Ship cooperation behavior identification method and system based on multi-ship geometrical relationship analysis

Abstract

The invention provides a ship cooperation behavior recognition method and system based on multi-ship geometrical relationship analysis, and the method comprises the steps: inputting the related variables of two ships into a trained ship cooperation behavior recognition model, and outputting a result whether the two ships cooperate or not; the relative variables of the two ships comprise the nearest distance delta dis between the two ships, the relative speed delta vel between the two ships, and the relative course delta head, sin (delta head), cos (delta head), delta dis * sin (delta head), delta dis * cos (delta head), delta dis * sin (delta vel) and delta dis * cos (delta vel) between the two ships. The method has the advantage that the characteristics of the ship cooperation behavior can be effectively judged.

Description

Technical Field

[0001] The present application belongs to the field of ship identification technology, and specifically relates to a method and system for identifying ship coordination behavior based on multi-ship geometric relationship analysis. Background Art

[0002] Maritime transport plays an important role in global trade, with about 90% of trade volume relying on sea transport. With the continuous expansion of maritime trade, the number and types of ships and their frequent activities have increased, resulting in increasingly crowded waterways and more complex shipping situations. This trend has significantly increased the risk of navigation safety, and traditional port monitoring and management methods have been unable to cope with current challenges. To this end, more and more ports are gradually moving towards digitalization and intelligence, with the goal of achieving a smarter, greener and more efficient operating environment. In this context, the identification of collaborative behavior of ships has become an important task to improve port operating efficiency and ensure shipping safety. The method of identifying ship cooperation behavior realizes real-time monitoring of ship interaction and cooperation modes by analyzing the spatial and dynamic relationships between ships. This study can not only effectively identify the collaborative behavior of ships, but also provide shipping risk warning and emergency management support, laying a solid foundation for the development of smart ports.

[0003] Multi-ship behavior includes interactions between ships, such as anti-collision operations and collaborative operations. Waterways near ports significantly increase the risk of collision due to complex navigation environments and high-density traffic conditions. By quantifying collision assessment coefficients and generating collision heat maps, it is possible to more effectively identify key conflict areas, assess navigation complexity and operational risks, and thus improve navigation safety. In addition, collaborative behavior mainly involves tugboats assisting large ships in stable navigation near ports. Tugboats play a vital role in the safe berthing and unberthing of large cargo ships and tankers. Due to the large size of these ships and the shallow waters in the port area, the response of the rudder is limited, and tugboats are needed to help maintain a stable course and achieve steering. In emergency situations, such as failure of the ship's main engine, steering gear or power system, tugboats are even more critical and can help ships avoid risks.

[0004] For example, in March 2024, the Dali lost power and collided with the Francis Scott Key Bridge in Baltimore, causing the bridge to collapse, severely affecting transportation and impacting the stability of the supply chain. After guiding the Dali to the main channel, the tugboat returned to the port and did not escort the freighter across the bridge, which was one of the causes of the accident. Therefore, ship behavior recognition plays an important role in ensuring navigation safety, optimizing port operations, and reducing maritime risks.

[0005] At present, the research on multi-ship behavior identification mainly focuses on collision avoidance strategies, while less attention is paid to collaborative behaviors (such as coordinated operations and inter-ship communication). In collision avoidance research, the main methods include calculating conflict factors based on static and dynamic characteristics of ships, or formulating rule constraints according to the International Regulations for Preventing Collisions at Sea (COLREGs). For example, researchers have proposed methods to detect near-collisions of ships using AIS data, and introduced a vessel conflict ranking operator (VCRO) to assess the risk of encounters. Some studies have also extended collision avoidance methods to other related problems to improve their adaptability and performance. However, although research on multi-ship collaboration has been explored, it is usually more generalized. For example, researchers have analyzed multi-ship interactive behaviors, but have not yet fully involved specific collaborative actions. Summary of the invention

[0006] In order to overcome the defects of the prior art, this application proposes a method for identifying ship coordination behavior based on multi-ship geometric relationship analysis, including:

[0007] Input the variables related to the two ships into the trained ship coordination behavior recognition model, and output the result of whether the two ships are cooperating;

[0008] The variables related to the two ships include: the closest distance δ between the two ships dis , the relative speed between the ships δ vel , the relative heading between the two ships δ head 、sin(δ head ), cos(δ head ),δ dis ·sin(δ head ),δ dis ·cos(δ head ),δ dis ·sin(δ vel ) and δ dis ·cos(δ vel ).

[0009] As an improvement of the above method, the ship coordination behavior identification model is a logistic regression model.

[0010] As an improvement of the above method, the closest distance δ between the two ships dis The calculation methods include:

[0011]

[0012] l j =|length j α j |

[0013] α i =|γij -θ i |

[0014] α j =|180°+θ j -γ ij |

[0015] δ ij =GreatCircleDistance(p i ,p j )

[0016]

[0017] Where i is the number of the larger of the two ships; j is the number of the smaller of the two ships; φ i represents the latitude of ship i; λ i represents the longitude of ship i; p i represents the position coordinates of ship i; GreatCircleDistance(·,·) represents the formula for finding the great circle distance between two ships; θ i Indicates the course of ship i relative to the ground; length i Indicates half the length of ship i; width i represents half the width of ship i.

[0018] As an improvement to the above method, the relative speed δ between the ships vel The calculation methods include:

[0019] δ vel =||v ij ||2

[0020] v ij =v j -v i

[0021] v i =(ν i ·cos(θ i ),ν i ·sin(θ i ))

[0022] v j =(ν j ·cos(θ j ),ν j ·sin(θ j ))

[0023] Among them, ν i represents the ground speed of ship i; ||·||2 represents the second norm.

[0024] As an improvement of the above method, the relative heading δ between the two ships head The calculation methods include:

[0025] δ head =|θ i -θ j |.

[0026] The present application also provides a ship coordination behavior identification system based on multi-ship geometric relationship analysis, which is implemented based on the above method, and the system includes:

[0027] The coordination behavior recognition module is used to input the variables related to the two ships into the trained ship coordination behavior recognition model and output the result of whether the two ships are cooperating.

[0028] As an improvement of the above system, the system further comprises:

[0029] The ship coordination behavior recognition model training module is used to train the ship coordination behavior recognition model based on historical data.

[0030] As an improvement of the above system, the system further comprises:

[0031] The variable calculation module is used to calculate the variables related to the two ships, including: the closest distance δ between the two ships dis , the relative speed between the ships δ vel , the relative heading between the two ships δ head 、sin(δ head ), cos(δ head ),δ dis ·sin(δ head ),δ dis ·cos(δ head ),δ dis ·sin(δ vel ) and δ dis ·cos(δ vel ).

[0032] Compared with the prior art, the advantages of this application are:

[0033] 1. Introducing periodic processing of relative heading: In order to effectively capture the cooperative behavior between the towing vessel and the cargo ship or tanker, the relative heading angle (δ head ), replacing the conventional heading angle. Compared with the traditional Fourier series method, this method can reduce the number of parameters and effectively express the periodic characteristics of the heading angle, thereby better identifying periodic cooperative behaviors.

[0034] 2. Processing of nonlinear relationships: By introducing interaction terms (such as δ dis ·sin(δhead ) and δ dis ·cos(δ head ), etc.), this method can handle the nonlinear relationship between variables such as range, speed and heading angle, and improves the accuracy of the model in complex cooperation scenarios, especially in the identification of cooperative behavior between tugboats and cargo ships or tankers.

[0035] 3. L1 regularization and model simplification: Applying L1 regularization in the logistic regression model can not only enhance the sparsity of the model and reduce overfitting, but also automatically select the most relevant features and improve the interpretability of the model. This method helps to extract the most important factors from complex ship cooperation behavior data and improve the generalization ability and accuracy of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 The figure shows the framework diagram of the ship coordination behavior identification method based on multi-ship geometric relationship analysis;

[0037] Figure 2 Shown is a geometric representation of the feature;

[0038] Figure 3 Shown is a relative speed calculation diagram;

[0039] Figure 4(a) shows a partial schematic diagram of cargo ships and oil tankers in a port coordination thermodynamic diagram;

[0040] Figure 4(b) shows a schematic diagram of the towing vessel part in the coordination thermal diagram of a port;

[0041] Figure 4(c) shows a schematic diagram of the dense area of ​​cargo ships and oil tankers in the thermal diagram of a port;

[0042] Figure 4(d) shows a schematic diagram of the towing vessel concentration area in the coordination heat map of a port;

[0043] Figure 4(e) shows a panoramic view of the thermal map of a port. DETAILED DESCRIPTION

[0044] The technical solution of the present application is described in detail below with reference to the accompanying drawings.

[0045] This application proposes a method and system for identifying ship coordination behavior based on multi-ship geometric relationship analysis. It enhances the nonlinear processing capability of the logistic regression model through feature mapping, realizes the identification of ship coordination behavior under low-label data, and can accurately identify key coordination positions and behavior patterns. Combined with the optimized algorithm architecture for specific scenarios, it can achieve efficient and accurate identification of ship coordination behavior, facilitate real-time monitoring and dynamic response, and provide strong technical support for the safe operation and emergency response of smart ports.

[0046] The purpose of this application is to solve the problem of imperfect identification of collaborative behaviors in the current high-density ship environment in ports and waterways. With the continuous increase in global maritime transport volume, waterways are becoming increasingly crowded, the types and numbers of ships are increasing, and the navigation dynamics are becoming more complex, increasing the risk of conflicts and accidents between ships. Traditional port monitoring methods mainly rely on manual monitoring and basic rule-making, and it is difficult to respond to complex ship collaboration scenarios in real time. At the same time, existing collision avoidance research mainly focuses on single-ship collision avoidance and risk assessment, lacks in-depth analysis of multi-ship collaborative interactive behaviors, and cannot fully support the efficient management needs of intelligent ports. Therefore, the identification method for multi-ship behavior, especially collaborative and coordinated behavior, has become a key technical requirement for the construction of intelligent ports.

[0047] This application is based on the geometric relationship between multiple ships. By introducing spatial and dynamic feature parameters to construct an analysis model, it can identify the interaction and coordination patterns between ships in real time and effectively judge the characteristics of ship collaborative behavior. For example, by analyzing the geometric characteristics of the relative position, speed, and heading of the ships, the method of this application can quickly identify common collaborative behaviors such as tugboats assisting large ships in berthing and leaving ports. In addition, the collaborative heat map and behavior warning generated by this identification method can effectively help port dispatchers quickly identify key conflict areas and potential risks, while improving navigation safety, supporting dynamic monitoring and early warning of ship behavior, and enhancing the efficiency of emergency response.

[0048] The application of this application is not only to improve the safety and reliability of maritime transportation, but more importantly to provide technical support for the operation and management of smart ports. Its method is highly automated and real-time, which can significantly reduce the labor cost of port management, and will be widely used in major ports and waterways around the world in the future, providing reliable technical support for the intelligent development of maritime transportation.

[0049] Example 1

[0050] like Figure 1 As shown, the method for identifying ship cooperation behavior based on multi-ship geometric relationship analysis provided by the present application includes: first, historical data is processed through steps such as trajectory ship extraction, trajectory interpolation and time alignment to provide high-quality input for subsequent model training. Then, the data is labeled using expert knowledge, and nonlinear regression and logistic regression are performed in combination with prior knowledge to estimate key parameters and train the logistic regression model. Finally, after the model is trained, it can be directly applied to real-time data, and classification is used to determine whether the ship is in cooperative behavior, thereby achieving real-time monitoring and behavior identification. The cooperative behavior of the present application includes pushing and pulling ships working together with tugboats, or tugboats cooperating with large ships, etc.

[0051] 1. Problem Modeling

[0052] Multi-ship behavior mainly depends on the interaction characteristics between ship pairs, including relative position, speed and heading differences. The classification task can be regarded as learning a function M to estimate whether two ships are cooperating based on these differences. Therefore, the classification expression of this application can be expressed as:

[0053] M((δ dis ) -1 ,(δ vel ) -1 ,g(δ head ))→{0,1} (1)

[0054] Where M is based on the position difference between the two ships (δ dis ), speed difference (δ vel ) and heading difference (δ head ), learn to predict whether two ships are cooperating (1) or not cooperating (0). Other parameters used in this application and their explanations are shown in Table 1.

[0055] Table 1 Parameters used in this application and their meanings

[0056]

[0057] Distance δ dis : This is the shortest distance between the two ships. Since distance alone cannot fully describe the complexity of cooperative behavior, it is necessary to combine other parameters for analysis. Generally, as the distance between ships increases, the likelihood of cooperation decreases. In mathematical models, this relationship is expressed as cooperative behavior being inversely proportional to distance.

[0058] Relative speed δ vel : The rate of change of the distance between the two ships is also a key factor in the model, which is calculated based on the true heading and actual speed of the ships. Since it is difficult for ships to perform complex cooperative operations when the speed difference is large, the higher the relative speed, the lower the possibility of cooperation. Therefore, this parameter is inversely proportional to the cooperative behavior.

[0059] Relative heading δ head : reflects the heading change between the two ships and determines the degree of heading adjustment required for cooperative action. Relative heading is defined here as an angle and its direction, with a value range of [0,2π]. In cooperative operations such as towing or pushing, even if the headings are opposite, the possibility of cooperation cannot be ruled out. Considering the periodic characteristics of heading, assume a periodic function g(δ head ) to describe the variable.

[0060] 2. Feature calculation

[0061] Cooperative behavior usually occurs when a tugboat assists a larger vessel. disWhen δ is used, the size of the ship needs to be taken into account. We model each ship as an ellipse based on its length and width, and consider the center of the ship as the center of the ellipse. dis The calculation is as follows Figure 2 First, use formula (2) to calculate the azimuth angle γ of ship i relative to ship j: ij , whose range is [0,360). At the same time, calculate the great circle distance δ between the two ships ij Next, use equations (4) and (5) to calculate the angle α i and α j , represents the heading and azimuth of each ship γ ij The angle between them. Because these angles are used for trigonometric functions and absolute value calculations, they can be simplified. In this application, ship i is considered a larger ship and ship j is a tugboat. For larger ships, the closest distance is determined by the azimuth angle γ ij The intersection with the boundary of the ellipse is determined by (Equation (6)). For a tugboat, which usually contains an engine casing, we use its heading at the azimuth angle γ ij The projection on is taken as the closest distance (Formula (7)). Finally, the hull distance is subtracted from the distance between the two ships to obtain the closest distance δ between the two ships. dis (Formula (8)).

[0062]

[0063] δ ij =GreatCircleDistance(p i ,p j ) (3)

[0064] α i =|γ ij -θ i | (4)

[0065] α j =|180°+θ j -γ ij | (5)

[0066]

[0067] l j =|length j α j | (7)

[0068]

[0069] Among them, length i Indicates half the length of ship i; width i represents half the width of ship i.

[0070] The relative speed between two ships is calculated as Figure 3 As shown. We treat the velocity as a vector and calculate the difference between these vectors. First, decompose the velocity into sine and cosine components (Formulas (9) and (10)), and then calculate the norm of the velocity difference vector (Formulas (11) and (12)):

[0071] v i =(ν i ·cos(θ i ),ν i ·sin(θ i )) (9)

[0072] v j =(ν j ·cos(θ j ),ν j ·sin(θ j )) (10)

[0073] v ij =v j -v i (11)

[0074] δ vel =||v ij ||2(12)

[0075] Here, ||·||2 represents the second norm.

[0076] The relative heading between the two ships is calculated directly from the heading difference, as shown in formula (13):

[0077] δ head =|θ i -θ j | (13)

[0078] 3. Logistic regression modeling

[0079] Logistic regression is a widely used statistical method for binary classification tasks. The model uses a logistic function to map a linear combination of input features to a value between 0 and 1, which can be interpreted as a probability. The classic logistic regression model is expressed as:

[0080]

[0081] Among them, p(y|x) represents the probability of output y under a given input vector x, w is the weight vector, e is the base of the natural logarithm, and w T x represents a linear combination of the input features. Building this linear relationship is key when applying logistic regression to a specific problem.

[0082] When applying logistic regression to identify cooperative behavior, the focus is on capturing the periodicity of "relative heading" and the nonlinear relationship between variables while maintaining the original form. For example, when a tugboat cooperates with a larger ship, its push and pull actions from all directions show periodicity, which is mainly reflected in the "relative heading". It is difficult to capture this periodic behavior by directly using the heading angle as a variable. Therefore, this application uses sine and cosine decomposition (see Algorithm 1 in Table 2, training step 2: sin(δ head ) and cos(δ head )) instead of the commonly used Fourier series method to reduce the number of parameters and capture the periodic nature of heading.

[0083] The effects of distance and relative speed on cooperative behavior are closely related to relative heading. Using these three variables directly in logistic regression cannot capture nonlinear relationships and may have a negative impact on behavior identification. For example, even if the distance remains unchanged, the impact of tugboats on large ships will vary significantly depending on the direction. Therefore, in addition to using δ dis ,δ vel and δ head In addition to the sine and cosine decomposition of , it also includes interaction terms (see Algorithm 1, training step 2 in Table 2), such as δ dis ·sin(δ head ),δ dis ·cos(δ head ),δ dis ·sin(δ vel ) and δ dis ·cos(δ vel ) to enhance the model's ability to handle the nonlinear relationships between these key variables.

[0084] In addition, L1 regularization is used to encourage the sparsity of the model, thereby selecting the most relevant features, improving interpretability and reducing overfitting. Specifically, the model parameters β0 to β9 and λ are used to capture these relationships, and the model is trained on a labeled dataset (see Algorithm 1, training in Table 2). After training, the constructed formula can be directly used for behavior classification (see Algorithm 1, testing in Table 2). This application focuses on the cooperation between cargo ships, tankers, and tugboats because these interactions are more typical and easy to annotate.

[0085] Table 2 Logistic regression model training and testing algorithm

[0086]

[0087] This application uses the original data provided by NOAA (National Oceanic and Atmospheric Administration) to test the above method. The data covers a one-year period near a port, with a longitude range of -118.36 to -118.0 and a latitude range of 33.55 to 33.80. After completing the preliminary track cleaning, the ship track is divided into segments of 20 points. If there are other track points within 3 kilometers of the 5th point of the target ship, they are considered to be the same scene. Ten representative scenes were selected for marking cooperative behaviors, such as tugboats assisting large ships to enter the port or turn. At the same time, some scenes with frequent tugboat activities are used for test analysis. Cooperative behavior is marked as 1 and non-cooperative behavior is marked as 0. The preparation and division of these data sets provide a solid foundation for research.

[0088] This application evaluates the ship coordination behavior recognition performance through the following key indicators: accuracy, precision, recall and F1 value. The calculation method of each indicator is as follows:

[0089] Accuracy: measures the proportion of correct predictions to the total number of predictions. The calculation formula is:

[0090]

[0091] Among them, T P is the number of true positives, T N is the number of true negatives, F P is the number of false positives, F N is the number of false negatives.

[0092] Precision: Evaluates the accuracy of positive predictions, reflecting the proportion of correctly predicted positive instances. The calculation formula is:

[0093]

[0094] Recall: It evaluates the model's ability to identify all positive instances, indicating the proportion of correctly identified positive instances to all actual positive instances. The calculation formula is:

[0095]

[0096] F1-score: The harmonic mean of precision and recall, providing a balance between the two, especially suitable for situations where the category distribution is uneven. The calculation formula is:

[0097]

[0098] These indicators comprehensively measure the performance of the model in identifying different ship coordination behaviors.

[0099] In maritime operations, tugboats play a vital role in assisting cargo ships and tankers in port operations. The method provided in this application focuses on the cooperation dynamics between these ships. When tugboats collaborate with cargo ships and tankers, the stability of the cargo and the maneuverability of the ship must be ensured, which requires precise control of the role and position of each tugboat. The cooperation between tugboats is different from the cooperation between tugboats and cargo ships or tankers. Therefore, we divide the data from ten annotated scenarios into two categories according to the type of ship cooperation. Of the 2,381 trajectory point pairs, 818 point pairs involve cooperation between tugboats and cargo ships or tankers (275 cooperative points and 543 non-cooperative points), while 1,563 point pairs involve cooperation between tugboats (173 cooperative points and 1,390 non-cooperative points). In order to ensure the quality of training, a division of 90% training and 10% testing is used for the less common tugboat and cargo ship / tanker pairs; and for the more frequent tugboat and tugboat pairs, a division of 80% training and 20% testing is used.

[0100] Table 3 shows the learned parameters of tug-tug cooperation, indicating that the values ​​of these parameters are generally small. In particular, parameters 1, 2, and 3 are all negative, which is consistent with our hypothesis that the closer the ships are, the smaller the difference in speed and direction, and the greater the possibility of cooperative behavior. Parameters 4 and 5 have opposite signs and can be approximated as a tangent relationship, indicating a periodic relationship between cooperation and angle. They are the largest among all parameters, indicating that angle plays a key role in tug-tug interaction, where push and pull assistance in different directions is crucial.

[0101] Table 3 Parameters between tugboats

[0102]

[0103] In contrast, in the cooperation between cargo ship (or tanker) and tugboat, Table 4 shows that the most important features are distance, speed and angle. After applying regularization, the sine and cosine components of the angle absorb the effects of parameters 4 and 5 through interaction term parameters 6 to 9, simplifying these parameters. By comparing the parameters of these two types of cooperation, the parameters of tugboat-tugboat cooperation are smaller, indicating that the model for this scenario is relatively simple and may be underfitting. In this case, further adding regularization terms to limit the parameter size may reduce the performance of the model because it may not fully capture the complexity of the data.

[0104] Table 4 Parameters between tugboats and large ships (tank ships and cargo ships)

[0105]

[0106] To evaluate the performance of the model, we conducted an accuracy evaluation on the test set, as shown in Table 5. The accuracy of tugboat-tugboat cooperation is high, reaching 93.930, but the F1 score is relatively low, at 75.862. This difference may be due to the complexity of the interaction between tugboats. For example, when tugboats are stationed at the entrance of the port, it may be very difficult to determine whether they are cooperating until they actually start to cooperate with the cargo ship or tanker, even if they are close to each other. On the other hand, the model performs very well in identifying the cooperation of tugboats with cargo ships or tankers, with an accuracy of 93.902 and an F1 score of 91.228. This shows that the model is particularly effective in identifying simpler cooperation scenarios where tugboats assist larger ships.

[0107] Table 5 Model recognition accuracy

[0108]

[0109] Cooperative behavior mainly involves collision avoidance and coordinated actions between ships. Some researchers have proposed an excellent method to calculate the collision factor using the KF-VCRO model. This application directly applies this method to identify high-frequency collision areas near the Port of Los Angeles, and uses our proposed method to identify cooperative behavior. Due to the complexity and variability of cooperative behavior, directly using KF-VCRO may mistakenly identify cooperative points as potential collision trajectory pairs. Therefore, if the distance between the two ships is less than 1 km, we first determine whether it is a collision behavior.

[0110] To investigate further, we visualized the trajectories identified as cooperative behaviors using heat maps, showing cargo ships, tankers, and tugboats, respectively, e.g. Figure 4(a)-Figure 4(e) By comparing the offshore heatmaps in Figure 4(a) and Figure 4(b), we can observe that cooperative behavior occurs in two anchorages ① and ②, although not frequently, indicating that coordination in these areas is more limited. In addition, in area ② of Figure 4(b), a tugboat is seen heading out to sea, indicating that the tugboat is deployed to assist other ships that encounter emergencies in deep waters.

[0111] Compared with the offshore area, the density of cooperative behavior is higher in the port area, as shown in Figure 4(c) and Figure 4(d). In area ① of Figure 4(c), the sparse trajectory points indicate a berthing area, as confirmed by Figure 4(c) and Figure 4(d), but there are fewer cooperative activities, indicating that these areas are mainly used by small cargo ships and tankers and do not require much assistance from tugboats. In addition, area ② in Figure 4(c) shows that tugboats begin to assist ships even when they are at anchor, which may involve installing towing equipment or boarding and piloting tasks. By comparing area ① of Figure 4(d) with Figure 4(c), the density of tugboats in the same area is significantly increased, indicating that large ships require a lot of assistance during the berthing process. Considering area ② of Figure 4(c) and the berthing area of ​​the port, this area may be a berthing point for tugboats or an area where tugboats assist other types of ships to enter the port, which deserves further analysis to confirm.

[0112] Example 2

[0113] The present application also provides a ship coordination behavior identification system based on multi-ship geometric relationship analysis, which is implemented based on the above method, and the system includes:

[0114] The coordination behavior recognition module is used to input the variables related to the two ships into the trained ship coordination behavior recognition model and output the result of whether the two ships are cooperating.

[0115] The ship coordination behavior recognition model training module is used to train the ship coordination behavior recognition model based on historical data.

[0116] The variable calculation module is used to calculate the variables related to the two ships, including: the closest distance δ between the two ships dis , the relative speed between the ships δ vel , the relative heading between the two ships δ head 、sin(δ head ), cos(δ head ),δ dis ·sin(δ head ),δ dis ·cos(δ head ),δ dis ·sin(δ vel ) and δ dis ·cos(δ vel ).

[0117] The present application may also provide a computer device, comprising: at least one processor, a memory, at least one network interface and a user interface. The various components in the device are coupled together through a bus system. It is understood that the bus system is used to achieve connection and communication between these components. In addition to the data bus, the bus system also includes a power bus, a control bus and a status signal bus.

[0118] The user interface may include a display, a keyboard or a pointing device, such as a mouse, a trackball, a touch pad or a touch screen.

[0119] It is understood that the memory in the embodiments disclosed in the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus random access memory (DRRAM). The memories described herein are intended to include, but are not limited to, these and any other suitable types of memories.

[0120] In some embodiments, the memory stores the following elements, executable modules or data structures, or a subset thereof, or an extended set thereof: an operating system and applications.

[0121] The operating system includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., which are used to implement various basic services and process hardware-based tasks. The application includes various application programs, such as a media player (Media Player), a browser (Browser), etc., which are used to implement various application services. The program for implementing the method of the embodiment of the present disclosure can be included in the application.

[0122] In the above embodiment, the processor may also call a program or instruction stored in the memory, specifically, a program or instruction stored in an application program, and is used to:

[0123] Execute the steps of the above method.

[0124] The above method can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor may be a general processor, a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The above-disclosed methods, steps and logic block diagrams can be implemented or executed. The general processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the above-disclosed method can be directly embodied as a hardware decoding processor to execute, or the hardware and software modules in the decoding processor are combined to execute. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0125] It is understood that the embodiments described in the present application can be implemented by hardware, software, firmware, middleware, microcode or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in the present application or a combination thereof.

[0126] For software implementation, the technology of the present application can be implemented by executing the functional modules (such as procedures, functions, etc.) of the present application. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or outside the processor.

[0127] The present application may also provide a non-volatile storage medium for storing a computer program. When the computer program is executed by a processor, each step in the above method embodiment can be implemented.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present application and are not intended to limit it. Although the present application is described in detail with reference to the embodiments, a person skilled in the art should understand that any modification or equivalent replacement of the technical solution of the present application does not depart from the spirit and scope of the technical solution of the present application and should be included in the scope of the claims of the present application.

Claims

1. A method for identifying ship coordination behavior based on multi-ship geometric relationship analysis, comprising: Input the variables related to the two ships into the trained ship coordination behavior recognition model, and output the result of whether the two ships are cooperating; The variables related to the two ships include: the closest distance δ between the two ships dis , the relative speed between the ships δ vel , the relative heading between the two ships δ head 、sin(δ head ), cos(δ head ),δ dis ·sin(δ head ),δ dis ·cos(δ head ),δ dis ·sin(δ vel ) and δ dis ·cos(δ vel ).

2. The method for identifying ship coordination behavior based on multi-ship geometric relationship analysis according to claim 1 is characterized in that: The ship coordination behavior identification model is a logistic regression model.

3. The method for identifying ship coordination behavior based on multi-ship geometric relationship analysis according to claim 1 is characterized in that: The minimum distance δ between the two ships dis The calculation methods include: l j =|length j ·α j | a i =|γ ij -θ i | a j =|180°+θ j -c ij | δ ij =GreatCircleDistance(p i ,p j ) Where i is the number of the larger of the two ships; j is the number of the smaller of the two ships; φ i represents the latitude of ship i; λ i represents the longitude of ship i; p i represents the position coordinates of ship i; GreatCircleDistance(·,·) represents the formula for finding the great circle distance between two ships; θ i Indicates the course of ship i relative to the ground; length i Indicates half the length of ship i; width i represents half the width of ship i.

4. The method for identifying ship coordination behavior based on multi-ship geometric relationship analysis according to claim 3 is characterized in that: The relative speed δ between the ships vel The calculation methods include: d vel =||v ij ||2 v ij =v j -v i v i =(n i ·cos(θ i ),n i ·sin(θ i )) v j =(n j ·cos(θ j ),n j ·sin(θ j )) Among them, ν i represents the ground speed of ship i; ||·||2 represents the second norm.

5. The method for identifying ship coordination behavior based on multi-ship geometric relationship analysis according to claim 4 is characterized in that: The relative heading δ between the two ships head The calculation methods include: d head =|θ i -θ j |。 6. A ship coordination behavior recognition system based on multi-ship geometric relationship analysis, implemented based on any of the methods described in claims 1-5, characterized in that: The system comprises: The coordination behavior recognition module is used to input the variables related to the two ships into the trained ship coordination behavior recognition model and output the result of whether the two ships are cooperating.

7. The ship coordination behavior identification system based on multi-ship geometric relationship analysis according to claim 6 is characterized in that: It is characterized in that The system further comprises: The ship coordination behavior recognition model training module is used to train the ship coordination behavior recognition model based on historical data.

8. The ship coordination behavior identification system based on multi-ship geometric relationship analysis according to claim 6 is characterized in that: It is characterized in that The system further comprises: The variable calculation module is used to calculate the variables related to the two ships, including: the closest distance δ between the two ships dis , the relative speed between the ships δ vel , the relative heading between the two ships δ head 、sin(δ head ), cos(δ head ),δ dis ·sin(δ head ),δ dis ·cos(δ head ),δ dis ·sin(δ vel ) and δ dis ·cos(δ vel ).

Images