A method and apparatus for predicting behavior of a ship

Abstract

The application discloses a ship behavior prediction method and device, and the method comprises the following steps: collecting target ship trajectory data; inputting the target ship trajectory data into a trained ship pattern recognition model to identify the traffic mode of the target ship trajectory data; selecting a plurality of first historical ship trajectory data from a historical ship trajectory data set with the same traffic mode as the target ship trajectory data; training a ship behavior prediction model based on the first historical ship trajectory data to obtain a trained ship behavior prediction model; inputting the first historical ship trajectory data into the trained ship behavior prediction model to predict each first historical ship trajectory data; and calculating the longitude, latitude, speed and heading of the target ship in a future period of time based on the predicted each first historical ship trajectory data. The application can realize more accurate and more interpretable ship behavior prediction.

Description

Technical Field

[0001] This invention relates to the field of ship behavior prediction technology, and in particular to a ship behavior prediction method and apparatus. Background Technology

[0002] With the growth of shipping operations, the inland waterway traffic environment is becoming increasingly complex, making accurate prediction of vessel behavior crucial. Driven by intelligent and information-based technologies, proactive traffic management is emerging as a trend, and vessel behavior inference technology is key to achieving this goal.

[0003] Currently, vessel traffic management relies primarily on passive monitoring, which suffers from problems such as delayed response, low efficiency, and difficulties in emergency handling. There is an urgent need to improve regulatory effectiveness through proactive management. Proactive maritime traffic management can effectively improve regulatory efficiency, optimize resource allocation, and reduce accident risks. However, the low density and complexity of massive AIS (Automatic Identification System) data make vessel behavior inference a technical bottleneck, especially given the differences in vessel movement patterns and traffic characteristics in inland waterways, which further increases the difficulty of analysis. Existing methods struggle to extract vessel behavior characteristics accurately and in real time, affecting the efficiency of traffic situation warnings and decision-making. Therefore, how to utilize data mining techniques to accurately infer vessel behavior and improve regulatory efficiency and safety has become a core issue that urgently needs to be addressed.

[0004] To address the challenges in ship behavior inference, existing research has employed methods such as cluster analysis for traffic pattern segmentation. However, these methods are inefficient when handling large-scale data and fail to fully utilize static features. Furthermore, dynamic-based Kalman filtering and Markov chain methods are widely used for ship behavior prediction, but they rely on idealized state assumptions and struggle to adapt to complex real-world scenarios. While machine learning methods have been applied, their applicability and interpretability vary across different scenarios. Neural network-based methods perform well when handling complex data, but their widespread application is limited by the difficulty of model building, high dataset requirements, and lack of interpretability. These shortcomings indicate that while existing research provides preliminary technical support for ship behavior inference, challenges remain regarding efficiency, accuracy, and interpretability. Summary of the Invention

[0005] This invention provides a method and apparatus for predicting ship behavior, making the prediction of ship behavior more accurate and more interpretable.

[0006] According to one aspect of the present invention, a method for predicting ship behavior is provided, comprising:

[0007] Collect target vessel trajectory data, which includes at least the vessel's sailing time, latitude and longitude, speed, and heading;

[0008] The target vessel trajectory data is input into the trained vessel pattern recognition model to identify the traffic pattern of the target vessel trajectory data; the same traffic pattern includes vessel trajectory data with the same navigation route;

[0009] From a set of historical ship trajectory data that have the same traffic pattern as the target ship trajectory data, select several first historical ship trajectory data; the first historical ship trajectory data includes at least the sailing time, latitude and longitude, speed and heading.

[0010] A ship behavior prediction model is trained based on the first historical ship trajectory data to obtain the trained ship behavior prediction model.

[0011] The first historical ship trajectory data is input into the trained ship behavior prediction model to predict each of the first historical ship trajectory data.

[0012] Based on the predicted trajectory data of each of the first historical ships, the latitude, longitude, speed, and heading of the target ship for a future period of time are calculated.

[0013] Optionally, before inputting the target ship trajectory data into the trained ship pattern recognition model, the method further includes:

[0014] Historical ship trajectory data is acquired, and the historical ship trajectory data is divided into multiple traffic modes. The historical ship trajectory data includes at least the ship's sailing time, latitude and longitude, speed and heading.

[0015] The traffic patterns of each historical ship trajectory data point are labeled, and a first dataset is constructed using the labeled historical ship trajectory data.

[0016] The ship pattern recognition model is trained based on the first dataset to obtain the trained ship pattern recognition model.

[0017] Optionally, before acquiring historical ship trajectory data, the method further includes:

[0018] Historical vessel automatic identification system (AIS) data is collected, and the historical vessel AIS data is preprocessed to obtain historical vessel trajectory data; the preprocessing includes at least data cleaning and trajectory extraction.

[0019] Optionally, the preprocessing of the historical vessel automatic identification system data to obtain the historical vessel trajectory data includes:

[0020] Data cleaning is performed on the historical Automatic Identification System (AIS) data.

[0021] Interpolation algorithms are used to repair missing data points in the historical Automatic Identification System (AIS) data.

[0022] The historical vessel automatic identification system data, after data cleaning, is processed into trajectory segments to obtain multiple historical vessel trajectory data.

[0023] Optionally, the process of dividing the ship trajectory data into multiple traffic modes includes:

[0024] Determine the departure and arrival areas for ship navigation based on the geographical information of the waterway;

[0025] The starting and ending points of the historical ship trajectory data are matched with the departure area and the arrival area, and the historical ship trajectory data whose starting point belongs to the departure area and whose ending point belongs to the arrival area are divided into parent traffic modes.

[0026] The historical ship trajectory data with the same navigation route in the parent traffic mode are divided into the same sub-traffic mode, and each sub-traffic mode can be used as a traffic mode.

[0027] Optionally, the step of selecting a number of first historical ship trajectory data from a set of historical ship trajectory data with the same traffic pattern as the target ship trajectory data includes:

[0028] Calculate the distance between the target vessel trajectory data and the historical vessel trajectory data with the same traffic pattern, and select the n closest historical vessel trajectory data from the historical vessel trajectory data set as the first historical vessel trajectory data.

[0029] Optionally, training the ship behavior prediction model based on the first historical ship trajectory data to obtain the trained ship behavior prediction model includes:

[0030] A second dataset is constructed using the first historical ship trajectory data;

[0031] The ship behavior prediction model is trained based on the second dataset. The ship behavior prediction model is based on a Gaussian process regression model. The prediction process of the Gaussian process regression model reflects the trend of historical ship trajectories following the same traffic pattern.

[0032] Optionally, the step of inputting the first historical ship trajectory data into the trained ship behavior prediction model to predict each of the first historical ship trajectory data includes:

[0033] The predicted sailing time of the first historical ship trajectory data is input into the trained ship behavior prediction model to predict the latitude, longitude, speed and heading of each first historical ship trajectory data at each moment within the predicted sailing time; the predicted sailing time is the time period that the target ship needs to predict.

[0034] Optionally, based on each predicted first historical ship trajectory data, the latitude, longitude, speed, and heading of the target ship over a future period are calculated, including:

[0035] By averaging the latitude, longitude, speed, and heading values ​​at each moment during the predicted navigation time for each of the first historical ship trajectory data, the latitude, longitude, speed, and heading of the target ship at each moment during the predicted navigation time can be obtained.

[0036] According to another aspect of the present invention, a ship behavior prediction device is provided, comprising:

[0037] The data acquisition unit is used to collect target vessel trajectory data, which includes at least the vessel's sailing time, latitude and longitude, speed, and heading.

[0038] The first training unit is used to input the target ship trajectory data into the trained ship pattern recognition model to identify the traffic pattern of the target ship trajectory data; the same traffic pattern includes ship trajectory data with the same navigation route;

[0039] The data selection unit is used to select a number of first historical ship trajectory data from a set of historical ship trajectory data that have the same traffic pattern as the target ship trajectory data; the first historical ship trajectory data includes at least the sailing time, latitude and longitude, speed and heading;

[0040] The second training unit is used to train the ship behavior prediction model based on the first historical ship trajectory data to obtain the trained ship behavior prediction model.

[0041] The first prediction unit is used to input the first historical ship trajectory data into the trained ship behavior prediction model and predict each of the first historical ship trajectory data.

[0042] The second prediction unit is used to calculate the latitude, longitude, speed and heading of the target ship in the future period based on each predicted first historical ship trajectory data.

[0043] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0044] At least one processor; and

[0045] A memory communicatively connected to the at least one processor; wherein,

[0046] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the ship behavior prediction method according to any embodiment of the present invention.

[0047] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the ship behavior prediction method according to any embodiment of the present invention.

[0048] The technical solution of this invention first uses a ship pattern recognition model to identify the navigation route of the ship, and then selects historical ship trajectories that are the same as the navigation route of the ship to train the ship behavior prediction model. The historical ship trajectories contain the latitude, longitude, speed, and heading data of the ship at each moment, so that the trained ship behavior prediction model can identify the latitude, longitude, speed, and heading data of the ship at each specific moment in the navigation route, thereby accurately predicting the behavioral characteristics of the target ship. In addition, in this application, the first historical ship trajectory data is used as the input of the ship behavior prediction model. Its prediction process reflects the trend of the target ship trajectory following the historical ship trajectory on the same navigation route. Then, the latitude, longitude, speed, and heading of the target ship in the future period of time are predicted by the historical ship trajectory data, so that the prediction results have good interpretability.

[0049] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a flowchart of a ship behavior prediction method provided in Embodiment 1 of the present invention;

[0052] Figure 2 This is a flowchart of a ship behavior prediction method provided in Embodiment 2 of the present invention;

[0053] Figure 3 This is a schematic diagram of an embodiment of the present invention for traffic mode segmentation of trajectory data;

[0054] Figure 4 This is a schematic diagram of traffic pattern recognition and ship behavior prediction in this invention;

[0055] Figure 5 This is a flowchart of one embodiment of ship behavior prediction in this invention;

[0056] Figure 6 This is a structural diagram of a ship behavior prediction device provided in Embodiment 3 of the present invention;

[0057] Figure 7 This is a schematic diagram of the structure of an electronic device that implements the ship behavior prediction method of this invention. Detailed Implementation

[0058] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0059] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0060] Example 1

[0061] Figure 1 The flowchart illustrates a ship behavior prediction method provided in Embodiment 1 of the present invention. Figure 1 As shown, the method includes:

[0062] S101. Collect target vessel trajectory data, which includes at least the vessel's sailing time, latitude and longitude, speed, and heading.

[0063] It should be noted that the target vessel trajectory data is the trajectory already traversed on the current navigation route. In this embodiment, the vessel trajectory data consists of several trajectory data points. For example, the vessel's current latitude, longitude, speed, and heading can be collected every 30 seconds or 1 minute. The collected data points reflect a series of data points on the vessel's navigation path, thus forming the vessel trajectory data. Each data point contains the vessel's current navigation time, latitude, longitude, speed, and heading.

[0064] S102. Input the target ship trajectory data into the trained ship pattern recognition model to identify the traffic pattern of the target ship trajectory data; the same traffic pattern includes ship trajectory data with the same navigation route.

[0065] Among them, the ship pattern recognition model is used to identify the traffic patterns of the target ship's trajectory data, and the traffic patterns reflect the current navigation route of the target ship.

[0066] In this embodiment, multiple historical ship trajectory data can be used to train the ship pattern recognition model. Each historical ship trajectory data is labeled with a traffic pattern tag, so that the trained ship pattern recognition model can identify the traffic pattern of the target ship trajectory data.

[0067] S103. Select several first historical ship trajectory data from the set of historical ship trajectory data with the same traffic pattern as the target ship trajectory data; the first historical ship trajectory data includes at least the sailing time, latitude and longitude, speed and heading.

[0068] S104. Train the ship behavior prediction model based on the first historical ship trajectory data to obtain the trained ship behavior prediction model.

[0069] To make the prediction of ship behavior interpretable, this embodiment uses historical ship trajectory data with the same traffic patterns as the target ship trajectory data as a training set to train the ship behavior prediction model, so that the target ship trajectory follows the trend of historical ship trajectories on the same navigation route.

[0070] It should be noted that, since the traffic pattern recognition model can only identify the navigation route of the target vessel, but cannot identify the latitude, longitude, speed, and heading of the target vessel at every moment in the future, this application also needs to train the vessel behavior prediction model to predict the navigation process of the target vessel in the future.

[0071] S105. Input the first historical ship trajectory data into the trained ship behavior prediction model to predict each of the first historical ship trajectory data.

[0072] S106. Based on the predicted trajectory data of each of the first historical ships, calculate the latitude, longitude, speed and heading of the target ship for a future period of time.

[0073] It should be noted that this embodiment uses historical ship trajectory data with the same traffic patterns as the target ship trajectory data as input to the ship behavior prediction model, and calculates the target ship's latitude, longitude, speed, and heading over a future period based on multiple output results. Specifically, the method used to calculate the target ship's latitude, longitude, speed, and heading over a future period can be a weighted average of multiple output results, i.e., the target ship's trajectory follows the trend of historical ship trajectories along the same navigation route.

[0074] The technical solution of this invention first uses a ship pattern recognition model to identify the navigation route of the ship, and then selects historical ship trajectories that are the same as the navigation route of the ship to train the ship behavior prediction model. The historical ship trajectories contain the latitude, longitude, speed, and heading data of the ship at each moment, so that the trained ship behavior prediction model can identify the latitude, longitude, speed, and heading data of the ship at each specific moment in the navigation route, thereby accurately predicting the behavioral characteristics of the target ship. In addition, in this application, the first historical ship trajectory data is used as the input of the ship behavior prediction model. Its prediction process reflects the trend of the target ship trajectory following the historical ship trajectory on the same navigation route. Then, the latitude, longitude, speed, and heading of the target ship in the future period of time are predicted by the historical ship trajectory data, so that the prediction results have good interpretability.

[0075] Example 2

[0076] Figure 2 This is a flowchart of a ship behavior prediction method provided in Embodiment 2 of the present invention. Figure 2 As shown, the method includes:

[0077] S201. Collect historical ship automatic identification system data, preprocess the historical ship automatic identification system data to obtain the historical ship trajectory data; the preprocessing includes at least data cleaning and trajectory extraction.

[0078] During the transmission and acquisition of Automatic Identification System (AIS) data, issues such as signal obstruction, equipment malfunction, and delays may occur, leading to varying degrees of data loss and errors. Therefore, data preprocessing should be performed before using AIS data.

[0079] In one embodiment, preprocessing the historical vessel automatic identification system data to obtain the historical vessel trajectory data specifically includes:

[0080] Data cleaning is performed on the historical Automatic Identification System (AIS) data.

[0081] Interpolation algorithms are used to repair missing data points in the historical Automatic Identification System (AIS) data.

[0082] The historical vessel automatic identification system data, after data cleaning, is processed into trajectory segments to obtain multiple historical vessel trajectory data.

[0083] Specifically, for AIS data containing erroneous data due to input errors or reception errors, the following should be removed: Data with an MMSI field not exceeding nine digits should be discarded; duplicate data with a speed greater than 2 knots but whose latitude and longitude positions are not updated should be discarded; and data with a heading field outside the 0-360° range should be discarded. Other common erroneous data include position drift and missing data, which can be repaired using other methods.

[0084] Interpolation algorithms are used to repair data defects that exist in the data itself or are caused by data cleaning. This application interpolates based on the mode of the time interval between adjacent trajectory points for each trajectory, taking into account the AIS transmission intervals of different vessels. Considering the characteristics of waterway trajectory data, this application uses a linear interpolation method for data completion.

[0085] Track segmentation refers to separating ship tracks for different voyages. Due to the large time span of the research data, the same ship may make multiple round trips within the research area; therefore, it is necessary to separate AIS data for the same ship belonging to different voyages. The index at the track segmentation point is determined based on a threshold time interval between adjacent track points. Generally, if the time interval between adjacent track points exceeds 900 seconds, they are considered to belong to different voyages, and the track needs to be segmented. In the segmented AIS data, the first point is the ship's track start point, and the last point is the ship's track end point.

[0086] S202. Obtain historical ship trajectory data, divide the historical ship trajectory data into multiple traffic modes, and the historical ship trajectory data shall include at least the ship's sailing time, latitude and longitude, speed and heading.

[0087] In one embodiment, dividing historical ship trajectory data into multiple traffic modes specifically includes:

[0088] Determine the departure and arrival areas for ship navigation based on the geographical information of the waterway;

[0089] The starting and ending points of the historical ship trajectory data are matched with the departure area and the arrival area, and the historical ship trajectory data whose starting point belongs to the departure area and whose ending point belongs to the arrival area are divided into parent traffic modes.

[0090] The historical ship trajectory data with the same navigation route in the parent traffic mode are divided into the same sub-traffic mode, and each sub-traffic mode can be used as a traffic mode.

[0091] It should be noted that traffic patterns refer to the overall state of vessel traffic within a waterway. For example, a large number of vessels traveling along the same route from the same departure area to the same arrival area constitute a traffic pattern, and different traffic patterns form a complex maritime traffic network. Traffic pattern classification refers to dividing vessel trajectories into different trajectory clusters based on vessel AIS data and using data mining and other techniques to characterize typical maritime traffic patterns. There are significant differences in vessel traffic characteristics between different trajectory clusters, while the differences in vessel trajectory traffic characteristics within the same cluster are small.

[0092] In this embodiment, the parent traffic pattern refers to the set of all ship trajectories that share the same departure and arrival areas; the sub-traffic pattern is the set of ship trajectories that share the same departure and arrival areas but have different routes. The sub-traffic pattern is contained within the parent traffic pattern.

[0093] An example of a two-stage traffic pattern segmentation method based on origin-end point survey and trajectory clustering is as follows: Figure 3 As shown. By Figure 3 It can be seen that for the three entrance / exit water areas in the example, three judgments, A, B, and C, were set.

[0094] The regions, through the order and pairing of trajectories within the regions, yield six trajectory clusters: trajectory 1 from A to B, trajectory 2 from B to A, trajectory 3 from A to C, trajectory 4 from B to C, trajectories 5 and 7 from C to A, and trajectory 6 from C to B. Trajectories 5 and 7 from C to A have different routes and can be separated through trajectory clustering. After two divisions, seven sub-traffic patterns are ultimately obtained.

[0095] S203. Label the traffic patterns of each of the historical ship trajectory data, and construct a first dataset using the labeled historical ship trajectory data.

[0096] S204. Train the ship pattern recognition model based on the first dataset to obtain the trained ship pattern recognition model.

[0097] The acquired historical ship trajectory data can be labeled with different types, and a first dataset can be constructed using the historical ship trajectory data. The ship pattern recognition model can then be trained using the first dataset.

[0098] In this embodiment, features of historical ship trajectory data are first extracted. These features include at least the ship's sailing time, latitude and longitude, speed, and heading. They may also include statistical information such as bow direction, turning angle, acceleration, and their mean, equation, maximum, minimum, and median. Physical characteristics of the ship, such as its length and beam, are also included. The dataset is then divided into training and testing sets in an 8:2 ratio.

[0099] In this embodiment, the ship pattern recognition model adopts an ensemble learning decision tree traffic pattern prediction model, that is, multiple different models are trained using a training set, such as CART, RF, AdaBoost, GB, and XGBoost, and these multiple models are used as base learners. Then, a meta-learner is used to integrate the outputs of these base learners to form the final prediction result.

[0100] S205. Collect target vessel trajectory data, wherein the target vessel trajectory data includes at least the vessel's sailing time, latitude and longitude, speed, and heading.

[0101] It should be noted that the target vessel trajectory data is also trajectory data composed of several collected data points. Each data point contains the vessel's sailing time, latitude and longitude, speed, and heading data.

[0102] S206. Input the target ship trajectory data into the trained ship pattern recognition model to identify the traffic pattern of the target ship trajectory data; the same traffic pattern includes ship trajectory data with the same navigation route.

[0103] It should be noted that the traffic pattern recognition model in this embodiment can only identify the target vessel's navigation route, but cannot identify the target vessel's latitude, longitude, speed, and heading at every moment within a future time period. Therefore, this application also needs to train a vessel behavior prediction model to predict the target vessel's navigation process within a future time period. Figure 4 As shown, the traffic pattern recognition model is used to identify the ship's navigation route, and the ship behavior prediction model is used to predict the ship's navigation behavior over a period of time in the future. The navigation behavior includes the latitude, longitude, speed, and heading data at each moment.

[0104] S207. Calculate the distance between the target ship trajectory data and the historical ship trajectory data with the same traffic pattern, and select the n closest historical ship trajectory data from the historical ship trajectory data set as the first historical ship trajectory data.

[0105] The first historical ship trajectory data includes at least the sailing time, latitude and longitude, speed and heading.

[0106] S208. Construct a second dataset using the first historical ship trajectory data.

[0107] S209. Train the ship behavior prediction model based on the second dataset. The ship behavior prediction model is based on a Gaussian process regression model. The prediction process of the Gaussian process regression model reflects the trend of historical ship trajectories following the same traffic pattern.

[0108] It should be noted that, for the collected target vessel trajectory data, it is assumed that the target vessel trajectory data records the target vessel at {t1, t... n If we want to predict the target ship's position within a given time period (t), we can use the data within that time period. n+1 , t n+1+k For trajectory data within a given time period, multiple historical ship trajectory data points with the closest Euclidean distance to the target ship trajectory data can be selected from historical ship trajectory data with the same traffic patterns. Specifically, {t1, t...} can be calculated separately. n Within a given time period, the Euclidean distances from the target vessel data point to different historical vessel trajectory data points are used to select the historical vessel trajectory data points with the smallest distance values ​​as the first historical vessel trajectory data points; the Euclidean distances from each first historical vessel trajectory data point are then obtained. n+1 , t n+1+k The trajectory data within the specified time period is collected, and the flight time, latitude, longitude, speed, and heading are obtained for each data point. (The last part, "t", appears to be an incomplete sentence or fragment.) n+1 , t n+1+k Gaussian modeling is performed on the latitude, longitude, speed, and heading of data points within a given time period to train a Gaussian process regression model. In the Gaussian process regression model, the kernel function for the latitude and longitude parameters can be a linear kernel, a radial basis function kernel, or a white noise kernel; the kernel function for the speed parameter can be a linear kernel, a radial basis function kernel, or a periodic kernel; and the kernel function for the heading parameter can be a linear kernel, a radial basis function kernel, or a white noise kernel, as detailed below. Figure 5 As shown.

[0109] S210. Input the predicted sailing time of the first historical ship trajectory data into the trained ship behavior prediction model to predict the latitude, longitude, speed and heading of each of the first historical ship trajectory data at each moment during the predicted sailing time.

[0110] S211. The latitude, longitude, speed and heading values ​​at each moment of the predicted navigation time in each of the predicted first historical ship trajectory data are averaged to obtain the latitude, longitude, speed and heading of the target ship at each moment of the predicted navigation time.

[0111] It should be noted that for a trained Gaussian process regression model, the independent variable {t} of the first historical ship trajectory data can be used. n+1, t n+1+k The time period is used as input to a Gaussian process regression model to predict the trajectory of each historical ship in {t}. n+1 , t n+1+k The latitude, longitude, speed, and heading of the target vessel at each moment within a given time period are obtained; by averaging the latitude, longitude, speed, and heading of multiple historical vessel trajectory data, the target vessel's position within {t} can be determined. n+1 , t n+1+k The latitude, longitude, speed, and heading at each moment within the time period.

[0112] This invention achieves efficient segmentation of inland waterway vessel traffic patterns through origin-end point surveys and cluster analysis. A vessel pattern recognition model identifies the vessel's navigation route, and historical vessel trajectories along the same route are selected to train the vessel behavior prediction model. These historical trajectories contain the vessel's latitude, longitude, speed, and heading data at each moment, enabling the trained model to accurately identify the vessel's latitude, longitude, speed, and heading at specific times along the navigation route, thus accurately predicting the target vessel's behavioral characteristics. Furthermore, this application uses first-generation historical vessel trajectory data as input to the vessel behavior prediction model. The prediction process reflects the trend of the target vessel's trajectory following historical vessel trajectories along the same navigation route, thereby predicting the target vessel's latitude, longitude, speed, and heading over a future period using this historical trajectory data, resulting in highly interpretable prediction results.

[0113] Example 3

[0114] Figure 6 This is a schematic diagram of the structure of a ship behavior prediction device provided in Embodiment 3 of the present invention. Figure 3 As shown, the device includes:

[0115] The data acquisition unit 601 is used to acquire target vessel trajectory data, which includes at least the vessel's sailing time, latitude and longitude, speed, and heading.

[0116] The first training unit 602 is used to input the target ship trajectory data into the trained ship pattern recognition model to identify the traffic pattern of the target ship trajectory data; the same traffic pattern includes ship trajectory data with the same navigation route;

[0117] The data selection unit 603 is used to select a number of first historical ship trajectory data from a set of historical ship trajectory data that have the same traffic pattern as the target ship trajectory data; the first historical ship trajectory data includes at least the sailing time, latitude and longitude, speed and heading;

[0118] The second training unit 604 is used to train a ship behavior prediction model based on the first historical ship trajectory data to obtain the trained ship behavior prediction model.

[0119] The first prediction unit 605 is used to input the first historical ship trajectory data into the trained ship behavior prediction model to predict each of the first historical ship trajectory data.

[0120] The second prediction unit 606 is used to calculate the latitude, longitude, speed and heading of the target ship in the future period based on each predicted first historical ship trajectory data.

[0121] The ship behavior prediction device provided in the embodiments of the present invention can execute the ship behavior prediction method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0122] Example 4

[0123] Figure 7 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0124] like Figure 7 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0125] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0126] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as a ship behavior prediction method.

[0127] In some embodiments, a ship behavior prediction method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the ship behavior prediction method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform a ship behavior prediction method by any other suitable means (e.g., by means of firmware).

[0128] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0129] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0130] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0131] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0132] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0133] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0134] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0135] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for predicting ship behavior, characterized in that, include: Collect target vessel trajectory data, which includes at least the vessel's sailing time, latitude and longitude, speed, and heading; The target vessel trajectory data is input into the trained vessel pattern recognition model to identify the traffic pattern of the target vessel trajectory data; the same traffic pattern includes vessel trajectory data with the same navigation route; From a set of historical ship trajectory data with the same traffic pattern as the target ship trajectory data, select several first historical ship trajectory data; the first historical ship trajectory data includes at least sailing time, latitude and longitude, speed, and heading; specifically, calculate the distance between the target ship trajectory data and the historical ship trajectory data with the same traffic pattern, and select the n closest historical ship trajectory data from the set of historical ship trajectory data as the first historical ship trajectory data; A ship behavior prediction model is trained based on the first historical ship trajectory data to obtain the trained ship behavior prediction model; specifically: a second dataset is constructed using the first historical ship trajectory data; the ship behavior prediction model is trained based on the second dataset, and the ship behavior prediction model is a Gaussian process regression model. The prediction process of the Gaussian process regression model reflects the trend of historical ship trajectories following the same traffic pattern in the target ship trajectory. The first historical ship trajectory data is input into the trained ship behavior prediction model to predict each first historical ship trajectory data; specifically, the predicted sailing time of the first historical ship trajectory data is input into the trained ship behavior prediction model to predict the latitude, longitude, speed and heading of each first historical ship trajectory data at each moment within the predicted sailing time; the predicted sailing time is the time period that the target ship needs to predict. Based on the predicted trajectory data of each of the first historical ships, the latitude, longitude, speed, and heading of the target ship for a future period of time are calculated.

2. The ship behavior prediction method according to claim 1, characterized in that, Before inputting the target ship trajectory data into the trained ship pattern recognition model, the following steps are also included: Historical ship trajectory data is acquired, and the historical ship trajectory data is divided into multiple traffic modes. The historical ship trajectory data includes at least the ship's sailing time, latitude and longitude, speed and heading. The traffic patterns of each historical ship trajectory data point are labeled, and a first dataset is constructed using the labeled historical ship trajectory data. The ship pattern recognition model is trained based on the first dataset to obtain the trained ship pattern recognition model.

3. The ship behavior prediction method according to claim 2, characterized in that, Prior to acquiring historical ship trajectory data, the following steps are also included: Historical vessel automatic identification system (AIS) data is collected, and the historical vessel AIS data is preprocessed to obtain historical vessel trajectory data; the preprocessing includes at least data cleaning and trajectory extraction.

4. The ship behavior prediction method according to claim 3, characterized in that, The preprocessing of the historical vessel automatic identification system data to obtain the historical vessel trajectory data includes: Data cleaning is performed on the historical Automatic Identification System (AIS) data. Interpolation algorithms are used to repair missing data points in the historical Automatic Identification System (AIS) data. The historical vessel automatic identification system data, after data cleaning, is processed into trajectory segments to obtain multiple historical vessel trajectory data.

5. The ship behavior prediction method according to claim 2, characterized in that, The process of dividing the ship trajectory data into multiple traffic modes includes: Determine the departure and arrival areas for ship navigation based on the geographical information of the waterway; The starting and ending points of the historical ship trajectory data are matched with the departure area and the arrival area, and the historical ship trajectory data whose starting point belongs to the departure area and whose ending point belongs to the arrival area are divided into parent traffic modes. The historical ship trajectory data with the same navigation route in the parent traffic mode are divided into the same sub-traffic mode, and each sub-traffic mode can be used as a traffic mode.

6. The ship behavior prediction method according to claim 1, characterized in that, Based on the predicted trajectory data of each of the first historical vessels, the latitude, longitude, speed, and heading of the target vessel for a future period of time are calculated, including: By averaging the latitude, longitude, speed, and heading values ​​at each moment during the predicted navigation time for each of the first historical ship trajectory data, the latitude, longitude, speed, and heading of the target ship at each moment during the predicted navigation time can be obtained.

7. A ship behavior prediction device, characterized in that, include: The data acquisition unit is used to collect target vessel trajectory data, which includes at least the vessel's sailing time, latitude and longitude, speed, and heading. The first training unit is used to input the target ship trajectory data into the trained ship pattern recognition model to identify the traffic pattern of the target ship trajectory data; the same traffic pattern includes ship trajectory data with the same navigation route; The data selection unit is used to select a number of first historical ship trajectory data from a set of historical ship trajectory data that have the same traffic pattern as the target ship trajectory data; the first historical ship trajectory data includes at least sailing time, latitude and longitude, speed, and heading; specifically, it calculates the distance between the target ship trajectory data and the historical ship trajectory data with the same traffic pattern, and selects the n closest historical ship trajectory data from the set of historical ship trajectory data as the first historical ship trajectory data; The second training unit is used to train a ship behavior prediction model based on the first historical ship trajectory data to obtain the trained ship behavior prediction model; specifically, it constructs a second dataset using the first historical ship trajectory data; trains the ship behavior prediction model based on the second dataset, wherein the ship behavior prediction model is based on a Gaussian process regression model; the prediction process of the Gaussian process regression model reflects the trend of historical ship trajectories following the same traffic pattern. The first prediction unit is used to input the first historical ship trajectory data into the trained ship behavior prediction model to predict each of the first historical ship trajectory data; specifically, it inputs the predicted sailing time of the first historical ship trajectory data into the trained ship behavior prediction model to predict the latitude, longitude, speed, and heading of each of the first historical ship trajectory data at each moment within the predicted sailing time; the predicted sailing time is the time period that the target ship needs to predict. The second prediction unit is used to calculate the latitude, longitude, speed and heading of the target ship in the future period based on each predicted first historical ship trajectory data.

Images