influence of different water flow speeds, flow directions, wind speeds, and wind directions on changes in the ship's position, course, and speed, such as the increase or decrease in ship speed due to different wind speeds in different directions.

[0028] According to an embodiment of the present invention, the introduction of a discrete-time Markov chain model to predict the navigation trajectory of a ship within a preset time period based on the second positioning data, the navigation status data of the ship, and the influence of the ship's operating status, to obtain predicted navigation trajectory data, specifically involves: The navigation state of the vessel in the target waters is discretized into a state space, where each state is determined by a combination of the vessel's position, heading, and speed. Based on the second positioning data and the vessel's navigation state data, the initial state of the vessel at the current moment is determined. Based on the navigation state data sequence of historical ships, the probability of transitioning from the current state to the next possible state in the state space is statistically analyzed, and a one-step state transition probability matrix is ​​constructed. Using the one-step state transition probability matrix as the model parameter of the Markov chain, and taking the current state of the sailing ship as the initial state, the probability distribution of the sailing ship's state in multiple discrete time steps in the future is calculated through the one-step state transition probability matrix. Based on the influence of the real-time navigation environment factors on the ship's motion state, the probability distribution of each state transition is corrected to obtain the corrected state probability distribution. From the corrected state probability distribution, the state with the highest probability of occurrence at each future time step is selected as the predicted state of the navigating vessel at that time step. Connect the predicted states at all predicted time steps from the current moment to the end of a preset future time period to form a continuous state sequence consisting of the states with the highest probability. Extract the state components representing the ship's position from this sequence to generate the predicted navigation trajectory data of the ship in the preset future time period.

[0029] It should be noted that the transition probability integrates the inherent ship handling laws calculated by the ship handling characteristic model, as well as the influence of environmental disturbances on the ship's motion state calculated by real-time navigation environment data. The process of correcting the probability distribution for each state transition is as follows: the system calculates the "promoting" or "inhibiting" effects of these external disturbances on different navigation states (such as specific position, heading, and speed combinations) based on real-time acquired navigation environment data such as wind and current, and quantifies this into a probability offset vector for each target state. Then, this offset vector is weighted or probabilistically superimposed with a one-step state transition probability matrix derived from historical behavior statistics, thereby directly adjusting the value of each transition element in the probability matrix. Finally, since the adjusted probability distribution may not satisfy the condition that the sum is 1, the system immediately performs normalization processing on the entire probability distribution to ensure that the sum of the probabilities of transitioning from the current state to all possible next states after correction is strictly equal to 1, thus obtaining a corrected state probability distribution that integrates the influence of the real-time environment.

[0030] Figure 3 A flowchart illustrating the present invention for determining the collision risk between navigating vessels is shown.

[0031] According to an embodiment of the present invention, the step of calculating the encounter probability, nearest encounter distance, and time to reach the nearest encounter point between vessels navigating in the target waters based on the predicted navigation trajectory data to obtain encounter data, and determining the collision risk between vessels based on the encounter data, specifically involves: Based on the predicted navigation trajectory data, calculate the potential intersection points of the predicted trajectories of any two ships within a preset time period in the future, and for each intersection point, calculate the time difference between the arrival of the two ships at the intersection point. If the time difference is not less than the preset safe meeting time window threshold, it is determined that there is a meeting risk between the two ships at the meeting point. The meeting point is marked as a valid meeting point. For each valid meeting point, the probability distribution of the predicted navigation trajectory of the two ships at the meeting point is calculated, and the probability of the two ships actually meeting at the meeting point is calculated based on the probability distribution to obtain the meeting probability. From the effective meeting points, find the minimum geometric distance between the two ships' trajectories as the nearest meeting distance, and calculate the time required for the two ships to reach the nearest meeting point in their current state, thus obtaining the time to reach the nearest meeting point. The encounter probability, the nearest encounter distance, and the time to reach the nearest encounter point are used as encounter data, and the collision risk level between ships is determined based on the encounter data.

[0032] It should be noted that the calculated encounter probability, nearest encounter distance, and time to the nearest encounter point are compared with several preset risk level thresholds. By comprehensively evaluating the combination of these three indicators, the severity of the potential collision threat is determined.

[0033] According to an embodiment of the present invention, if the collision risk is greater than a preset value, the predicted flight trajectory data, second positioning data, collision risk, and pre-set safe encounter distance constraints are analyzed based on the A* algorithm to determine a collision avoidance path. The collision avoidance path includes a heading, speed, and a sequence of turning points, specifically: Vessels with a collision risk greater than the preset value are designated as collision avoidance planning vessels. Based on the second positioning data and predicted navigation trajectory data, the electronic nautical chart model of the target waters is discretized into grid nodes. The grid position in front of the collision avoidance planning vessel is taken as the starting node, and the grid position in the preset collision avoidance waypoint set is taken as the target node. Construct a path search cost function for the A* algorithm. The cost function includes the actual travel cost from the starting node to the current node and the expected travel cost from the current node to the target node estimated by the heuristic function. The pre-set safe encounter distance constraint is transformed into a penalty term in the cost function. In the grid model, the grid occupied by the predicted future navigation trajectory data of the collision avoidance planning vessel and surrounding vessels is marked as a dynamic obstacle area. Based on the A* algorithm, starting from the starting node, the total cost of adjacent grid nodes is iteratively evaluated. The node with the minimum total cost is added to the optimal path set, and the current node is updated to generate a collision avoidance path composed of a sequence of grid nodes. Based on the sequence of grid nodes traversed by the initial collision avoidance path, calculate the heading changes between adjacent nodes and the required speed adjustments, and output a collision avoidance path containing the heading, speed, and turning point sequence.

[0034] It should be noted that the integration of international maritime collision avoidance rules into Algorithm A in this claim is mainly reflected in the transformation of the core constraints of the rules into quantitative evaluation criteria for algorithm search. Specifically, when constructing the cost function for path search, the algorithm does not only target the shortest path, but also transforms the "safe encounter distance constraint" in the 1972 International Convention for the Prevention of Collisions at Sea (ICCCL) into a penalty term in the cost function. When the minimum distance between the predicted path and the predicted navigation trajectory data of surrounding vessels is less than the safe encounter distance, the navigation cost of that path node is increased. This means that when evaluating each candidate path node, the algorithm not only calculates its geometric distance cost, but also evaluates its dynamic positional relationship with surrounding vessels. If a candidate node is too close to the future predicted position of surrounding vessels (i.e., does not meet the preset safe encounter distance threshold), the algorithm considers that the node violates the "maintain a safe distance" rule clause and imposes a penalty cost on it, thus prioritizing its exclusion in the path search. Through this mechanism, Algorithm A can automatically plan a collision-free avoidance path in geometric space that meets the basic safety requirements of maritime regulations, thereby achieving intelligent integration and compliance with international rules.

[0035] Figure 4 The diagram shows a block diagram of a ship collision avoidance identification and path planning system based on ship-shore cooperation according to the present invention.

[0036] A second aspect of the present invention also provides a ship collision avoidance identification and path planning system based on ship-shore cooperation. The system includes: a memory 401, a processor 402, and a communication interface 403. The memory includes a ship collision avoidance identification and path planning method program based on ship-shore cooperation. The communication interface is used for data connection and communication between the memory and the processor. When the ship collision avoidance identification and path planning method program based on ship-shore cooperation is executed by the processor, it implements the steps of the ship collision avoidance identification and path planning method based on ship-shore cooperation as described in any of the above claims.

[0037] This invention discloses a ship collision avoidance identification and path planning method and system based on ship-shore cooperation. It acquires multi-source data such as ship navigation status, positioning, and panoramic video by constructing a ship-shore cooperative data network. The video data is used to calibrate positioning information, and combined with a ship maneuvering model, future navigation trajectories are predicted. Based on the predicted trajectories, the probability of encounter between ships, the nearest encounter distance, and the time to collision are calculated to assess collision risk. When the risk exceeds a threshold, the optimal collision avoidance path, including heading, speed, and turning points, is planned based on the A* algorithm, integrating the predicted trajectory, calibrated positioning, risk value, and safety constraints. This invention achieves accurate collision risk perception and automatic generation of safe collision avoidance paths.

[0038] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0039] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0040] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A ship collision avoidance identification and path planning method based on ship-shore cooperation, characterized in that, Includes the following steps: A ship-shore collaborative data network is constructed based on the shipborne terminal system of the demonstration vessel, the shore-based control center and the cloud platform. Multi-source heterogeneous data of ships navigating in the target waters are acquired based on the ship-shore collaborative data network. The multi-source heterogeneous data includes navigation status data of the ships, first positioning data and panoramic video data of the demonstration vessel. Based on the panoramic video data of the demonstration ship, the first positioning data of the vessel navigating in the target waters is calibrated, and the second positioning data of the calibrated data is output. Based on the second positioning data and the navigation status data of the vessel, the navigation trajectory of the vessel within a preset time period is predicted in combination with the vessel maneuvering characteristic model, and the predicted navigation trajectory data is obtained. Based on the predicted navigation trajectory data, the probability of encounter between vessels navigating in the target waters, the nearest encounter distance, and the time to reach the nearest encounter point are calculated to obtain encounter data. Based on the encounter data, the collision risk between vessels is determined. If the collision risk is greater than a preset value, the predicted flight trajectory data, the second positioning data, the collision risk, and the preset safe encounter distance constraints are analyzed based on the A* algorithm to determine the collision avoidance path, which includes the heading, speed, and turning point sequence.

2. The ship collision avoidance identification and path planning method based on ship-shore cooperation according to claim 1, characterized in that, The system constructs a ship-shore collaborative data network based on the demonstration vessel's onboard terminal system, shore-based control center, and cloud platform. This network acquires multi-source heterogeneous data of vessels navigating in the target waters. This multi-source heterogeneous data includes vessel navigation status data, first positioning data, and panoramic video data from the demonstration vessel. Specifically: The demonstration vessel is equipped with a shipborne terminal system that includes radar, satellite positioning modules and a panoramic camera at the bow. The shipborne terminal system of the demonstration vessel establishes a communication connection with the shipborne terminals of vessels navigating in the target waters through the AIS system and VDES system, forming a preliminary inter-ship collaborative network. The preliminary ship-to-ship collaborative network is connected to the shore-based control center in the target waters via shore-based AIS base stations and VHF communication networks to build a ship-to-shore collaborative network. The data collected by the ship-to-shore collaborative network is uploaded to the cloud platform to form a ship-to-ship, shore-to-ship, and cloud-to-cloud collaborative data network. Analyze the data transmission quality of vessels navigating in the target waters and the data reception quality of the demonstration vessel to determine the data collection points of the demonstration vessel for vessels navigating in the target waters. The demonstration vessel, based on the shipborne terminal system at the data collection point, acquires the ship navigation status data and first positioning data transmitted back by AIS and VDES from the ships navigating in the target waters. The ship navigation status data includes the ship's position, heading, speed, and heading, which characterize the ship's dynamics, as well as the ship's engine operating status, such as the speed, power, and fuel consumption. The first positioning data consists of the ship's AIS positioning data and satellite positioning data. The demonstration vessel's onboard terminal system uses panoramic cameras to acquire panoramic video data of the vessel's navigation, which is then aggregated to form a multi-source heterogeneous dataset.

3. The ship collision avoidance identification and path planning method based on ship-shore cooperation according to claim 2, characterized in that, The analysis of data transmission quality of vessels navigating in the target waters and data reception quality of demonstration vessels, to determine the data collection points of the demonstration vessels for vessels navigating in the target waters, specifically includes: The signal strength, packet loss rate, and transmission delay of ships navigating in the target waters when transmitting data back to the demonstration ship, as well as the signal reception strength and signal-to-noise ratio of the demonstration ship when receiving data, are obtained. The data transmission quality score between each ship navigating in the target waters and the demonstration ship is calculated in a comprehensive manner. Based on the real-time location information of the demonstration vessel, the historical navigation trajectory of the demonstration vessel in the target waters is obtained, the historical navigation trajectory is divided according to a preset grid, the average quality score of the navigation vessel data received by the demonstration vessel in each grid is calculated, and a data reception quality distribution map is formed. The K-means clustering algorithm is introduced. k grids are randomly selected from the data reception quality distribution map as initial cluster centers. Then, the Euclidean distance from the quality score feature vector of each grid in the distribution map to each initial cluster center is calculated. Each grid is assigned to the cluster containing the nearest cluster center, resulting in k data quality clusters. Select several clusters with the highest average grid quality score within each cluster from k clusters, and determine their corresponding spatial regions as candidate regions for high-quality data collection. Based on the boundary coordinates of the candidate areas for high-quality data acquisition, and combined with the real-time distribution density of ships, coordinate points are calculated within each candidate area to enable the demonstration ship to maintain high-quality communication connections with the maximum number of ships. These coordinate points are then determined as the optimal locations for the demonstration ship to collect data from ships in the target waters.

4. The ship collision avoidance identification and path planning method based on ship-shore cooperation according to claim 1, characterized in that, The process involves calibrating the first positioning data of a vessel navigating in the target waters based on panoramic video data from the demonstration vessel, outputting calibrated second positioning data, and then using this second positioning data and the vessel's navigation status data, combined with a vessel maneuvering characteristic model, to predict the vessel's trajectory within a preset time period, resulting in predicted trajectory data. Specifically: Compare the AIS positioning data in the first positioning data with the satellite positioning data, and identify the navigation vessels whose positioning deviation is greater than a preset threshold as vessels to be corrected for positioning data. Based on the panoramic video data of the demonstration ship, the navigation targets in the video are identified by machine vision recognition algorithm, and the measured visual distance between the ship to be located is corrected by calculating the data to be located based on the panoramic video data. Based on the current satellite positioning data of the demonstration vessel, and according to the measured visual distance, the first positioning data of the vessel to be positioned is corrected, and the calibrated second positioning data is output. If the positioning deviation of the first positioning data is not greater than a preset threshold, the average value of the AIS positioning data and the satellite positioning data is output as the second positioning data. A ship maneuvering characteristic model is constructed. The second positioning data, ship heading, speed and bow direction are used as the initial state input of the model. Based on the ship's tonnage, draft, ship type parameters and current speed, maneuvering characteristic parameters including ship turning performance, stroke and rudder response time are calculated. Acquire real-time navigation environment data, including water flow velocity, flow direction, wind speed, and wind direction, as external disturbance vector input, and calculate the impact of external disturbance on the ship's motion state; A discrete-time Markov chain model is introduced to predict the navigation trajectory of the ship within a preset time period based on the second positioning data, the navigation status data of the ship, and the influence of the ship's operating status, thus obtaining the predicted navigation trajectory data.

5. The ship collision avoidance identification and path planning method based on ship-shore cooperation according to claim 4, characterized in that, The introduction of a discrete-time Markov chain model, based on the second positioning data, the navigation status data of the vessel, and the influence of the vessel's operating status, predicts the navigation trajectory of the vessel within a preset time period, resulting in predicted navigation trajectory data, specifically: The navigation state of the vessel in the target waters is discretized into a state space, where each state is determined by a combination of the vessel's position, heading, and speed. Based on the second positioning data and the vessel's navigation state data, the initial state of the vessel at the current moment is determined. Based on the navigation state data sequence of historical ships, the probability of transitioning from the current state to the next possible state in the state space is statistically analyzed, and a one-step state transition probability matrix is ​​constructed. Using the one-step state transition probability matrix as the model parameter of the Markov chain, and taking the current state of the sailing ship as the initial state, the probability distribution of the sailing ship's state in multiple discrete time steps in the future is calculated through the one-step state transition probability matrix. Based on the influence of the real-time navigation environment factors on the ship's motion state, the probability distribution of each state transition is corrected to obtain the corrected state probability distribution. From the corrected state probability distribution, the state with the highest probability of occurrence at each future time step is selected as the predicted state of the navigating vessel at that time step. Connect the predicted states at all predicted time steps from the current moment to the end of a preset future time period to form a continuous state sequence consisting of the states with the highest probability. Extract the state components representing the ship's position from this sequence to generate the predicted navigation trajectory data of the ship in the preset future time period.

6. The ship collision avoidance identification and path planning method based on ship-shore cooperation according to claim 1, characterized in that, The process of calculating the encounter probability, nearest encounter distance, and time to reach the nearest encounter point between vessels navigating in the target waters based on the predicted navigation trajectory data, and determining the collision risk between vessels based on the encounter data, specifically involves: Based on the predicted navigation trajectory data, calculate the potential intersection points of the predicted trajectories of any two ships within a preset time period in the future, and for each intersection point, calculate the time difference between the arrival of the two ships at the intersection point. If the time difference is not less than the preset safe meeting time window threshold, it is determined that there is a meeting risk between the two ships at the meeting point. The meeting point is marked as a valid meeting point. For each valid meeting point, the probability distribution of the predicted navigation trajectory of the two ships at the meeting point is calculated, and the probability of the two ships actually meeting at the meeting point is calculated based on the probability distribution to obtain the meeting probability. From the effective meeting points, find the minimum geometric distance between the two ships' trajectories as the nearest meeting distance, and calculate the time required for the two ships to reach the nearest meeting point in their current state, thus obtaining the time to reach the nearest meeting point. The encounter probability, the nearest encounter distance, and the time to reach the nearest encounter point are used as encounter data, and the collision risk level between ships is determined based on the encounter data.

7. The ship collision avoidance identification and path planning method based on ship-shore cooperation according to claim 1, characterized in that, If the collision risk exceeds a preset value, the predicted flight trajectory data, second positioning data, collision risk, and pre-set safe encounter distance constraints are analyzed based on the A* algorithm to determine a collision avoidance path. The collision avoidance path includes a heading, speed, and a sequence of turning points, specifically: Vessels with a collision risk greater than the preset value are designated as collision avoidance planning vessels. Based on the second positioning data and predicted navigation trajectory data, the electronic nautical chart model of the target waters is discretized into grid nodes. The grid position in front of the collision avoidance planning vessel is taken as the starting node, and the grid position in the preset collision avoidance waypoint set is taken as the target node. Construct a path search cost function for the A* algorithm. The cost function includes the actual travel cost from the starting node to the current node and the expected travel cost from the current node to the target node estimated by the heuristic function. The pre-set safe encounter distance constraint is transformed into a penalty term in the cost function. In the grid model, the grid occupied by the predicted future navigation trajectory data of the collision avoidance planning vessel and surrounding vessels is marked as a dynamic obstacle area. Based on the A* algorithm, starting from the starting node, the total cost of adjacent grid nodes is iteratively evaluated. The node with the minimum total cost is added to the optimal path set, and the current node is updated to generate a collision avoidance path composed of a sequence of grid nodes. Based on the sequence of grid nodes traversed by the initial collision avoidance path, calculate the heading changes between adjacent nodes and the required speed adjustments, and output a collision avoidance path containing the heading, speed, and turning point sequence.

8. A ship collision avoidance identification and path planning system based on ship-shore cooperation, characterized in that, The ship collision avoidance identification and path planning system based on ship-shore cooperation includes a storage unit and a processor. The storage unit includes a ship collision avoidance identification and path planning method program based on ship-shore cooperation. When the ship collision avoidance identification and path planning method program based on ship-shore cooperation is executed by the processor, it implements the steps of the ship collision avoidance identification and path planning method based on ship-shore cooperation as described in any one of claims 1 to 7.

Images