Ship trajectory prediction method and device based on large model, equipment, medium and program product

Abstract

The present disclosure provides a large model-based ship trajectory prediction method, device, equipment, storage medium and program product. The method comprises: in response to a trajectory prediction instruction for a ship, obtaining a historical navigation trajectory of the ship, an initial constrained navigation range of the ship and meteorological data of a target navigation area; the meteorological data is imported into a grid-based geographic space index map to obtain a meteorological geographic space index map; based on the initial constrained navigation range of the ship, the historical navigation trajectory of the ship and the meteorological geographic space index map, the predicted trajectory of the ship within a future preset time length is determined; the predicted navigation position of the nth time point is determined by the following steps: inputting the nth constrained navigation range of the ship, the (n-1)th predicted navigation position of the ship and the meteorological geographic space index map into a pre-trained language model to obtain the nth candidate navigation position and the predicted probability corresponding to each nth candidate navigation position; based on the predicted probability, the nth predicted navigation position is determined from the nth candidate navigation position.

Description

Technical Field

[0001] This disclosure relates to the field of intelligent navigation technology, and more specifically to a method, apparatus, equipment, medium, and program product for predicting ship trajectories based on a large model. Background Technology

[0002] Ship trajectory prediction is the process of forecasting a ship's future trajectory based on its historical trajectory. Common methods involve using deep neural networks to predict ship movement, but these methods are insufficient for fitting complex patterns in ship time-series data and suffer from information loss when processing spatial information. Furthermore, the instability and diversity of ocean weather introduce significant noise into the trajectory data, further exacerbating the decline in prediction accuracy and making it difficult to accurately predict ship navigation trajectories. Summary of the Invention

[0003] In view of the above problems, this disclosure provides a method, apparatus, device, medium and program product for ship trajectory prediction based on a large model.

[0004] According to a first aspect of this disclosure, a ship trajectory prediction method based on a large model is provided, comprising: in response to a trajectory prediction command for a ship, acquiring the ship's historical navigation trajectory, the ship's initial constrained navigation range, and meteorological data of a target navigation area, wherein the target navigation area is an area that the ship plans to navigate within a preset future time period, the historical navigation trajectory is a navigation trajectory generated by the ship within a first preset time period before entering the target navigation area, and the initial constrained navigation range is a navigation range constrained by the ship's initial motion elements and physical performance when the ship receives the trajectory prediction command; importing the meteorological data into a gridded geospatial index map to obtain a meteorological geospatial index map, wherein the meteorological geospatial index map indicates the feasible and prohibited areas of the target navigation area; and determining the ship's trajectory within the target navigation area based on the ship's initial constrained navigation range, the ship's historical navigation trajectory, and the meteorological geospatial index map. The predicted trajectory within a preset time period includes the predicted navigation positions at N time points. The predicted navigation position at the nth time point is determined by the following steps: inputting the nth constrained navigation range of the ship, the (n-1)th predicted navigation position of the ship, and the meteorological geospatial index map into a pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position. The nth constrained navigation range is determined based on the (n-1)th predicted navigation position of the ship. When n is 1, the nth constrained navigation range is the initial constrained navigation range. Based on the prediction probability, the nth predicted navigation position is determined from the multiple nth candidate navigation positions.

[0005] According to embodiments of this disclosure, the initial motion elements include: initial position, initial course, and initial speed; the physical properties include: limiting acceleration, limiting angular velocity, and limiting angular velocity; and the initial constrained navigation range is determined by the following steps: acquiring the ship's initial speed, initial position, initial course, limiting acceleration, limiting angular velocity, and limiting angular velocity; determining the ship's straight-line travel distance within a second preset time period based on the initial speed and limiting acceleration; constructing a feasible course range for the ship based on the ship's initial course and limiting angular velocity; determining the ship's predicted navigation area in the target navigation area based on the initial position, the straight-line travel distance, and the limiting angular velocity; and constructing the constrained navigation range based on the feasible course range and the predicted navigation area.

[0006] According to embodiments of this disclosure, the aforementioned geographic index map includes grid identifiers, the aforementioned meteorological data includes multiple meteorological elements, each meteorological element includes a meteorological location and a meteorological value sequence, and the aforementioned importation of the meteorological data into a gridded geographic spatial index map to obtain a meteorological geographic spatial index map includes: for any meteorological element, determining the grid identifier of the aforementioned meteorological element based on the geographic location corresponding to the aforementioned grid identifier and the aforementioned meteorological location; selecting the maximum value in the aforementioned meteorological value sequence as a meteorological feature value and filling it into the grid corresponding to the aforementioned grid identifier to obtain the aforementioned meteorological geographic spatial index map.

[0007] According to embodiments of this disclosure, the above-mentioned inputting the nth constrained navigation range of the ship, the (n-1)th predicted navigation position of the ship, and the meteorological geospatial index map into a pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position includes: removing the prohibited areas from the meteorological geospatial index map from the nth constrained navigation range to obtain a feasible navigation range, and determining the feasible navigation meteorological data corresponding to the feasible navigation range in the meteorological geospatial index map; inputting the nth feasible navigation range, the (n-1)th predicted navigation position of the ship, and the feasible navigation meteorological data into the pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position.

[0008] According to embodiments of this disclosure, the method further includes: when the nth predicted navigation position does not meet preset conditions, adjusting the (n-1)th predicted navigation position based on the (n-2)th predicted navigation position, the (n-1)th constrained navigation range, and the meteorological geospatial index map to obtain the adjusted (n-1)th predicted navigation position and the adjusted nth constrained navigation range; and determining multiple adjusted nth predicted navigation positions based on the adjusted nth constrained navigation range, the adjusted (n-1)th predicted navigation position, and the meteorological geospatial index map.

[0009] According to embodiments of this disclosure, the aforementioned preset conditions include at least one of the following: the directional error between the nth heading and the (n-1)th heading is less than or equal to a preset error; the nth predicted navigation position is within the aforementioned feasible area; the predicted probability corresponding to the nth predicted navigation position is greater than or equal to a preset probability value; and the difference between the predicted probability corresponding to the nth predicted navigation position and the predicted probability corresponding to the (n-1)th predicted navigation position is greater than or equal to a preset difference. The nth heading is determined by the nth predicted navigation position and the (n-1)th predicted navigation position, and the (n-1)th heading is determined by the (n-1)th predicted navigation position and the (n-2)th predicted navigation position.

[0010] The second aspect of this disclosure provides a ship trajectory prediction device based on a large model, comprising: an acquisition module, configured to, in response to a trajectory prediction command for a ship, acquire the ship's historical navigation trajectory, the ship's initial constrained navigation range, and meteorological data of a target navigation area, wherein the target navigation area is the area the ship plans to navigate within a preset future time period, the historical navigation trajectory is the navigation trajectory generated by the ship within a first preset time period before entering the target navigation area, and the initial constrained navigation range is the navigation range constrained by the ship's initial motion elements and physical performance when the ship receives the trajectory prediction command; importing the meteorological data into a gridded geospatial index map to obtain a meteorological geospatial index map, wherein the meteorological geospatial index map indicates the feasible and prohibited areas of the target navigation area; and determining the ship's trajectory within the target navigation area based on the ship's initial constrained navigation range, the ship's historical navigation trajectory, and the meteorological geospatial index map. The predicted trajectory within a preset time period includes the predicted navigation positions at N time points. The predicted navigation position at the nth time point is determined by the following steps: inputting the nth constrained navigation range of the ship, the (n-1)th predicted navigation position of the ship, and the meteorological geospatial index map into a pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position. The nth constrained navigation range is determined based on the (n-1)th predicted navigation position of the ship. When n is 1, the nth constrained navigation range is the initial constrained navigation range. Based on the prediction probability, the nth predicted navigation position is determined from the multiple nth candidate navigation positions.

[0011] A third aspect of this disclosure provides an electronic device comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method described above.

[0012] A fourth aspect of this disclosure also provides a computer-readable storage medium having a computer program or instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.

[0013] The fifth aspect of this disclosure also provides a computer program product, including a computer program or instructions that, when executed by a processor, implement the steps of the above-described method.

[0014] According to embodiments of this disclosure, meteorological data is imported into a gridded geospatial index map to achieve a gridded representation of the meteorological data. Furthermore, constrained navigation ranges and historical ship trajectories are introduced to achieve a spatiotemporal coupled representation of ship trajectories and meteorological data, enabling the model to accurately handle the dynamic impact of complex marine environments on navigation trajectories. Simultaneously, by constructing dual navigation restrictions—restricted areas and physical performance constraints—through the meteorological geospatial index map and constrained navigation ranges, the pre-trained language model only needs to search within feasible areas at each step, reducing the search range and thus reducing computational memory usage. At the same time, the dual navigation restrictions reduce potential collisions and unsafe factors related to hazardous weather elements, improving the safety of predicted trajectories. Attached Figure Description

[0015] The foregoing contents, as well as other objects, features, and advantages of this disclosure, will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0016] Figure 1 The illustration shows an application scenario of the ship trajectory prediction method and apparatus based on a large model according to an embodiment of the present disclosure.

[0017] Figure 2 A flowchart illustrating a ship trajectory prediction method based on a large model according to an embodiment of the present disclosure is shown schematically.

[0018] Figure 3 A schematic diagram illustrating a navigation grid visualization according to an embodiment of the present disclosure is shown.

[0019] Figure 4 A schematic diagram illustrates the structure of a large-model-based ship trajectory prediction device according to an embodiment of the present disclosure; and

[0020] Figure 5 A block diagram schematically illustrates an electronic device suitable for implementing a large-model-based ship trajectory prediction method according to an embodiment of the present disclosure. Detailed Implementation

[0021] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0022] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0023] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0024] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0025] With the development of artificial intelligence technology, large language models are becoming increasingly important in time series processing tasks. Although they have shown potential in time series processing, in the field of ship trajectory prediction, there is a lack of effective strategies for integrating meteorological elements, spatiotemporal characteristics, and multimodal data. This fails to fully leverage the ability of large models to process complex multimodal data in the maritime field, making it difficult to meet the needs of accurate trajectory prediction.

[0026] This disclosure provides a ship trajectory prediction method based on a large model, comprising: in response to a trajectory prediction command for a ship, acquiring the ship's historical navigation trajectory, the ship's initial constrained navigation range, and meteorological data of the target navigation area, wherein the target navigation area is the area the ship plans to navigate within a preset future time period, the historical navigation trajectory is the navigation trajectory generated by the ship within a first preset time period before entering the target navigation area, and the initial constrained navigation range is the navigation range constrained by the ship's initial motion elements and physical performance when the ship receives the trajectory prediction command; importing the meteorological data into a gridded geospatial index map to obtain a meteorological geospatial index map, wherein the meteorological geospatial index map indicates the feasible and prohibited areas of the target navigation area; and determining the ship's trajectory within a preset future time period based on the ship's initial constrained navigation range, the ship's historical navigation trajectory, and the meteorological geospatial index map. The predicted trajectory includes the predicted navigation positions at N time points. The predicted navigation position at the nth time point is determined by the following steps: inputting the ship's nth constrained navigation range, the ship's (n-1)th predicted navigation position, and the meteorological geospatial index map into a pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position. The nth constrained navigation range is determined based on the ship's (n-1)th predicted navigation position. When n is 1, the nth constrained navigation range is the initial constrained navigation range. The nth predicted navigation position is determined from the multiple nth candidate navigation positions based on the prediction probability.

[0027] Figure 1 The illustration shows an application scenario of the ship trajectory prediction method and apparatus based on a large model according to an embodiment of the present disclosure.

[0028] like Figure 1 As shown, application scenario 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 serves as a medium for providing a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0029] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social media platform software, etc. (for example only).

[0030] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0031] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0032] It should be noted that the ship trajectory prediction method based on a large model provided in this disclosure can generally be executed by server 105. Correspondingly, the ship trajectory prediction device based on a large model provided in this disclosure can generally be located in server 105. The ship trajectory prediction method based on a large model provided in this disclosure can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Correspondingly, the ship trajectory prediction device based on a large model provided in this disclosure can also be located in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105.

[0033] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0034] The following will be based on Figure 1 The described scene, through Figures 2-3 The ship trajectory prediction method based on a large model, as disclosed in the embodiments, is described in detail.

[0035] Figure 2 A flowchart illustrating a ship trajectory prediction method based on a large model according to an embodiment of the present disclosure is shown schematically.

[0036] like Figure 2 As shown, the ship trajectory prediction method based on a large model in this embodiment includes operations S210 to S230.

[0037] In operation S210, in response to a trajectory prediction command for the ship, the ship's historical navigation trajectory, the ship's initial constrained navigation range, and meteorological data of the target navigation area are acquired.

[0038] The target navigation area is the area that the ship plans to navigate within a preset time period in the future. The historical navigation trajectory is the navigation trajectory generated by the ship within the first preset time period before entering the target navigation area. The initial constrained navigation range is the navigation range constrained by the initial motion elements and the ship's physical performance when the ship receives the trajectory prediction command.

[0039] According to embodiments of this disclosure, motion elements are a set of physical quantities describing the instantaneous motion state of a ship, and physical performance is a set of inherent motion capabilities of a ship due to its own structure, power, and maneuvering characteristics.

[0040] According to embodiments of this disclosure, historical navigation trajectories of ships can be collected through an Automatic Identification System (AIS) and converted into a unified format to establish a basic data structure. Elements in the historical navigation trajectory include timestamps, latitude and longitude, speed, heading, and ship type. The timestamps can be in Unix format for easier subsequent timestamp calculation. The historical navigation trajectories can be deduplicated based on the ship's unique identification number and timestamp to remove duplicate records. Furthermore, the data can be sorted in ascending order based on the ship's unique identification number and timestamp to ensure the temporal continuity of the trajectory.

[0041] According to embodiments of this disclosure, historical navigation trajectories collected by the Automatic Identification System (AIS) may have missing data. The historical navigation trajectory can be divided into independent segments at fixed time intervals (e.g., 20 minutes) to avoid invalid interpolation across long time periods. Furthermore, a target timestamp sequence can be generated for subsequent alignment with meteorological and environmental data. For missing parameters such as position and speed in the historical navigation trajectory, cubic spline interpolation can be used to achieve a smooth transition. Cubic spline interpolation ensures the continuity of the first and second derivatives of the interpolation curve, thus conforming to the smooth trajectory characteristics of ship navigation. When there are fewer than two valid data points in the parameters of the historical navigation trajectory, the target timestamp exceeds the valid data time range, or the cubic spline interpolation calculation is abnormal, it will automatically switch to linear interpolation (i.e., linear fitting based on the two most recent valid data points before and after the target timestamp). The interpolation result is rounded to a preset three decimal places to improve data accuracy. Finally, the normalized segment is divided into sub-trajectories at a fixed number of points (15 points / segment) to ensure consistent input length.

[0042] According to embodiments of this disclosure, outlier detection and filtering can also be performed on the ship's historical navigation trajectory, initial constrained navigation range, and meteorological data of the target navigation area to identify and eliminate noise data exceeding reasonable ranges, ensuring the authenticity of the trajectory. Normal values ​​for speed are 0-40 knots, heading is 0-360 degrees, bow is 1-360 degrees (allowing 511 as an invalid value according to industry standards), latitude is -90-90 degrees, longitude is -180-180 degrees, length is 0-400 meters, beam is 0-60 meters, and draft is 0-25 meters. For each of the aforementioned parameters, a Boolean mask for field values ​​exceeding the valid range can be generated based on the latitude and longitude range of the target sea area. The bow field requires additional exclusion of values ​​exceeding the range other than 511. Furthermore, corresponding abnormal records can be deleted based on the mask, and the number of abnormal values ​​for each field can be counted to generate an outlier filtering report, facilitating data quality traceability.

[0043] According to embodiments of this disclosure, the historical navigation trajectory of a ship can also be mapped to a standardized geographic grid, thereby achieving a quantitative expression of spatial features. This disclosure employs an H3 hexagonal grid index system (i.e., a geospatial index map) that supports multi-resolution spatial partitioning. Based on the latitude and longitude coordinates of the trajectory points, the H3 grid identifier at the corresponding resolution is calculated, achieving discretized encoding of geographical locations. Multi-resolution grid data is generated to meet the accuracy requirements of different prediction tasks, supporting multi-level spatial analysis.

[0044] In operation S220, meteorological data is imported into a gridded geospatial index map to obtain a meteorological geospatial index map.

[0045] The meteorological geospatial index map indicates the feasible and prohibited areas of the target navigation area.

[0046] According to embodiments of this disclosure, a feasible area is a set of grids in a meteorological geospatial index map where the meteorological risk is lower than or equal to a set threshold. Environmental parameters such as wind speed, wave speed, and visibility within this area are all within the limits allowed by the ship's physical performance, allowing for safe navigation. A prohibited area is a set of grids in the meteorological geospatial index map where the meteorological risk is higher than the set threshold and where there are objects in the geospatial index map that obstruct navigation.

[0047] According to embodiments of this disclosure, meteorological data may include 10m east-west wind speed components, 10m north-south wind speed components, 10m instantaneous maximum wind speed, effective wave height, wave direction, etc., and further, data preprocessing operations such as outlier detection and missing value imputation can be performed on the meteorological data. Specifically, the data source is a NetCDF (.nc) format file, and the multidimensional meteorological data of time-latitude-longitude-elements is converted into a two-dimensional structured data table, thereby ensuring that each record corresponds to a single time-single location-multiple elements relationship. Core fields are extracted, including time, latitude and longitude, and various meteorological elements. The time field and latitude and longitude field names are unified to be the same as those of the ship data, and invalid fields with all extremely small values ​​are deleted to ensure the consistency of fields in multi-source data.

[0048] According to the embodiments of this disclosure, when correcting meteorological data, for non-negative elements such as instantaneous maximum wind speed and wave height, the negative values ​​in the original values ​​can be corrected to 0 by using the physical threshold constraint of formula (1) to ensure that the corrected data conforms to the physical meaning.

[0049]

[0050] Where N is the corrected value and M is the original value.

[0051] According to the embodiments of this disclosure, for angle-type variables such as wind direction and wave direction, the value θ of the direction variable can be constrained to the range of 0-360° by the modulo operation of formula (2), and a normalized value θ' is generated to avoid ambiguity in the direction representation.

[0052]

[0053] Where θ is the value of the direction variable. This represents the value of the normalized direction variable.

[0054] According to embodiments of this disclosure, for extreme values ​​exceeding a reasonable range, a mean-filling method can be used for correction. This involves replacing outliers with the mean of the element within a global or spatiotemporal group, thereby reducing noise interference. For missing values ​​in meteorological data, a combination of hierarchical interpolation and targeted interpolation strategies can be employed, based on the spatiotemporal correlation of the meteorological data, to achieve accurate filling.

[0055] For example, meteorological data can be divided into spatiotemporal groups according to date, latitude range, and longitude range. For missing values, the mean of the same group is used first. If there is no valid data in the same group, the global mean of the same date, the global mean of the same latitude and longitude range, or the global mean is used in descending order. For single-point data with fixed latitude and longitude, a complete time series is constructed at 20-minute intervals. Time interpolation is used to impute missing values ​​in the continuous time dimension. The formula (3) below shows the formula for imputing missing values ​​in the continuous time dimension using time interpolation.

[0056]

[0057] Where V represents the imputed missing values, and t represents the time corresponding to the missing values. The most recent valid time before the missing value. The most recent valid time after the missing value. To and The corresponding element value, To and The corresponding element value.

[0058] According to the embodiments of this disclosure, when interpolating and filling directional variables (such as wind direction, wave direction, etc.), the value of the original directional variable can be converted into radians by the following formula (4).

[0059]

[0060] in, For the value of the direction variable, Radians represent the direction variable.

[0061] Then, the horizontal and vertical components are calculated using formulas (5) and (6).

[0062]

[0063]

[0064] in, For the horizontal component of the direction variable, For the vertical component of the direction variable, Radians represent the direction variable.

[0065] Furthermore, linear time interpolation is applied to the horizontal and vertical components respectively, and the following formula (7) shows the radian of the direction variable after linear time interpolation.

[0066]

[0067] in, The interpolated direction variable is in radians. The horizontal component of the interpolated direction variable. This is the vertical component of the interpolated direction variable.

[0068] Finally, the value of the interpolated direction variable is converted back to the value using formula (10).

[0069]

[0070] in, The value of the direction variable after interpolation. This is the radian value of the interpolated direction variable.

[0071] According to embodiments of this disclosure, by converting the direction variable into radians, then calculating the horizontal and vertical components, interpolating them, and then converting them back into the direction variable, interpolation errors caused by angle jumps can be avoided.

[0072] In operation S230, the ship's predicted trajectory within a preset time period is determined based on the ship's initial constrained navigation range, the ship's historical navigation trajectory, and the meteorological geospatial index map.

[0073] The predicted trajectory includes predicted navigation positions at N time points. The predicted navigation position at the nth time point is determined through the following steps: the nth constrained navigation range of the ship, the (n-1)th predicted navigation position of the ship, and the meteorological geospatial index map are input into a pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position. The nth constrained navigation range is determined based on the (n-1)th predicted navigation position of the ship. When n is 1, the nth constrained navigation range is the initial constrained navigation range. The nth predicted navigation position is determined from the multiple nth candidate navigation positions based on the prediction probability.

[0074] For example, the historical navigation track includes 10 points, which can be recorded as follows: - After receiving the trajectory prediction command for the ship, it can... - The system predicts the ship's first predicted navigation position within the target navigation area's predicted trajectory. After the first round of predictions is completed, this first predicted navigation position can be used as input for the second predicted navigation position in the next round. Furthermore, key constraints in each round can be preserved through structured descriptions (i.e., constrained navigation range, meteorological geospatial index map, and historical navigation trajectory) to avoid error accumulation.

[0075] According to embodiments of this disclosure, the time granularity of both the points in the historical navigation trajectory and the predicted navigation position is 20 minutes, that is, the time interval between the first predicted navigation position and the second predicted navigation position is 20 minutes, so as to ensure the continuity of the time sequence.

[0076] According to embodiments of this disclosure, meteorological data is imported into a gridded geospatial index map to achieve a gridded representation of the meteorological data. Furthermore, constrained navigation ranges and historical ship trajectories are introduced to achieve a spatiotemporal coupled representation of ship trajectories and meteorological data, enabling the model to accurately handle the dynamic impact of complex marine environments on navigation trajectories. Simultaneously, by constructing dual navigation restrictions—restricted areas and physical performance constraints—through the meteorological geospatial index map and constrained navigation ranges, the pre-trained language model only needs to search within feasible areas at each step, reducing the search scope and thus lowering computational memory requirements. Furthermore, the dual navigation restrictions reduce potential collisions and unsafe factors related to hazardous weather elements, improving the safety of predicted trajectories.

[0077] According to embodiments of this disclosure, the initial motion elements include: initial position, initial course, and initial speed; the physical properties include: limiting acceleration, limiting angular velocity, and limiting angular velocity; and the initial constrained navigation range is determined by the following steps: acquiring the ship's initial speed, initial position, initial course, limiting acceleration, limiting angular velocity, and limiting angular velocity, and determining the ship's straight-line travel distance within a second preset time period based on the initial speed and limiting acceleration; constructing the ship's feasible course range based on the ship's initial course and limiting angular velocity; determining the ship's predicted navigation area in the target navigation area based on the initial position, straight-line travel distance, and limiting angular velocity; and constructing the constrained navigation range based on the feasible course range and the predicted navigation area.

[0078] According to the embodiments of this disclosure, the initial heading is the heading of the ship at its initial position, and the initial speed is the real-time speed of the ship at its initial position.

[0079] According to the embodiments of this disclosure, the straight-line travel distance of a ship within a second preset time period can be calculated by the following formula (9).

[0080]

[0081] in, This is the straight-line travel distance. For the second preset duration, Initial speed, To limit acceleration.

[0082] According to embodiments of this disclosure, considering the limitations of a ship's turning capability, a fan-shaped activity area (i.e., a feasible course range) constrained by the maximum turning angle in the left and right directions can be established. The ship's limited turning angle can be expressed as... Then at the initial time The feasible heading range is ,in, Initial time The course, To limit the turning angle of ships.

[0083] According to embodiments of this disclosure, the coordinates of boundary points for the extreme cases of left / right turns in the furthest distance trajectory can be calculated using a circular trajectory model. The left turn boundary limits the steering angular velocity. Yaw to the left, trajectory is radius The arc (R is the straight-line sailing distance), end position Calculations were performed using circular interpolation in a spherical coordinate system. The position at the end of the right yaw was calculated symmetrically. This forms a shape with the current position as the vertex and the left and right boundary angles as... , radius is fan-shaped area This allows for the establishment of constrained navigation ranges.

[0084] According to an embodiment of this disclosure, the above-mentioned geographic index map includes grid identifiers, and the meteorological data includes multiple meteorological elements. Each meteorological element includes a meteorological location and a meteorological value sequence. Importing the meteorological data into the gridded geographic spatial index map to obtain the meteorological geographic spatial index map includes: for any meteorological element, determining the grid identifier of the meteorological element based on the geographic location and meteorological location corresponding to the grid identifier; selecting the maximum value in the meteorological value sequence as the meteorological feature value and filling it into the grid corresponding to the grid identifier to obtain the meteorological geographic spatial index map.

[0085] According to embodiments of this disclosure, gridded integration of meteorological environmental data can be achieved based on an H3 spatial indexing system. First, a multi-resolution H3 grid identifier is calculated for each latitude and longitude point according to a preset latitude and longitude range (i.e., the latitude and longitude range of the target navigation area). Meteorological element data at each time point are then bound to the corresponding grid identifier to generate structured data with multi-resolution H3 grid identifiers. Furthermore, independent gridded data files are generated according to different H3 resolutions, thereby achieving data aggregation at different spatial scales and providing standardized input for subsequent forecasting tasks.

[0086] For example, the meteorological data includes two meteorological elements: wind direction and wave height. The wind direction meteorological element includes multiple meteorological locations, and each meteorological location corresponds to a meteorological value sequence over a continuous period of time. For the meteorological element wind direction, the meteorological value sequence corresponding to meteorological location [1,1] is [5,6,7]. The meteorological location corresponds to the geographical location of grid identifier "1", so the grid identifier of the meteorological element is also "1". Further, the maximum value "7" in the meteorological value sequence is selected as the meteorological feature value and filled into the grid corresponding to grid identifier "1". The above operation is repeated continuously to obtain the meteorological geospatial index map.

[0087] According to the embodiments of this disclosure, the data can also be matched with historical flight paths. For each flight point in the historical flight path, the meteorological element closest to its geographical location is extracted to complete the precise match. When the precise match fails, the neighboring grid can be found based on the spatial topology relationship of the meteorological geospatial index map. All environmental grids with meteorological data are traversed, the distance to the center of the flight point grid is calculated and sorted in ascending order, and the top 3 environmental grids with the smallest distance are selected as candidate grids. The meteorological element record closest to the AIS record timestamp in the 3 candidate grids is extracted as the matching candidate data.

[0088] According to the embodiments of this disclosure, for angle-type meteorological elements such as wind direction and wave direction in meteorological data, each angle value (i.e., meteorological value) in the meteorological value sequence can be converted into vector components according to formula (10) to formula (12), and the average value of the vector components is taken and then converted back to the angle to ensure the continuity of direction.

[0089]

[0090]

[0091]

[0092] in, Let n be the average level vector component of the meteorological value series, and n be the number of meteorological values ​​in the series. This represents the horizontal vector component of the i-th meteorological value. The average vertical vector component of the meteorological value series. Let i be the horizontal vector component of the i-th meteorological value. This represents the average angle value of the meteorological value series.

[0093] According to embodiments of this disclosure, the above-mentioned inputting the nth constrained navigation range of the ship, the (n-1)th predicted navigation position of the ship, and the meteorological geospatial index map into a pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position includes: removing prohibited areas from the meteorological geospatial index map from the nth constrained navigation range to obtain a feasible navigation range, and determining the feasible navigation meteorological data corresponding to the feasible navigation range in the meteorological geospatial index map; inputting the nth feasible navigation range, the (n-1)th predicted navigation position of the ship, and the feasible navigation meteorological data into the pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position.

[0094] According to embodiments of this disclosure, the overlapping portion of the nth constrained navigation range and the restricted area (polygon) can be identified using specialized geometric calculation tools. When the nth constrained navigation range covers land or reef groups, these non-navigable overlapping areas can be accurately located. Through spatial clipping, the detected restricted areas are removed from the nth constrained navigation range, resulting in a feasible navigation range containing only navigable waters. Furthermore, each grid within the feasible navigation range is masked, with 0 representing a restricted grid and 1 representing a navigable grid, forming mask data that can be directly recognized by the computer model. Since the restricted area boundaries on electronic nautical charts are typically composed of polygonal line segments, direct clipping may result in jagged edges on the navigation area boundaries. Therefore, curve fitting techniques can be used to smooth the boundaries, making the edges of the navigation area more continuous and natural. If a grid includes both land and sea, it can be defined as a navigable area to adapt to varying mission requirements.

[0095] According to an embodiment of this disclosure, the method further includes: if the nth predicted navigation position does not meet the preset conditions, adjusting the (n-1)th predicted navigation position based on the (n-2)th predicted navigation position, the (n-1)th constrained navigation range, and the meteorological geospatial index map to obtain the adjusted (n-1)th predicted navigation position and the adjusted nth constrained navigation range; and determining multiple adjusted nth predicted navigation positions based on the adjusted nth constrained navigation range, the adjusted (n-1)th predicted navigation position, and the meteorological geospatial index map.

[0096] According to embodiments of this disclosure, the aforementioned preset conditions include at least one of the following: the directional error between the nth heading and the (n-1)th heading is less than or equal to a preset error; the nth predicted navigation position is within the feasible area; the predicted probability corresponding to the nth predicted navigation position is greater than or equal to a preset probability value; and the difference between the predicted probability corresponding to the nth predicted navigation position and the predicted probability corresponding to the (n-1)th predicted navigation position is greater than or equal to a preset difference. The nth heading is determined by the nth predicted navigation position and the (n-1)th predicted navigation position, and the (n-1)th heading is determined by the (n-1)th predicted navigation position and the (n-2)th predicted navigation position.

[0097] The predicted probability corresponding to the nth predicted navigation position can be calculated using the following formula (13).

[0098] (13)

[0099] in, The predicted probability for the nth predicted navigation position. Historical trend similarity (which can be determined by comparing the similarity of the heading and speed change patterns between the nth predicted navigation position and the 10 historical points). Meteorological condition fit (which can be determined by the meteorological condition fit of the nth predicted navigation position). The constraint satisfaction (can be determined by identifying the physical and geographical constraints of the nth predicted navigation position).

[0100] According to the embodiments of this disclosure, the directional errors of the nth and (n-1)th headings are less than or equal to a preset error, which means that the ship's heading has no sudden change. The preset probability value can be, for example, 0.7, and the preset difference can be, for example, 0.2. By ensuring that the difference between the predicted probability corresponding to the nth predicted navigation position and the predicted probability corresponding to the (n-1)th predicted navigation position is greater than or equal to the preset difference, it can be ensured that the prediction is indeed stable. Before each round of prediction, the pre-trained language model can read the structured description of the prediction points of the previous round in advance and verify the rationality verification results. If there are unqualified items, the parameters of the prediction points of the previous round are adjusted back, and the active sector of the current prediction point is calculated based on the corrected previous round points to avoid the error being passed on to the next round.

[0101] For example, the pre-trained language model can first be set as a senior ship trajectory prediction expert, proficient in ship dynamics, marine meteorology, and H3 grid coding. The ship trajectory needs to be predicted point-by-point according to the following rules: Input historical navigation trajectories (including k points), including structured data of ship attributes and meteorological data. The inference requirement is a sequential judgment based on ship attributes. First, define the ship's corresponding limiting acceleration (maximum acceleration) and limiting turning angle (default maximum turning angle). Second, combine the meteorological data of the ship's current navigation position to correct the constraint parameters to calculate the predicted navigation range (active sector) of the first predicted navigation position; that is, first calculate the initial range, then trim the restricted area. Third, within the trimmed active sector, combine the historical navigation trajectory and meteorological data to generate three candidate navigation positions and their corresponding prediction probabilities for the first predicted navigation position. Fourth, generate a structured description of the first predicted navigation position in a fixed format, containing basic information and constraint association information, for use in the next round of predicting the second predicted navigation position. Finally, output the latitude and longitude, H3 grid identifier, and corresponding structured description of the first predicted navigation position.

[0102] Figure 3 A schematic diagram illustrating a navigation grid visualization according to an embodiment of the present disclosure is shown.

[0103] like Figure 3As shown, the grid with ship patterns on a blue background represents the ship's historical navigation trajectory. For the current navigation position, the circle drawn by the dashed line represents the ship's initial constrained navigation range at the current position, the solid line represents the range of the nemesis course determined by the ship's limited turning angle, the grid with ship patterns on an orange background represents the nth predicted navigation position up to the nth time step, the grid with ship patterns on a red background represents the Nth predicted navigation position (i.e., the final position), the black grid represents the restricted area, the grid corresponding to the wave pattern represents the large wave area of ​​the target navigation area, and the grid corresponding to the wind pattern represents the large wind area of ​​the target navigation area.

[0104] According to embodiments of this disclosure, when storing historical navigation trajectories, predicted trajectories, and constrained navigation, a three-dimensional data container containing time, space, and constraint conditions can be constructed for storage. This container can include constrained navigation ranges for n future time points, thus forming a dynamic data chain of time series to meet the time accuracy requirements of short-term navigation prediction. The original area information can retain the original circular and sector geometric data without removing restricted areas for subsequent verification and backtracking. Restricted area masks, in the form of a two-dimensional array, clearly mark whether navigation is permitted at the corresponding time point for each grid. Key information related to navigation restrictions, these parameters determining the ship's maneuverability boundaries, the version number and update time of the electronic chart data ensure the use of the latest restricted area information, and accuracy indicators during spatial clipping ensure the reliability of data processing. The final spatiotemporal constraint data can be output as a three-dimensional tensor of time steps × number of grids × constraint features, thus deeply integrating with the previously processed spatiotemporal grid data. This structured data reflects both the physical motion limits of the ship and the geographical constraints, providing a scientifically sound and safe input for pre-trained language models.

[0105] Based on the aforementioned ship trajectory prediction method based on large models, this disclosure also provides a ship trajectory prediction device based on large models. The following will combine... Figure 4 The device is described in detail.

[0106] Figure 4 A schematic diagram of a large-model-based ship trajectory prediction device according to an embodiment of the present disclosure is shown.

[0107] like Figure 4 As shown, the ship trajectory prediction device 400 based on a large model in this embodiment includes an acquisition module 410, an import module 420, and a determination module 430.

[0108] The acquisition module 410 is used to acquire, in response to a trajectory prediction command for the ship, the ship's historical navigation trajectory, the ship's initial constrained navigation range, and meteorological data of the target navigation area. The target navigation area is the area the ship plans to navigate within a preset time period in the future. The historical navigation trajectory is the navigation trajectory generated by the ship within a first preset time period before entering the target navigation area. The initial constrained navigation range is the navigation range constrained by the initial motion elements and the ship's physical performance when the ship receives the trajectory prediction command. In one embodiment, the acquisition module 410 can be used to perform the operation S210 described above, which will not be repeated here.

[0109] Import module 420 is used to import meteorological data into a gridded geospatial index map to obtain a meteorological geospatial index map, which indicates the feasible and prohibited areas of the target navigation area. In one embodiment, import module 420 can be used to perform the operation S220 described above, which will not be repeated here.

[0110] The determination module 430 is used to determine the predicted trajectory of the ship within a preset time period based on the ship's initial constrained navigation range, the ship's historical navigation trajectory, and the meteorological geospatial index map. The predicted trajectory includes the predicted navigation positions at N time points. The predicted navigation position at the nth time point is determined by the following steps: inputting the ship's nth constrained navigation range, the ship's (n-1)th predicted navigation position, and the meteorological geospatial index map into a pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position. The nth constrained navigation range is determined based on the ship's (n-1)th predicted navigation position. When n is 1, the nth constrained navigation range is the initial constrained navigation range. The nth predicted navigation position is determined from the multiple nth candidate navigation positions based on the prediction probability. In one embodiment, the determination module 430 can be used to perform the operation S230 described above, which will not be repeated here.

[0111] According to embodiments of this disclosure, any plurality of modules among the acquisition module 410, import module 420, and determination module 430 may be combined into one module, or any one of these modules may be split into multiple modules. Alternatively, at least a portion of the functionality of one or more of these modules may be combined with at least a portion of the functionality of other modules and implemented in one module. According to embodiments of this disclosure, at least one of the acquisition module 410, import module 420, and determination module 430 may be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable means of integrating or packaging circuitry, or implemented in software, hardware, or firmware, or in any appropriate combination of any of these three implementation methods. Alternatively, at least one of the acquisition module 410, import module 420, and determination module 430 may be at least partially implemented as a computer program module, which, when run, can perform corresponding functions.

[0112] Figure 5 A block diagram schematically illustrates an electronic device suitable for implementing a large-model-based ship trajectory prediction method according to an embodiment of the present disclosure.

[0113] like Figure 5 As shown, an electronic device 500 according to an embodiment of the present disclosure includes a processor 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage portion 508 into a random access memory (RAM) 503. The processor 501 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 501 may also include onboard memory for caching purposes. The processor 501 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0114] RAM 503 stores various programs and data required for the operation of electronic device 500. Processor 501, ROM 502, and RAM 503 are interconnected via bus 504. Processor 501 performs various operations of the method flow according to embodiments of the present disclosure by executing programs in ROM 502 and / or RAM 503. It should be noted that the programs may also be stored in one or more memories other than ROM 502 and RAM 503. Processor 501 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in said one or more memories.

[0115] According to embodiments of this disclosure, the electronic device 500 may further include an input / output (I / O) interface 505, which is also connected to a bus 504. The electronic device 500 may also include one or more of the following components connected to the input / output (I / O) interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output (I / O) interface 505 as needed. A removable medium 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 510 as needed so that computer programs read from it can be installed into the storage section 508 as needed.

[0116] This disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs that, when executed, implement the method according to the embodiments of this disclosure.

[0117] According to embodiments of this disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, such as including, but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this disclosure, the computer-readable storage medium may include ROM 502 and / or RAM 503 and / or one or more memories other than ROM 502 and RAM 503 described above.

[0118] Embodiments of this disclosure also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code enables the computer system to implement the large-model-based ship trajectory prediction method provided in embodiments of this disclosure.

[0119] When the computer program is executed by the processor 501, it performs the functions defined in the system / apparatus of this disclosure embodiments. According to embodiments of this disclosure, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0120] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 509, and / or installed from a removable medium 511. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0121] In such an embodiment, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is executed by processor 501, it performs the functions defined in the system of this disclosure embodiment. According to embodiments of this disclosure, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0122] According to embodiments of this disclosure, program code for executing the computer programs provided in embodiments of this disclosure can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can execute entirely on a user's computing device, partially on a user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0123] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0124] Those skilled in the art will understand that the features described in the various embodiments of this disclosure can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this disclosure. In particular, the features described in the various embodiments of this disclosure can be combined and / or combined in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or combinations fall within the scope of this disclosure.

[0125] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.

Claims

1. A ship trajectory prediction method based on a large model, characterized in that, The method includes: In response to a trajectory prediction command for a ship, the system acquires the ship's historical navigation trajectory, the ship's initial constrained navigation range, and meteorological data for the target navigation area. The target navigation area is the area the ship plans to navigate within a preset time period in the future. The historical navigation trajectory is the navigation trajectory generated by the ship within a first preset time period before entering the target navigation area. The initial constrained navigation range is the navigation range constrained by the ship's initial motion elements and physical performance when the ship receives the trajectory prediction command. The meteorological data is imported into a gridded geospatial index map to obtain a meteorological geospatial index map, which indicates the feasible and prohibited areas of the target navigation area. Based on the ship's initial constrained navigation range, the ship's historical navigation trajectory, and the meteorological geospatial index map, the predicted trajectory of the ship within the preset future time period is determined. The predicted trajectory includes the predicted navigation positions at N time points, wherein the predicted navigation position at the nth time point is determined through the following steps: The nth constrained navigation range of the ship, the (n-1)th predicted navigation position of the ship, and the meteorological geospatial index map are input into a pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position. The nth constrained navigation range is determined based on the (n-1)th predicted navigation position of the ship. When n is 1, the nth constrained navigation range is the initial constrained navigation range. The nth predicted navigation position is determined from the plurality of nth candidate navigation positions based on the predicted probability.

2. The method according to claim 1, characterized in that, The initial motion elements include: initial position, initial course, and initial speed; the physical properties include: limited acceleration, limited turning angular velocity, and limited turning angle of the ship; and the initial constrained navigation range is determined through the following steps: The initial speed, initial position, initial course, limited acceleration, limited turning angular velocity, and limited turning angle of the ship are obtained, and the straight-line travel distance of the ship within a second preset time period is determined based on the initial speed and the limited acceleration. The feasible course range of the ship is constructed based on the ship's initial course and the ship's limited turning angle; Based on the initial position, the straight-line travel distance, and the limited turning angular velocity, the predicted navigation area of ​​the ship in the target navigation area is determined; The constrained navigation range is constructed based on the feasible course range and the predicted navigation area.

3. The method according to claim 1, characterized in that, The geographic index map includes grid identifiers, and the meteorological data includes multiple meteorological elements, each of which includes a meteorological location and a meteorological value sequence. The process of importing the meteorological data into the gridded geospatial index map to obtain a meteorological geospatial index map includes: For any meteorological element, the grid identifier of the meteorological element is determined based on the geographical location corresponding to the grid identifier and the meteorological location; The maximum value in the meteorological value sequence is selected as the meteorological feature value and filled into the grid corresponding to the grid identifier to obtain the meteorological geospatial index map.

4. The method according to claim 1, characterized in that, The process of inputting the nth constrained navigation range of the ship, the (n-1)th predicted navigation position of the ship, and the meteorological geospatial index map into a pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position includes: The prohibited areas in the meteorological geospatial index map are removed from the nth constrained navigation range to obtain the feasible navigation range, and the feasible navigation meteorological data corresponding to the feasible navigation range in the meteorological geospatial index map are determined. The nth feasible navigation range, the (n-1)th predicted navigation position of the ship, and the feasible navigation meteorological data are input into the pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position.

5. The method according to claim 4, characterized in that, The method further includes: If the nth predicted navigation position does not meet the preset conditions, the (n-2)th predicted navigation position is adjusted based on the (n-1)th constrained navigation range and the meteorological geospatial index map to obtain the adjusted (n-1)th predicted navigation position and the adjusted nth constrained navigation range. Multiple adjusted nth predicted navigation positions are determined based on the adjusted nth constrained navigation range, the adjusted (n-1)th predicted navigation position, and the meteorological geospatial index map.

6. The method according to claim 5, characterized in that, The preset conditions include at least one of the following: The directional error between the nth and (n-1)th headings is less than or equal to a preset error; the nth predicted navigation position is within the feasible area; the predicted probability corresponding to the nth predicted navigation position is greater than or equal to a preset probability value; and the difference between the predicted probability corresponding to the nth predicted navigation position and the predicted probability corresponding to the (n-1)th predicted navigation position is greater than or equal to a preset difference value. The nth heading is determined by the nth and (n-1)th predicted navigation positions, and the (n-1)th heading is determined by the (n-1)th and (n-2)th predicted navigation positions.

7. A ship trajectory prediction device based on a large model, characterized in that, The device includes: The acquisition module is used to acquire, in response to a trajectory prediction command for a ship, the ship's historical navigation trajectory, the ship's initial constrained navigation range, and meteorological data of the target navigation area. The target navigation area is the area that the ship plans to navigate within a preset time period in the future. The historical navigation trajectory is the navigation trajectory generated by the ship within a first preset time period before entering the target navigation area. The initial constrained navigation range is the navigation range constrained by the ship's initial motion elements and physical performance when the ship receives the trajectory prediction command. The import module is used to import the meteorological data into a gridded geospatial index map to obtain a meteorological geospatial index map, which indicates the feasible and prohibited areas of the target navigation area. The determination module is used to determine the predicted trajectory of the ship within a predetermined future time period based on the ship's initial constrained navigation range, the ship's historical navigation trajectory, and the meteorological geospatial index map. The predicted trajectory includes the predicted navigation positions at N time points, wherein the predicted navigation position at the nth time point is determined through the following steps: The nth constrained navigation range of the ship, the (n-1)th predicted navigation position of the ship, and the meteorological geospatial index map are input into a pre-trained language model to obtain multiple nth candidate navigation positions and the prediction probability corresponding to each nth candidate navigation position. The nth constrained navigation range is determined based on the (n-1)th predicted navigation position of the ship. When n is 1, the nth constrained navigation range is the initial constrained navigation range. The nth predicted navigation position is determined from the plurality of nth candidate navigation positions based on the predicted probability.

8. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 6.

Images