Ship communication link handover prediction method and system for LEO sparse constellation

Abstract

The present application relates to a kind of ship communication link switching prediction method and system for LEO sparse constellation, belong to information technology field, solve the problem of existing technology in LEO sparse constellation scene, switching threshold solidification, gap pre-judgment is missing.Method includes: obtaining constellation parameters and ship data;According to constellation parameters and ship data, constellation and ship coupling modeling is carried out, determines candidate satellite, according to constellation parameters and ship data, calculate the elevation switching threshold corresponding to target ship and current satellite, determine the link switching time corresponding to target ship;At link switching time, target ship carries out communication link switching, and establishes communication connection with candidate satellite. Realize the dynamic adaptation of ship communication link switching threshold and the accurate prediction of link switching time, link switching and resource scheduling collaborative simulation, accurately reproduce real communication performance.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method and system for predicting ship communication link switching for LEO sparse constellations. Background Technology

[0002] With the booming development of global maritime trade, ocean shipping, and offshore operations, the demand for high-speed and stable satellite communication services is becoming increasingly urgent for ships. Low Earth Orbit (LEO) satellite constellations, due to their advantages of low latency and high bandwidth, have become a core solution for maritime communication. However, in the early stages of LEO constellation deployment, limitations imposed by factors such as the number of satellites launched and orbital layout resulted in sparse coverage, leading to prominent problems for ship communication, including short satellite visibility times and frequent link switching.

[0003] Existing satellite communication simulation technologies for ships at sea are mainly designed for scenarios with complete constellation coverage or single satellite links, making it difficult to adapt to the special communication requirements of sparse LEO constellations, and presenting the following technical problems: 1) Fixed handover thresholds lead to insufficient scenario adaptability. Existing simulations use a fixed satellite elevation angle threshold (e.g., 5°) to trigger link handover, failing to consider dynamic factors such as ship speed, service priority, and constellation coverage density. For example, high-speed ships in sparsely covered areas are prone to delayed handovers due to excessively low handover thresholds, while low-priority services waste resources when resources are scarce due to excessively high handover thresholds. The simulation results deviate significantly from actual scenarios.

[0004] 2) Lack of coverage gap prediction and passive, delayed link switching. Existing simulation techniques do not consider the coupling relationship between the orbital characteristics of LEO sparse constellations and the dynamic navigation of ships. They trigger link switching based solely on single parameters such as the current satellite elevation angle, failing to predict constellation coverage gaps (i.e., periods when satellites are not visible). When ships are at the edge of coverage gaps, communication interruptions are prone to occur due to untimely switching. The link interruption rate in simulations is as high as 15%-20%, which is poorly adapted to actual sparse constellation scenarios.

[0005] Therefore, there is an urgent need for a method and system for predicting ship communication link switching for LEO sparse constellations. This method should be able to dynamically adapt to multiple scenarios, accurately predict coverage gaps, and achieve dynamic adaptation of ship communication link switching thresholds and prediction of link switching times. It can be applied to actual ship navigation to improve the timeliness of ship communication link switching control, and can also be applied to satellite communication simulation to accurately reproduce real communication performance and provide a reference for constellation optimization. Summary of the Invention

[0006] Based on the above analysis, the embodiments of the present invention aim to provide a method and system for predicting ship communication link handover for LEO sparse constellations, in order to solve the problems of fixed handover thresholds, lack of coverage gap prediction, and passive link handover in the existing systems.

[0007] On one hand, embodiments of the present invention provide a method for predicting ship communication link handover for LEO sparse constellations, characterized by comprising: Obtain the constellation parameters of the LEO constellation and the ship data of the target vessel; the LEO constellation includes multiple satellites; Based on the constellation parameters and the ship data, constellation-ship coupling modeling is performed to determine candidate satellites; Based on the constellation parameters and the ship data, the elevation angle switching threshold corresponding to the target ship and the current satellite is calculated, and the link switching time corresponding to the target ship is determined; wherein, the current satellite is the satellite to which the target ship has a communication link; At the link switching time, the target ship switches its communication link and establishes a communication connection with the candidate satellite.

[0008] Furthermore, the constellation parameters include constellation deployment progress, ephemeris data of each satellite, and orbital period parameters; constellation-ship coupling modeling is performed based on the constellation parameters and the ship data to determine candidate satellites, including: Based on the constellation deployment progress, the ephemeris data, and the orbital period parameters, the visible satellites and their ECEF coordinates at multiple time steps are determined according to a preset time step. Based on the ship data, the ship's navigation trajectory is predicted using a ship maneuvering model. Based on the ship data, the ship's navigation trajectory, and the ECEF coordinates of each of the satellites, at each of the time steps, the feasibility of link communication between the target ship and each of the visible satellites is determined; The candidate satellites are determined based on the ship data, the constellation parameters, the ship's navigation trajectory, and the link communication feasibility between the target ship and each visible satellite at each time step.

[0009] Further, the ship data includes ship antenna parameters; based on the ship data, the ship's navigation trajectory, and the ECEF coordinates of each of the satellites, at each of the time steps, the feasibility of link communication between the target ship and each of the visible satellites is determined, including: Based on the ship's navigation trajectory and the ECEF coordinates of each of the satellites, at each of the time steps, the distance between the target ship and each of the visible satellites, the elevation angle of the satellite, and the azimuth angle of the satellite are calculated. At each time step, the geometric feasibility of the link between the target vessel and each visible satellite is determined based on the distance between each of the satellites and the elevation angle of the satellites. Based on the elevation angle, azimuth angle, and antenna parameters of each satellite, determine the feasibility of link pointing between the target ship and each visible satellite; Based on the geometric feasibility and the directional feasibility of each link, the feasibility of the link communication between the target ship and each of the visible satellites is determined.

[0010] Furthermore, the constellation parameters also include the satellite link quality of each satellite; the candidate satellites are determined based on the ship data, the constellation parameters, the ship's navigation trajectory, and the link communication feasibility between the target ship and each visible satellite at each time step, including: The ship's navigation trajectory, each of the orbital period parameters, and the link communication feasibility between the target ship and each of the visible satellites corresponding to each time step are fitted to obtain the link feasibility probability curve between the target ship and each of the visible satellites; Based on the feasibility probability curves of each link, determine the current link coverage gap; Based on the current link coverage gap, determine the link switching candidate time period, and calculate the candidate visibility duration of each visible satellite during the link switching candidate time period; The candidate satellites are determined based on the candidate visibility duration and the satellite link quality corresponding to each of the visible satellites.

[0011] Further, the ship's navigation trajectory, each of the orbital period parameters, and the feasibility of the link communication between the target ship and each of the visible satellites at each time step are fitted to obtain the probability curve of the link feasibility between the target ship and each of the visible satellites, including: The ship's navigation trajectory, each of the orbital period parameters, and each of the link communication feasibility are standardized to construct a link feasibility time series dataset; Feature extraction is performed on the aforementioned link feasibility time-series dataset to obtain a feasibility time-series feature matrix; The feasibility time-series feature matrix is ​​fitted, and the feasibility probability corresponding to each timestamp is calculated. All the timestamps and their corresponding feasibility probabilities constitute the link feasibility probability curve.

[0012] Further, based on the current link coverage gap, candidate time periods for link switching are determined, and the candidate visibility duration of each visible satellite during the candidate time period for link switching is calculated, including: Based on the current link coverage gap and in conjunction with the preset warning gap, candidate time periods for link handover are determined; During the link switching candidate period, each of the visible satellites is screened to determine the recommended switching satellite; Based on the feasibility probability curves of each link, the candidate visibility duration of the recommended switching satellite is calculated.

[0013] Furthermore, the ship data also includes ship speed and ship service priority; based on the constellation parameters and the ship data, the elevation angle switching threshold between the target ship and the current satellite is calculated, and the link switching time of the target ship is determined, including: Based on the ship data and the constellation parameters, extract the target ephemeris data and target orbital period parameters of the current satellite, as well as the candidate satellite link quality corresponding to the candidate satellite; Based on the ship data, the target ephemeris data, and the target orbital period parameters, calculate the rate of change of the target ship's elevation angle relative to the current satellite; The elevation angle switching threshold is calculated based on the ship's speed, the ship's service priority, the quality of the candidate satellite link, and the satellite ship's elevation angle change rate. Based on the starship elevation angle change rate, the moment when the current starship elevation angle of the target ship and the current satellite reaches the elevation angle switching threshold is calculated and used as the link switching moment.

[0014] Furthermore, based on the constellation deployment progress, the ephemeris data, and the orbital period parameters, the visible satellites and their ECEF coordinates at multiple time steps are determined, including: Quantify the deployment progress of the aforementioned constellations; Based on the quantified constellation deployment progress, the ephemeris data, and the orbital period parameters, a sparse constellation topology model is constructed, and constellation topology data for multiple time steps are obtained according to the preset time step. The sparse constellation topology model includes a data access layer, a physical computing layer, a topology construction layer, and an output service layer. The constellation topology data includes the visible satellites and the ECEF coordinates of the satellites.

[0015] Furthermore, the ship data includes ship AIS data; based on the ship data, using a ship maneuvering model, the ship's navigation trajectory is predicted, including: Obtain the current state vector from the ship's AIS data; The current state vector is input into the ship maneuvering model, and multiple predicted state vectors for each time step are obtained according to the preset time step. The ship's navigation trajectory is generated based on the current state vector and each predicted state vector.

[0016] On the other hand, embodiments of the present invention provide a ship communication link handover prediction system for LEO sparse constellations, including: The data acquisition module is used to acquire the constellation parameters of the LEO constellation and the ship data of the target vessel; the LEO constellation includes multiple satellites; The constellation-ship coupling modeling module is used to perform constellation-ship coupling modeling based on the constellation parameters and the ship data, and to determine candidate satellites. The dynamic switching decision module is used to calculate the elevation angle switching threshold between the target ship and the current satellite based on the constellation parameters and the ship data, and to determine the link switching time corresponding to the target ship; wherein, the current satellite is the satellite with which the target ship has a communication link; The dynamic switching decision module is also used to enable the target ship to switch communication links and establish a communication connection with the candidate satellite at the link switching time.

[0017] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects: 1. Calculate the elevation angle switching threshold based on constellation parameters and ship data to ensure dynamic adaptation between the elevation angle switching threshold and constellation status, ship dynamic characteristics and business requirements, thereby improving multi-scenario adaptability; determine the link switching time based on the elevation angle switching threshold to accurately predict coverage gaps, proactively prepare for switching, improve communication continuity, and enhance the ability to guarantee critical services.

[0018] 2. Based on the constellation parameters, a sparse constellation topology model is constructed to determine the constellation topology data. At the same time, based on ship data, a ship maneuvering model is used to predict the ship's navigation trajectory and determine the feasibility of link communication between the target ship and each visible satellite. Then, candidate satellites are determined. Combined with the prediction of the ship's future trajectory, the current link coverage gap is determined, and then candidate satellites are determined. This achieves accurate prediction of the current link coverage gap, ensures the feasibility of link switching to candidate satellites, and guarantees the quality of communication links.

[0019] 3. Based on ship data and the target ephemeris data and target orbital period parameters of the current satellite, calculate the satellite-ship elevation angle change rate. Combined with the ship's speed on the trajectory, ship service priority, and candidate satellite link quality, calculate the elevation angle switching threshold. This enables dynamic determination of the elevation angle switching threshold for different scenarios, improving multi-scenario adaptability. Furthermore, the link switching time is determined based on the elevation angle change rate, accurately predicting the link switching execution time and providing a time benchmark for resource preparation and scheduling before the switch.

[0020] In this invention, the above-described technical solutions can be combined with each other to achieve more preferred combinations. Other features and advantages of this invention will be set forth in the following description, and some advantages may become apparent from the description or be learned by practicing the invention. The objects and other advantages of this invention can be realized and obtained from what is particularly pointed out in the description and drawings. Attached Figure Description

[0021] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. Figure 1 This is a flowchart illustrating a method for predicting ship communication link switching for LEO sparse constellations in an embodiment of the present invention. Figure 2 A schematic diagram of the main modules of a ship communication link switching prediction system for LEO sparse constellations. Detailed Implementation

[0022] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0023] A specific embodiment of the present invention discloses a method for predicting ship communication link handover for LEO sparse constellations, such as... Figure 1 As shown, it includes: Step S1: Constellation parameters of the LEO constellation and ship data of the target vessel; the LEO constellation includes multiple satellites.

[0024] Acquire the constellation parameters of the LEO constellation and the vessel data of the target ship. The LEO constellation comprises multiple satellites. It should be noted that the constellation parameters and vessel data can originate from real-world measured data, manually set data, or a combination of both. For example, constellation parameters can be determined by real-time data from existing satellites and new satellite deployment plans. A combination of real-time acquisition and manual setting, along with real-time acquisition of vessel data and communication resource constraints, better reflects real-world scenarios; therefore, no restrictions are imposed here.

[0025] In this embodiment, constellation parameters include numerical and structured document-type data, specifically including constellation deployment progress, ephemeris data of each satellite, orbital period parameters, and satellite link quality of each satellite. For example, constellation deployment progress includes the number of launched satellites, the total number of planned satellites, the number of satellites in orbit, and orbital plane distribution (including the number of satellites in orbit and the number of planned satellites per orbital plane). Ephemeris data includes the orbital altitude, inclination, right ascension of the ascending node, and orbital insertion status of each satellite. It should be noted that when using this method to provide decision-making reference for ship communication link switching in real-world scenarios, the number of launched satellites refers to the actual number of satellites launched; when using this method to simulate ship communication link switching in constellation planning scenarios, the number of launched satellites also includes satellites preset as launched in the simulation, and the corresponding ephemeris data uses simulated ephemeris data. In other embodiments, constellation parameters also include satellite payload parameters (e.g., number of beams, bandwidth limit, transmit power) and single-timeslot bandwidth.

[0026] Furthermore, methods for obtaining constellation deployment progress include officially released deployment schedules and launch lists, which can be used for long-term constellation planning, such as the Starlink launch batches and number of satellites in orbit published on SpaceX's official website; TLE data (Celestrak / Space-Track) from space databases, which can be used for real-time status determination, such as the number of TLEs (number of satellites in orbit) under the "Starlink" category in Celestrak; satellite health status and orbit insertion confirmation from telemetry data, which can be used for accurate simulation, such as the "orbit insertion completed" and "payload activated" statuses of satellite platform telemetry; and third-party statistical space tracking websites, which can be used for rapid verification, such as constellation deployment progress statistics from N2YO.com and the Satellite Database website.

[0027] In this embodiment, ship data includes spatiotemporal sequence data, numerical data, text data, and interval data. Specifically, it includes ship AIS data, as well as ship antenna parameters, service QoS requirements, and ship service types. For example, ship AIS data includes the target ship's longitude / latitude, draft, ship speed, and ship heading towards land; ship antenna parameters include gain, pointing accuracy, and ship link quality; service QoS requirements include latency thresholds, service transmission rate requirements and / or bandwidth requirements, packet loss rate limits, and service data volume; and ship service types include navigation, emergency communication, high-definition video, and IoT sensing. Further, ship data also includes ship maneuver constraints (maximum turning angle, minimum speed) and ship service priorities (Level 1 service: emergency / navigation; Level 2 service: high-definition video; Level 3 service: IoT data).

[0028] Step S2: Perform constellation-ship coupling modeling based on the constellation parameters and the ship data to determine candidate satellites.

[0029] Based on constellation parameters and ship data, a precise coupling model of LEO sparse constellation and ship dynamics is performed. A multi-factor fusion algorithm is used to determine candidate satellites, specifically including steps S21-S24.

[0030] Step S21: Based on the constellation deployment progress, the ephemeris data, and the orbital period parameters, determine the visible satellites and their ECEF coordinates at multiple time steps according to a preset time step.

[0031] In this embodiment, based on the ephemeris data in the constellation parameters, the SGP4 orbital dynamics model is used, combined with the constellation deployment progress, to construct a dynamically updated sparse constellation topology model, outputting the coverage area of ​​each satellite and determining the constellation topology data for multiple time steps. The constellation topology data for each time step includes multiple visible satellites corresponding to that time step and the ECEF coordinates of each visible satellite. Step S21 specifically includes steps S211-S212.

[0032] Step S211: Quantify the constellation deployment progress.

[0033] The deployment progress of the constellation is quantified by converting it into structured data recognizable by the sparse constellation topology model. For example, deployment completion rate. The range of values ​​is ,in Indicates the number of satellites available in orbit. This represents the total number of planned satellites, used to adjust the weighting coefficients of the sparse constellation topology model during coupled modeling; orbital status. Where 1 indicates orbital activation and 0 indicates no launch or malfunction, used to mark the first... Whether each satellite participates in the calculation of the sparse constellation topology model; orbital plane completion rate. ,in , , They represent the first The orbital plane completion rate, number of satellites in orbit, and number of planned satellites can be used to construct sparse constellation topology models and regional deployment density by orbital plane. This is used to evaluate the constellation's coverage capability over a target area. Understandably, the target area refers to a city or a latitude-longitude band that is to be evaluated or of interest. Based on the target area's location and satellite ephemeris data and orbital period parameters, the number of satellites covering the target area can be determined based on historical data or planned satellite coverage. Target vessels refer to vessels that pass through the target area and are of interest in subsequent simulation calculations. When multiple vessels exist in the target area, each vessel can be sequentially designated as a target vessel.

[0034] Step S212: Based on the quantized constellation deployment progress, the ephemeris data, and the orbital period parameters, construct a sparse constellation topology model, and obtain constellation topology data for multiple time steps according to the preset time step; wherein, the sparse constellation topology model includes a data access layer, a physical computing layer, a topology construction layer, and an output service layer; the constellation topology data includes the visible satellites and the ECEF coordinates of the satellites.

[0035] This embodiment constructs a sparse constellation topology model using a layered construction and dynamic update mechanism, including a data access layer, a physical computation layer, a topology construction layer, and an output service layer. The physical computation results are combined with the deployment progress to form a constellation topology that can be output in real time, including the satellite IDs and ECEF coordinates of each visible satellite. It is understood that each satellite operates according to a preset orbit. In this embodiment, the visibility of each satellite to the target area is determined based on the target area, and the sparse constellation topology model is constructed accordingly.

[0036] The data access layer unifies the format of the acquired ephemeris data (including TLE data) and the quantized constellation deployment progress (including telemetry data) before inputting them into the physical calculation layer. Further, the data access layer includes a TLE parsing module, which parses TLE data into structured orbital elements (JSON format); a deployment progress module, which obtains the quantized constellation deployment progress and outputs an orbital status table to determine whether each satellite participates in subsequent calculations by the physical calculation layer; and a data verification module, which filters invalid TLEs and marks faulty satellites.

[0037] The physical calculation layer includes the SGP4 orbital dynamics model, which calculates satellite attitude and coverage, and outputs the physical state of each satellite. Further, the physical calculation layer includes the SGP4 calculation module, which, for satellites in the orbital state table with an "orbital activation" status, outputs the ECEF coordinates and velocity of each satellite based on the corresponding structured orbital elements and the input time information; the attitude calculation module, which outputs quaternions / Euler angles based on the ECEF coordinates and velocity; the coverage calculation module, which outputs coverage area parameters based on the ECEF coordinates and velocity; and the time synchronization module, which unifies the time base (UTC) for all calculations. SGP4 (Simplified General Perturbations 4) is a standard simplified perturbation model for near-Earth orbit satellite position and velocity prediction. Its core function is to calculate the spatial state of a satellite at any given time, starting from the orbital elements (TLE). For near-Earth orbit satellites with altitudes below 1000 km, short-term (within several days) prediction accuracy can reach the meter to hundred-meter level, meeting the needs of constellation topology dynamic updates and coverage calculations.

[0038] The topology construction layer, combined with the constellation deployment progress, constructs the topological relationships of the sparse constellation and dynamically updates the connection / coverage topology of available satellites. Further, the topology construction layer includes a satellite filtering module, which filters available satellites in orbit based on their deployment status (e.g., an orbital status of 1 indicates an active, available satellite in orbit); a topology graph construction module, which constructs a topology graph using satellites as nodes and their visibility (no Earth obstruction) as edges; a sparsity handling module, which filters invalid connections (e.g., satellites are too far apart, invisible, or the elevation angle between the target area and the satellite is insufficient), retaining the core topology; and a dynamic update module, which updates the status of nodes and edges in the topology graph at preset time steps (e.g., 10 seconds).

[0039] The output service layer converts the topology map into a business-usable format, generating constellation topology data for subsequent calculations. For example, it converts the data to GeoJSON (map visualization) or Protobuf (real-time transmission) formats. Further, the output service layer includes a format conversion module, an interface service module, and / or a visualization module. The format conversion module converts the topology map into a business-usable format; the interface service module provides a REST API / websocket interface to push constellation topology data in real time; and the visualization module projects the satellite coverage obtained from the physical calculation layer onto a 3D Earth model.

[0040] The sparse constellation topology model can update the topology map according to a preset time step and output the constellation topology data corresponding to each time step, including the corresponding multiple visible satellites and the ECEF coordinates of each visible satellite. In this embodiment, the prediction period and preset time step are set according to the prediction range and prediction accuracy requirements, and the number of time steps is determined. For example, if the preset time step is 10 seconds and the prediction period is 10 time steps, then the constellation topology data for 10 time steps will be output. It can be understood that each satellite operates according to its own orbit, and for the target area or target ship, the number of visible satellites corresponding to each time step will change over time.

[0041] Step S22: Based on the ship data, use the ship maneuvering model to predict the ship's navigation trajectory.

[0042] Based on ship data, ship maneuvering models are used to perform dynamic ship modeling, predict the position, speed, and navigation changes of target ships in the future, and generate ship navigation trajectories.

[0043] The ship's AIS data from the ship data is input into the ship maneuvering model. The ship maneuvering model outputs a predicted state vector for each time step within a preset prediction period, according to a preset time step. The predicted state vectors for each time step are then concatenated with the existing state vectors corresponding to the ship's AIS data to generate the ship's navigation trajectory. In this embodiment, step S22 specifically includes steps S221-S223. In other embodiments, the ship maneuvering model can employ existing technologies, such as a 6-DOF model or a wind, wave, and current model.

[0044] Step S221: Obtain the current state vector based on the ship's AIS data.

[0045] The current state vector is obtained based on the ship's AIS data. In this embodiment, the ship's AIS data includes longitude / latitude (…). ), ground speed SOG ( ), COG (Center of Horizontal Travel) ), the unit is degrees ( ), bow towards HDG ( ), ROT () These parameters are used sequentially to determine the target vessel's position, velocity, direction of travel, maneuvering status, and angular velocity. Assuming the target vessel moves in a horizontal plane, vertical motions such as pitch and roll are ignored, and only the horizontal position is considered. Speed ,course Four state variables are used to determine the current state vector based on the ship's AIS data. In the formula Indicates to Perform matrix transpose. The horizontal position of the target vessel can be obtained by converting its latitude and longitude. Furthermore, outlier filtering (such as position jumps or speed exceeding limits) can be performed on the vessel's AIS data before determining the current state vector.

[0046] Step S222: Input the current state vector into the ship maneuvering model, and obtain the predicted state vectors for multiple time steps according to the preset time step.

[0047] Ship maneuvering models are used to describe the motion patterns of ships under control commands (rudder angle, throttle), and combine them with ship AIS data to achieve coherent modeling of the ship's motion state.

[0048] In this embodiment, the prediction period and preset time step set in step S21 are used to determine the prediction state vectors corresponding to multiple time steps. The ship maneuvering model iteratively calculates the prediction state vector corresponding to each time step within the preset prediction period.

[0049] Specifically, based on the current state vector ( (Input the ship's maneuvering model) and predict the predicted state vector for the next time step. In the formula, Indicates based on the first The first time step state vector obtained Predict the state vector at each time step. This represents the nonlinear function of the ship's maneuvering model. This indicates control parameters, including ship turning rate and ship throttle, where the first... Time step state vector Determined based on the ship's AIS data.

[0050] This embodiment employs a first-order maneuvering model function, suitable for short-term prediction, balancing accuracy and computational efficiency. For example, the update equations for the four state variables are respectively... , , , In the formula, , These represent the target ship's desired speed and desired turning rate, respectively. , , , These represent the speed time constant, speed disturbance noise, heading time constant, and heading disturbance noise, respectively, corresponding to different types of ships (e.g., cargo ships, passenger ships, fishing boats). , , , The settings can differ and can be adjusted based on experience.

[0051] After obtaining the predicted state vector for the next time step, use that predicted state vector as input to predict the predicted state vector for the time step after that. Iterate this prediction process until the predicted state vector for each time step within the preset prediction period is obtained. For example, the current time step is... (No. (Time step), preset prediction period Preset time step The ship maneuvering model iteratively calculates future time points ( The predicted state vector of ) It is understandable Understandably, it is also possible to iteratively predict the prediction covariance vector at each time step (…). ) and the confidence level of the predicted state.

[0052] Furthermore, the ship maneuvering model also outputs a prediction covariance vector, which characterizes the prediction accuracy of the predicted state vector. Prediction covariance vector In the formula, Indicates based on the first Time step covariance vector The obtained number Time step prediction covariance vector Indicates the first The state transition matrix corresponding to a time step represents the sum of transformation relationships between the state vectors of two time steps. For example, it can be determined based on the update equations of the four state variables. express The transpose of the matrix, The process noise covariance matrix can be set based on engineering experience. For example, diagonal elements are , indicating horizontal position The noise weights are determined by factors such as speed and heading. Furthermore, a confidence ellipse is constructed using the predicted covariance, and the predicted state confidence is obtained by mapping the Mahalanobis distance to the chi-square distribution.

[0053] Furthermore, to improve prediction accuracy, after obtaining the predicted state vector for the next time step based on the ship maneuvering model, the ship's AIS data for the next time step is acquired. Based on the ship's AIS data for the next time step, the predicted state vector for the next time step is corrected using Kalman filtering. The filtered predicted state vector is then used as the final predicted state vector for the next time step and participates in the iterative prediction of the time step after that. Kalman filtering (KF) is used for trajectory smoothing and noise suppression. By fusing the ship's AIS data and the prediction results of the ship maneuvering model, the original predicted state vector of the ship maneuvering model is optimized. Through prediction-correction, the final predicted state vector for each time step is obtained. The ship's navigation trajectory is determined according to step S223. Through multi-source data fusion and algorithm optimization, accurate characterization, smooth correction, and short-term prediction of the ship's navigation state are achieved, generating a continuous dynamic navigation trajectory.

[0054] In this embodiment, considering the nonlinear characteristics of ship motion, an extended Kalman filter (EKF) is used for nonlinear approximation. Specifically, the filtered predicted state vector... In the formula, Indicates based on the first The first time step state vector obtained The original predicted state vector at each time step. Indicates the first Time step Kalman gain, Indicates the first The observation vector corresponding to the time step ship AIS data. The observation matrix represents the transformation matrix between the observation vector and the state vector acquired by AIS. In other embodiments, if the meaning of each element of the observation vector is consistent with that of the state vector, the observation matrix may not be required. In the formula, Indicates the first Time step prediction covariance vector Represents the observation matrix matrix transpose, This represents the measurement noise covariance matrix, based on the accuracy settings of the AIS data acquisition device. For example... diagonal elements are Furthermore, the predicted covariance vector can be updated by filtering. The filtered predicted covariance vector... In the formula, Represents the identity matrix.

[0055] By using Kalman filtering, the output of the ship maneuvering model, the first The original predicted state vector at each time step is fused with the existing state vector corresponding to AIS to correct the prediction results, reduce model prediction errors and acquisition device errors, and obtain a more accurate predicted state vector.

[0056] Step S223: Generate the ship's navigation trajectory based on the current state vector and each predicted state vector.

[0057] Based on the current state vector and the predicted state vectors at each time step within the prediction period, a ship's navigation trajectory is generated. Furthermore, based on the ship's AIS data, referring to the method for obtaining the current state vector in step S221, the historical state vectors of multiple time steps prior to the current time step can be determined. Then, in chronological order, the historical state vectors, the current state vector, and the predicted state vectors are concatenated to generate the ship's navigation trajectory.

[0058] Furthermore, the ship's navigation trajectory can be output in a standardized format and pushed to a visualization platform or provided as an API for upper-level systems to call. For example,

[0059] Furthermore, the ship maneuvering model and Kalman filter parameters used in this embodiment are obtained by optimizing historical measured AIS data or real-time collected measured data. For example, they are determined by fitting measured AIS data. , ,like , Based on measured AIS data, model and filtering parameters are optimized to improve the accuracy of ship navigation trajectories. AIS (Automatic Identification System) data, which automatically broadcasts ship identity, position, and speed information via the ship's VHF radio, is characterized by its real-time nature and wide coverage. Measured AIS data is typically updated every 1-10 seconds (the higher the ship's speed and the more frequent its maneuvers, the higher the update frequency). The raw data contains measurement noise (such as positioning drift and electromagnetic interference), requiring preprocessing to filter outliers before being used for model and filtering parameter optimization. In this embodiment, AIS data includes longitude / latitude (… ), the unit is degrees ( ), accuracy range is (about ), ground speed SOG ( ), the unit is section ( ), accuracy range is COG (Current Heading) ), the unit is degrees ( ), accuracy range is The bow faces HDG ( ), the unit is degrees ( ), accuracy range is ROT (Rating of Turning Point) ), the unit is degrees / minute ( ), accuracy range is These are used in sequence to determine the ship's position, speed, direction of travel, maneuvering status, and angular velocity status.

[0060] Step S23: Based on the ship data, the ship's navigation trajectory, and the ECEF coordinates of each of the satellites, determine the feasibility of link communication between the target ship and each of the visible satellites at each time step.

[0061] Based on the satellite ECEF coordinates obtained at multiple time steps in step S21 and the ship's navigation trajectory obtained in step S22, constellation and ship coupling modeling is performed to realize the dynamic connection between the satellite constellation topology and the target ship. The relative geometric relationship between the target ship and the on-orbit satellites is calculated in real time. Combined with the ship's antenna performance indicators (i.e., ship antenna parameters), the feasibility of establishing a communication link between the target ship and each satellite (i.e., link communication feasibility) is determined, which is used to screen satellites that the target ship can establish an effective communication link with. In this embodiment, the link communication feasibility is judged based on the visible satellites in the constellation topology data of the target ship. Furthermore, the calculation cycle of link communication feasibility is synchronized with the ship's AIS data update frequency and / or the constellation topology data update frequency. The link communication feasibility calculation is performed according to the preset time step to ensure data real-time performance and synchronization. Step S23 specifically includes steps S231-S234.

[0062] Step S231: Based on the ship's navigation trajectory and the ECEF coordinates of each of the satellites, at each of the time steps, calculate the distance between the target ship and each of the visible satellites, the elevation angle of the satellite, and the azimuth angle of the satellite.

[0063] Based on the ship's trajectory and the ECEF coordinates of each visible satellite at multiple time steps, the distance between the target ship and each visible satellite, the ship's elevation angle, and the ship's azimuth angle are calculated at each time step. The historical state vector, current state vector, and predicted state vector of the ship's trajectory can be used to determine the target ship's longitude / latitude, heading towards the ground, and attitude angles at multiple time steps based on the target ship's state vector at each time step.

[0064] The starship distance is the straight-line distance between a satellite and a target ship in the ECEF coordinate system. In this embodiment, the longitude / latitude of the target ship is first converted into the ship's ECEF coordinates. Using the ECEF coordinate system as a reference, the starship distance between the target ship and each visible satellite is calculated. The coordinate system is transformed through vector operations to eliminate the influence of the Earth's curvature on angle calculations. The conversion between longitude / latitude and ship ECEF coordinates can be calculated using the WGS-84 ellipsoid model or other existing technologies; no limitation is made here. In this embodiment, the draft is assumed to be 0, and vertical height is ignored during the conversion. In other embodiments, the ship's AIS data also includes the draft, which is converted into the ship's ECEF coordinates based on the ship's longitude / latitude and draft. For example, the starship distance... In the formula, Indicates the ECEF coordinates of the satellite. Indicates the ship's ECEF coordinates.

[0065] The satellite's elevation angle is the upward angle between the satellite and the target ship's horizontal plane, ranging from... After converting the satellite ECEF coordinates to the target ship's local horizontal coordinate system (NED coordinate system: North-East-Ground), the ship's elevation angle is calculated. For example, firstly, based on the satellite ECEF coordinates and the ship's ECEF coordinates, the vector from the target ship to the satellite is calculated. and unitized to obtain Secondly, based on the target ship's longitude / latitude and heading, calculate the target ship's local coordinate system unit vector and the northward unit vector. Eastward unit vector Earth-directed unit vector In the formula, The longitude / latitude of the target ship. The target ship's heading is determined; finally, the satellite's elevation angle is calculated based on the target ship's unit vector to the satellite and the target ship's ground-oriented unit vector. In the formula, express and The vector dot product, the result range is The negative sign indicates that the direction of elevation is converted from ground to sky, ensuring a consistent definition of elevation angle.

[0066] The azimuth angle of the satellite is the horizontal angle between the satellite and the target ship's due north direction, and its range is: The starship's azimuth is linked to the ship's course; when the ship's course changes, the starship's azimuth is updated accordingly. For example, first, the original value of the starship's azimuth is calculated. By taking the vector dot product of the target ship's unit vector to the satellite and the target ship's northward and eastward unit vectors, and using... The function obtains the original starship azimuth angle value, and then converts the original starship azimuth angle value into a corrected starship azimuth angle value. Maintain the starship's azimuth at Within the range.

[0067] Step S232: At each time step, determine the geometric feasibility of the link between the target ship and each visible satellite based on the distance between each satellite and the elevation angle of the satellite.

[0068] According to step S231, at each time step, the distance between the target vessel and each visible satellite and the satellite's elevation angle are calculated. Based on the link geometric feasibility conditions, the link geometric feasibility between the target vessel and each visible satellite is determined. In this embodiment, the link geometric feasibility conditions include elevation angle constraints, unobstructedness constraints, and distance constraints. If all three are satisfied, the link geometric feasibility between the target vessel and the satellite is considered feasible.

[0069] For example, the elevation angle constraint is that the starship's elevation angle is not less than an elevation angle threshold, where the elevation angle threshold is set according to the actual scenario, for example... The unobstructed distance constraint requires that there is no Earth obstruction between the satellites. For example, the tangent length between the satellite and the Earth's surface is determined using the satellite's ECEF coordinates. If the distance between the satellites is less than the tangent length, then there is no Earth obstruction between the target ship and the satellite. The distance constraint requires that the distance between the satellites does not exceed the maximum effective distance of the satellite communication link. The maximum effective distance of the satellite communication link is determined by satellite payload parameters (such as satellite payload power) and ship antenna parameters (such as antenna gain), typically within... The maximum effective distance of the starship communication link can be set according to the target ship and different satellites.

[0070] Step S233: Determine the feasibility of link pointing between the target ship and each of the visible satellites based on the elevation angle, azimuth angle, and antenna parameters of each satellite.

[0071] Based on the ship antenna parameters (such as pointing accuracy) in the target ship data and the star ship elevation angle and star ship azimuth angle corresponding to each visible satellite calculated in step S231 at each time step, it is determined whether the ship antenna can stably point to the satellite. In this embodiment, the pointing deviation angle of the ship antenna is calculated and compared with the pointing accuracy to determine the feasibility of link pointing, so as to ensure that the center of the ship antenna beam covers the satellite and that random pointing errors will not cause the satellite to deviate from the beam range.

[0072] First, calculate the theoretical pointing angle of the ship's antenna. In this embodiment, the theoretical pointing angle of the ship's antenna is consistent with the azimuth angle of the satellite. Second, calculate the elevation angle of the ship's antenna. In this embodiment, the elevation angle of the ship's antenna is consistent with the elevation angle of the satellite. Finally, calculate the pointing deviation angle of the ship's antenna. The calculation formula is as follows: In the formula Indicates the starship's azimuth. This indicates the starship's elevation angle. Due to the pointing error inherent in the ship's antenna, this is a random error and follows a normal distribution. ,in This indicates pointing accuracy; therefore, the pointing deviation angle of the ship's antenna must meet the following requirements. In the formula, HPBW represents the half-power beamwidth of the ship's antenna (typical value: ), This represents the error confidence interval (covering 99.7% of the random error).

[0073] Furthermore, the actual pointing angle of the antenna (including the theoretical pointing angle of the ship's antenna and the elevation angle of the ship's antenna) must be within the mechanical limits of the ship's antenna equipment (such as the azimuth angle of the starship). Angle of elevation If the calculated actual antenna pointing angle exceeds the mechanical limit range of the equipment, it indicates that there is a deviation in the calculation process, the calculation result is not feasible, and the ship data, starship parameters and calculation process need to be checked and recalculated.

[0074] Step S234: Determine the feasibility of the link communication between the target ship and each of the visible satellites based on the geometric feasibility and the link pointing feasibility of each link.

[0075] At each time step, if both the geometric feasibility and the pointing feasibility of the link between the target ship and the visible satellite are feasible, the link communication feasibility between the target ship and the visible satellite is determined to be feasible. If the geometric feasibility and / or pointing feasibility are infeasible, the link communication feasibility between the target ship and the visible satellite is determined to be infeasible. Link geometric feasibility and pointing feasibility can be assessed simultaneously, and the link communication feasibility can be determined last. Alternatively, they can be assessed sequentially. For example, first assess the geometric feasibility, then select visible satellites with feasible geometric feasibility, further assess the pointing feasibility of the selected visible satellites, and finally select those with feasible pointing feasibility, thus determining the link communication feasibility between the target ship and the newly selected visible satellites as feasible. It should be noted that the position (longitude / latitude) of the target ship at a future time can be determined based on the predicted state vector. Based on the constellation topology data at each time step (including the visible satellites at each time step and their corresponding ECEF coordinates), the link communication feasibility between the target ship and the visible satellites at that time step can be determined.

[0076] Furthermore, for cases where link communication is feasible, the link parameters are output, such as the distance between the target ship and each visible satellite, the elevation angle of the satellite, and the azimuth angle of the satellite.

[0077] Step S24: Determine the candidate satellites based on the ship data, the constellation parameters, the ship's navigation trajectory, and the link communication feasibility between the target ship and each visible satellite at each time step.

[0078] Based on ship data, constellation parameters, ship navigation trajectory, and the feasibility of link communication between the target ship and each visible satellite at each time step, a time series analysis model is used to fit the curve of link communication feasibility changing over time. The current link coverage gap (including gap start time and duration) is calculated, and candidate satellites are determined based on the candidate visibility duration and satellite link quality of each visible satellite, so that link switching can be performed at the start time of the current link coverage gap. Step S24 specifically includes steps S241-S244.

[0079] Step S241: Fit the ship's navigation trajectory, each of the orbital period parameters, and the link communication feasibility between the target ship and each of the visible satellites corresponding to each time step to obtain the link feasibility probability curve between the target ship and each of the visible satellites.

[0080] Based on the ship's navigation trajectory, orbital period parameters, and the feasibility of link communication between the target ship and each visible satellite at each time step, a curve showing the change in link communication feasibility over time is fitted to obtain the probability curve of link feasibility between the target ship and each visible satellite. Specifically, this includes steps S2411-S2413.

[0081] Step S2411: Standardize the ship's navigation trajectory, each of the orbital period parameters, and each of the link communication feasibility parameters to construct a link feasibility time series dataset.

[0082] First, the feasibility of link communication between the target ship and each visible satellite at each time step (discrete data) and the predicted future position data of the ship in its navigation trajectory (i.e., the predicted state vector corresponding to each time step, time series) are unified onto the UTC time axis, with the time step length consistent with the calculation period of link communication feasibility. In other embodiments, the feasibility of link communication between the target ship and each visible satellite at historical time steps and the longitude / latitude of the target ship can also be selected and unified onto the UTC time axis to construct a time series dataset of link feasibility.

[0083] Secondly, based on The principle is to remove abnormal data (such as sudden occlusion misjudgment or distance jumps) and retain valid data.

[0084] Finally, data is imputed for short-term missing data. In this embodiment, linear interpolation is used for data imputation to ensure the continuity of data in the link feasibility time-series dataset. Furthermore, if the proportion of missing data to total data exceeds the missing rate threshold (e.g., 5%), data re-collection is triggered (e.g., re-acquiring ship data and constellation parameters) to reconstruct the link feasibility time-series dataset.

[0085] In this embodiment, the link feasibility time series dataset includes time steps ( ), and the target ship's longitude / latitude corresponding to the time step ( ), feasibility of link communication () Starship elevation angle ( Starship azimuth ( In this embodiment, a structured JSON / CSV format is used.

[0086] By performing time alignment, outlier filtering, and data completion on the relevant data, a standardized time series dataset of link feasibility is formed, providing high-quality input for subsequent fitting.

[0087] Step S2412: Extract features from the link feasibility time series dataset to obtain a feasibility time series feature matrix.

[0088] The feasibility of link communication between the target ship and the satellite varies due to factors such as the satellite orbit period and ship navigation. Based on the link feasibility time series dataset, time series features, correlation features and derived features are extracted to form a feasibility time series feature matrix. This transforms unstructured dynamic information into quantitative indicators that can be used for time series fitting, providing support for subsequent time series fitting.

[0089] Specifically, the feasibility time-series feature matrix includes the link communication feasibility at each time point. Starship elevation angle Arranged according to time step into one-dimensional time series arrays, forming time series features. The satellite orbit period... Ship speed Rate of change of heading (Reflecting the disturbance of ship position to the feasibility of link communication), arranged into one-dimensional time series arrays according to time step, forming correlation features. It can be understood that this includes the orbital period of each satellite. Stable, each time step corresponds to the same ; Starship elevation angle variation rate Starship Distance Variation Rate (To assist in verifying occlusion risk), the features are arranged into a one-dimensional time series array according to the time step, forming derived features. The time series features, correlation features, and derived features are integrated into a time series feature matrix, where the row index is the timestamp and the column vector is... At the same time, for continuous features (such as Normalization processing (mapped to) (Interval), to eliminate the influence of dimensional differences on fitting accuracy.

[0090] Furthermore, appropriate weights can be set based on the strength of the array's impact on the feasibility of link communication. For example, the starship's elevation angle is a key dynamic information and is strongly correlated with the feasibility of link communication; its weight can be set to... ,in , Indicates the number of array elements.

[0091] Step S2413: Fit the feasibility time-series feature matrix and calculate the feasibility probability corresponding to each timestamp. All the timestamps and their corresponding feasibility probabilities constitute the link feasibility probability curve.

[0092] The feasibility time-series feature matrix is ​​fitted, and the feasibility probability corresponding to each timestamp is calculated. All timestamps and their corresponding feasibility probabilities constitute the link feasibility probability curve. In this embodiment, a modified Gaussian process regression (GPR) is used as the fitting basis function to balance fitting accuracy and computational efficiency. It should be noted that a timestamp is a smaller time granularity based on the time step. For example, if the time step is 10 seconds and the timestamp interval is 1 second, with 0:00:00 being the first time step and the first timestamp, then the next time step is 0:00:10, and the next timestamp is 0:00:01. In other embodiments, the timestamp interval can be 0.1 seconds, which can be set based on accuracy requirements.

[0093] Specifically, Gaussian process regression (GPR) is selected as the core fitting model. The kernel function uses a combination of a periodic kernel and a radial basis function kernel. The periodic kernel captures the periodicity of the satellite orbit, while the radial basis function fits the nonlinear disturbances caused by the target ship's motion. For example, the kernel function expression is as follows: ,in, , In the formula, Indicates a periodic kernel. Indicates radial base core, , The feasibility time series feature matrix represents the first... The, the One characteristic, , For periodic kernel parameters, , The radial basis kernel parameters are calibrated using the maximum likelihood estimation method.

[0094] The feasibility time-series feature matrix is ​​input into the GPR model to obtain the feasibility probability of the target ship and each visible satellite at each time stamp. ( Based on all timestamps and their corresponding feasibility probabilities, a link feasibility probability curve is constructed. Further, historical data is used to verify the fitting accuracy and optimize model parameters to improve prediction accuracy. In this embodiment, the fitting error characterizes the fitting accuracy. For example, if the fitting error (the deviation rate between actual feasibility and predicted feasibility) exceeds 5%, a sliding window iterative optimization of the kernel function parameters is used until the accuracy requirements are met. In other embodiments, the fitting error interval can also be derived based on the covariance matrix of the GPR model to reflect the fitting reliability.

[0095] Based on the feasibility time-series feature matrix, the mapping relationship between link feasibility and time is fitted, and a continuous link feasibility probability curve is output, realizing the transformation from discrete data to continuous curve.

[0096] Furthermore, to improve the adaptability of the link feasibility probability curve to the repeatability of satellite visibility patterns, before fitting the feasibility time-series feature matrix, based on the orbital period parameter, the feasibility time-series feature matrix is ​​divided into multiple feasibility time-series feature sub-matrices. Each feasibility time-series feature sub-matrice is fitted using the same model. After determining the corresponding single-cycle model parameters based on the fitting of each sub-matrix, the final model parameters are determined according to the mean. In other embodiments, the single-cycle model parameters with the smallest fitting error can also be selected as the final model parameters; this is not a limitation here. In this embodiment, the time length of one cycle of each satellite's operation (satellite cycle duration) is determined based on the orbital period parameter and used as the basis for dividing the feasibility time-series feature matrix. The feasibility time-series feature matrix is ​​divided into sub-matrices according to the satellite cycle duration. For example, the duration of 5 satellite cycles is used as the sub-matrix step size (i.e., the fitting time window), and the timestamp range of the feasibility time-series feature matrix ( The satellite cycle duration is The feasibility time series feature matrix is ​​divided into Each feasible temporal feature submatrix corresponds to a timestamp range of ( ). ), ( )....( ),in This indicates the time step. Furthermore, for scenarios involving high-speed ship maneuvers, when the ship's rate of change of course... During high-speed turning, a dynamic window fitting strategy is adopted. For example, the fitting time window is adjusted from 5 orbital cycles to 2 cycles to improve the adaptability to nonlinear disturbances and avoid deviations that are easy to occur in time-series fitting.

[0097] It should be noted that for multiple visible satellites, the sub-matrix can be divided according to the average satellite period of the multiple satellites, or a feasibility time-series feature matrix can be constructed for each visible satellite, and sub-matrix division can be performed separately. For example, for each visible satellite, based on the ship's navigation trajectory, the feasibility of link communication between the target ship and the visible satellite, and the orbital period parameter of the visible satellite, a corresponding link feasibility time-series dataset is constructed according to step S2411; for each visible satellite's link feasibility time-series dataset, the corresponding feasibility time-series feature matrix is ​​obtained according to step S2412; for each visible satellite's feasibility time-series feature matrix, based on the visible satellite's orbital period parameter, it is divided into multiple feasibility time-series feature sub-matrices; for each visible satellite's multiple feasibility time-series feature sub-matrices, the link feasibility probability curve for each visible satellite is obtained by fitting according to step S2413.

[0098] By employing a segmented fitting and correction strategy, the periodicity of satellite visibility and the nonlinearity of target ship motion are combined to form a high-precision curve showing the change in the feasibility of the target ship-satellite link over time.

[0099] Step S242: Determine the current link coverage gap based on the link feasibility probability curves of each link.

[0100] Based on the link feasibility probability curve, the single-satellite link coverage gap of each visible satellite is accurately calculated, and the single-satellite link coverage gap between the target ship and the current satellite is the current link coverage gap.

[0101] In this embodiment, firstly, by setting a visibility determination threshold (e.g., 0.5), the feasibility probability is converted into a binary feasibility probability. For example, the binary feasibility probability... The link feasibility probability curve is converted into a binary link feasibility change curve, which facilitates intuitive judgment of link feasibility.

[0102] Secondly, the binarized link feasibility change curves of each visible satellite are traversed. When there are consecutive timestamps with a binarized feasibility probability of 0, it is identified as a coverage gap, and the timestamp sequence corresponding to the gap is recorded. Complete the identification of coverage gaps.

[0103] Then, the gap parameters for determining the single-satellite link coverage gap of each visible satellite are calculated, including the start time. The first timestamp that satisfies the binarization feasibility probability of 0 and whose subsequent two consecutive time steps are also 0 is determined as the start time, avoiding instantaneous misjudgment and filtering out invisible disturbances within a single time step; the end time... The end time is determined by the timestamps where the binarization feasibility probability is 0 for the first two consecutive time steps and 1 for the subsequent timestamps; duration , Indicates the timestamp interval.

[0104] Furthermore, to improve prediction accuracy, the location information of the target ship at a certain timestamp can be predicted by combining the ship's navigation trajectory. Based on the current link coverage gap, the feasibility of link communication between the target ship and the current satellite is verified time-by-time, eliminating false gaps and updating the corresponding gap parameters. Specifically, the future position of the target ship is determined by combining the ship's navigation trajectory, and it is verified whether the ship's position exceeds the coverage range of all satellites within the time range of the gap duration, eliminating false gaps caused by fitting errors. For example, for the start time (timestamp) of the current link coverage gap, the position of the target ship at that start time is determined based on the ship's navigation trajectory. Based on the position of the current satellite at that start time, the feasibility of link communication between the target ship and the current satellite at that start time is determined according to step S23. If the link communication feasibility is feasible, the target ship does not exceed the coverage range of the current satellite, and it is identified as a false gap. The timestamp of that start time is excluded from the current link coverage gap, and the gap parameters are updated. It can be understood that for other single-satellite link coverage gaps, the gap parameters can be updated in the same way.

[0105] Furthermore, based on the target vessel's service disconnection threshold, the danger level of the current link coverage gap is determined. If the duration of the current link coverage gap exceeds the service disconnection threshold, the current link coverage gap is identified as a high-risk gap. In this embodiment, the service disconnection threshold for the target vessel is set based on the service level. For example, the service disconnection threshold for Level 1 services is 1 second, for Level 2 services it is 3 seconds, and for Level 3 services it is 5 seconds. If the target vessel includes Level 1 services, the service disconnection threshold is set to 1 second. When the duration of the current link coverage gap exceeds 1 second, it is marked as a high-risk gap, and information such as the vessel's position and the number of visible satellites within the duration is recorded. If the current link coverage gap is marked as a high-risk gap, an emergency handover plan trigger signal is generated simultaneously with the marking. This signal includes the start time, duration, and recommended handover satellite of the current link coverage gap, triggering the emergency handover plan. The recommended handover satellite can be determined by referring to step S23, identifying visible satellites within the duration of the current link coverage gap whose link communication feasibility is feasible for the target vessel, and using these satellites as recommended handover satellites.

[0106] Based on the fitted link feasibility probability curve, the future position of the target ship is determined by combining the ship's navigation trajectory. The gap parameters of the current link coverage gap are accurately calculated. The high-risk gap is identified by comparing with the service disconnection threshold and the emergency handover plan is triggered. This makes it easier to discover satellite handover needs in advance, predict handover time, improve handover effectiveness, and ensure link continuity and effectiveness.

[0107] In other embodiments, a coverage gap time-series distribution map can be formed with time as the horizontal axis and coverage status (binarized link feasibility probability) and risk level as the vertical axis. As a simulation result, it can intuitively present the changes in satellite coverage and single-satellite link coverage gap characteristics of the ship within the predicted future time window (e.g., 10-30s). The start time, duration, end time, and risk level information can also be simultaneously marked on the coverage gap time-series distribution map to provide visual support for emergency decision-making.

[0108] By standardizing the data, a standardized time-series dataset is constructed, features are extracted, normalized feature regions are obtained, and dimensionality and redundant information are eliminated. Probability intervals are obtained through fitting, completing the mapping from discrete data to continuous patterns. Emergency decision-making information that can be used for decision-making reference is output. Furthermore, after each step, the data accuracy or integrity is verified (e.g., data integrity after preprocessing ≥ 95%, fitting error ≤ 5%, gap judgment accuracy ≥ 99%). If the accuracy or integrity requirements are not met, a rollback mechanism is triggered (e.g., reprocessing, optimizing fitting parameters) to ensure the reliability of the output.

[0109] Furthermore, for short-cycle coverage gap scenarios where the duration of the current link coverage gap is close to the service disconnection threshold (e.g., the service disconnection threshold is 1 second, and the duration of the current link coverage gap is...), A quadratic fitting verification method is used to expand the fitting time window and combine it with satellite orbit prediction to correct the gap duration, thereby reducing the misjudgment rate. For example, the simulation performance indicators obtained through quadratic fitting verification are as follows:

[0110] In other embodiments, after constructing the link feasibility time-series dataset in step S2411, an LSTM neural network model can be used to directly output the start and end times, duration, and risk level prediction results of the coverage gap. The LSTM neural network model is trained based on historical starship visibility data and ship trajectory data. It is applicable to scenarios with complex constellation layouts and frequent ship maneuvers (such as near-shore multi-island areas and high-speed maneuvering ships). Its advantages include a 5% improvement in prediction accuracy compared to the original scheme, and better adaptation to dynamic and complex scenarios; its disadvantages include a 10% increase in model inference time compared to the original scheme, and certain additional requirements on hardware computing power.

[0111] Step S243: Based on the current link coverage gap, determine the link switching candidate time period and calculate the candidate visibility duration of each visible satellite during the link switching candidate time period.

[0112] Based on the current link coverage gap, candidate time periods for link switching are determined, and the candidate visibility duration of each satellite during the candidate link switching time period is calculated to provide a reference for subsequent selection of candidate satellites. In this embodiment, to ensure that candidate satellites can provide stable coverage before and after the current link coverage gap, candidate satellites are selected from available satellites in orbit. The selection criteria must be met simultaneously: geometric feasibility, timing compatibility, and performance compliance. Satellites with faults, obstructions, or insufficient performance are excluded.

[0113] Step S2431: Based on the current link coverage gap and in conjunction with the preset warning gap, determine the candidate time period for link switching.

[0114] In this embodiment, a warning gap is preset based on the advance warning requirement for high-risk gaps, and a candidate time period for link switching is determined according to the current link coverage gap and the warning gap. For example, the start time of the current link coverage gap... End time Based on the advance warning requirements for high-risk gaps (2 time steps), set the warning interval as follows: Then the candidate time period for link handover is ( In other embodiments, the preset warning gap includes a pre-warning gap and a post-warning gap, which are set according to the advance warning requirement for high-risk gaps and the stability requirement for high-risk gap switching, respectively. For example, the advance warning requirement for high-risk gaps is... High-risk gap switching stability requirements are Set the warning interval to be The interval after the warning is Then the candidate time period for link handover is ( ).

[0115] Step S2432: During the link switching candidate period, each of the visible satellites is screened to determine the recommended switching satellite.

[0116] During the link switching candidate period, visible satellites are screened to determine recommended switching satellites. In this embodiment, the screening of visible satellites includes geometric feasibility screening, temporal adaptation screening, and performance compliance screening. Satellites with feasible link communication with the target ship during the link switching candidate period are selected as recommended switching satellites. Specifically, geometric feasibility screening, based on step S232, screens visible satellites with feasible link geometry during the link switching candidate period, i.e., visible satellites that meet the following conditions: no Earth obstruction between the satellite and the ship (the distance between the satellite and the ship is less than the tangent length between the satellite and the Earth's surface), elevation angle constraint (the elevation angle of the satellite is not less than the elevation angle threshold), and distance constraint (the distance between the satellite and the ship is not greater than the maximum effective distance of the satellite communication link). Temporal adaptation screening, based on step S242, screens visible satellites that meet the condition of no single-satellite link coverage gap during the link switching candidate period, ensuring that a communication link can be established in advance before the current link coverage gap and that continuous coverage can be maintained after the current link coverage gap, ensuring a seamless link switching transition. Performance compliance screening, based on step S233, screens visible satellites with feasible link pointing during the link switching candidate period. Furthermore, the Starship link attenuation value can be calculated based on the satellite payload power and antenna gain. Visible satellites with a Starship link attenuation value not exceeding the Starship link attenuation threshold (e.g., 120dB) can be selected. Based on antenna performance and link stability, visible satellites with excessive pointing deviation or signal attenuation exceeding the standard can be excluded. For example, the Starship link attenuation value... In the formula, This indicates the payload power (the transmission power of the satellite payload). Indicates the satellite transmitting antenna gain. Indicates the target ship's receiving antenna gain. Indicates spatial link attenuation. , This represents the attenuation parameter, which is set to 32.45 in this embodiment. Indicates the distance to the starship. Indicates the satellite carrier frequency.

[0117] Step S2433: Calculate the candidate visibility duration of the recommended switching satellite based on the feasibility probability curves of each link.

[0118] Based on the feasibility probability curves of each link, the candidate visibility duration is calculated for the recommended switching satellites selected through screening. In this embodiment, the candidate visibility duration is the core indicator for screening candidate satellites, referring to the total duration during which the satellite meets the link communication feasibility requirement within the link switching candidate period. Specifically, the binarized link feasibility change curves of the recommended switching satellites are extracted from all the binarized link feasibility change curves. Based on the binarized link feasibility change curves of each recommended switching satellite, all visible sub-periods of each recommended switching satellite within the link switching candidate period are extracted. That is, for each recommended switching satellite, all binarized feasibility probabilities within the link switching candidate period are extracted. The consecutive timestamps are used as multiple visible sub-time periods. The durations of all visible sub-time periods are summed to obtain the candidate visible duration. For example, the first... The number of visible sub-time periods for each recommended switching satellite within the link switching candidate time period is ( ), then the first Candidate visibility duration for recommended satellite switching In the formula, , They represent the first The first recommended satellite switching option, the... The start and end times of each visible sub-period. Indicates the number of visible sub-time periods. This represents the time step, supplementing the end time with a time step to avoid truncation errors. In this embodiment, the unit is uniformly seconds (s), and the calculation is accurate to 0.1s. It is understood that the link switching candidate time period includes the current link coverage gap. Based on the principle for determining the current link coverage gap, the binarization feasibility probability of the current satellite within the current link coverage gap is calculated. Therefore, when extracting all visible sub-time periods within the candidate time period of link switching based on the binarized link feasibility change curve of each recommended switching satellite, the binarized link feasibility change curve of the current satellite can be ignored. That is, if the recommended switching satellite includes the current satellite, the current satellite will be removed from the recommended switching satellites to reduce the amount of computation.

[0119] Furthermore, to mitigate the impact of the optimal Starship elevation angle on link stability, elevation angle weights are set, and candidate visibility durations are determined using weighted summation. In the formula, Indicates the first The elevation angle weighting for each visible sub-period. For example, if the Starship elevation angle corresponding to the visible sub-period is within the optimal Starship elevation angle range (e.g., ... If the inter-ship link is highly stable, then... To increase the contribution of the optimal Starship elevation angle visible sub-period to the candidate visible duration; if the Starship elevation angle corresponding to the visible sub-period is not within the optimal Starship elevation angle range, then .

[0120] Step S244: Determine the candidate satellites based on the candidate visibility duration and the satellite link quality corresponding to each of the visible satellites.

[0121] Candidate satellites are determined based on the candidate visibility duration and satellite link quality corresponding to each visible satellite. In this embodiment, candidate satellites are determined based on the candidate visibility duration and satellite link quality corresponding to the screened recommended switching satellites. Satellite link quality is a secondary indicator in the candidate satellite selection process, quantified using a multi-dimensional weighted scoring method. The scoring dimensions include three main dimensions: geometric parameters, signal attenuation, and stability. A higher score indicates better link quality. In other embodiments, the signal-to-noise ratio (SNR) can be used as the satellite link quality, or the satellite link quality can be directly obtained as initial simulation data.

[0122] A dual-index progressive ranking strategy is adopted. First, the recommended satellites are ranked according to their candidate visibility duration. Then, they are fine-tuned based on their satellite link quality to ensure that the ranking balances coverage continuity and link stability, forming a recommended satellite list. In this embodiment, the top-ranked recommended satellite is selected as a candidate satellite, and its satellite ID is obtained. The recommended satellite list serves as a component of the simulation results, providing a basis for satellite switching decisions. For example, the satellite ranking priority scoring model is represented as follows: In the formula, Indicates the satellite sorting priority. This indicates the visibility duration weight, for example, 0.6. Indicates the duration the candidate is visible. This represents the satellite link quality weight, for example, 0.4. This represents the satellite link quality. In other embodiments, the satellite ranking priority can be calculated based on the candidate visibility duration, with satellite link quality as the quality weight, and the satellites can be ranked according to this priority. For example, satellite link quality can be scored out of 100, and the link quality weight can be... In the formula Indicates the first Satellite link quality of each satellite, satellite ranking priority .

[0123] Candidate satellite selection is the core basis for the current emergency response plan for gaps in link coverage. By selecting satellites that can provide stable coverage and meet the link quality standards during the candidate period of link switching, the candidate satellites are determined based on the candidate visibility duration and satellite link quality to ensure communication continuity after switching.

[0124] By integrating the SGP4 orbital dynamics model and Kalman filtering, precise coupling between the orbital characteristics of the LEO sparse constellation and the dynamic navigation of ships is achieved, providing real-time output of the relative geometric relationship between the satellite and the ship and the effective communication link. The coupling modeling error can be reduced to less than 3%. Based on satellite ephemeris data and ship navigation trajectories (ship trajectory prediction results), the gap parameters of the current link coverage gap (including start time, end time, duration, and hazard level) are predicted through a time-series analysis model. Through a multi-factor fusion algorithm for predicting the current link coverage gap, the prediction error can be reduced to no more than 100ms. The accurate prediction of coverage gaps allows sufficient time for link switching in advance, improving the initiative of link switching and increasing the link switching success rate to over 99.5%, reducing the risk of communication interruption by 80% compared to traditional technologies.

[0125] Step S3: Based on the constellation parameters and the ship data, calculate the elevation angle switching threshold between the target ship and the current satellite, and determine the link switching time corresponding to the target ship; wherein, the current satellite is the satellite with which the target ship has a communication link.

[0126] Based on constellation parameters and ship data, the elevation angle switching threshold between the target ship and the current satellite is calculated to determine the link switching time corresponding to the target ship. This breaks through the fixed threshold switching mode and achieves intelligent decision-making for link switching based on a dynamic threshold adaptive algorithm, taking into account ship data packets, ship speed, and ship service priority. Step S3 specifically includes steps S31-S34.

[0127] Step S31: Based on the ship data and the constellation parameters, extract the target ephemeris data and target orbital period parameters of the current satellite, and the candidate satellite link quality corresponding to the candidate satellite.

[0128] The target vessel's speed (in this embodiment, SOG) and service priority are obtained from its ship data, along with the current satellites connected to the target vessel via communication links, including the satellite IDs of the current satellites.

[0129] Based on the current satellite's satellite ID, extract the ephemeris data of the current satellite from the constellation parameters as the target ephemeris data, extract the orbital period parameters of the current satellite as the target orbital period parameters, and extract the satellite link quality of the candidate satellite from the constellation parameters based on the candidate satellite's satellite ID as the candidate satellite link quality.

[0130] Step S32: Calculate the rate of change of the target ship's elevation angle relative to the current satellite based on the ship data, the target ephemeris data, and the target orbital period parameters.

[0131] Based on the ship data, target ephemeris data, and target orbital period parameters, calculate the current satellite's ship elevation angle change rate. For example, In the formula, , These represent the current time step. , the previous time step The corresponding target vessel and the current satellite's current elevation angle. Based on the target ephemeris data and target orbital period parameters, determine the current satellite's position at the current time step. , the previous time step The corresponding ECEF coordinates of the target satellite are determined; based on the ship data, the target ship's position at the current time step is determined. , the previous time step Based on the corresponding longitude / latitude, ship's heading relative to the ground, and ship's attitude angle, refer to step S231 to calculate the current time step. , the previous time step The corresponding starship elevation angles.

[0132] Step S33: Calculate the elevation angle switching threshold based on the ship's speed, the ship's service priority, the candidate satellite link quality, and the satellite ship's elevation angle change rate.

[0133] First, the priorities of ship operations will be quantified. For example, ship operations priorities include primary, secondary, and tertiary operations, and the quantified priorities are as follows: , , .

[0134] Secondly, the elevation angle switching threshold is calculated based on the ship's speed, quantified ship service priority, candidate satellite link quality, and satellite-ship elevation angle change rate. For example, In the formula, Indicates the elevation angle switching threshold. This represents the basic elevation angle threshold, which is used in this embodiment. , , , , These represent the switching weight coefficients, which can be obtained through training with historical test data. In this embodiment, the value range is [value range missing]. , Indicates the ship's speed. Indicates the rate of change of the starship's elevation angle. This represents the quantified vessel service priority. If the target vessel has multiple service priorities, the highest priority is used. In other embodiments, the average priority can also be used. The signal-to-noise ratio (SNR) is used to represent the quality of the candidate satellite link in this embodiment.

[0135] Step S34: Based on the starship elevation angle change rate, calculate the moment when the current starship elevation angle of the target ship and the current satellite reaches the elevation angle switching threshold, and use it as the link switching moment.

[0136] Based on the satellite's elevation angle change rate, the moment when the target satellite's current elevation angle reaches the elevation angle switching threshold is calculated and taken as the link switching moment. For example, the link switching moment... At this point, the target ship and the current satellite's current elevation angle reach the elevation angle switching threshold, and a link switch is performed.

[0137] By adopting a dynamic threshold adaptive algorithm, four influencing factors are introduced: ship speed, elevation angle change rate, ship service priority, and candidate satellite link quality. A dynamic threshold calculation model is constructed to achieve intelligent adaptation at the time of link switching. Compared with the fixed threshold switching mode, the link interruption rate is reduced by more than 80%.

[0138] In other embodiments, to ensure applicability to complex link environments, and for situations where some influencing parameters in complex link environments cannot be accurately quantified, a fuzzy logic algorithm is used to determine the elevation angle switching threshold. Ship speed, elevation angle change rate, ship service priority, and candidate satellite link quality are fuzzified, and a dynamic switching threshold is output based on a preset fuzzy rule base, replacing the switching weight coefficient. , , , The calculation method using fuzzy logic algorithms reduces the accuracy of threshold adjustment by 3%-5% compared to the switching weight coefficient scheme, but it is easier to implement and does not require precise calibration of parameter thresholds. This makes it suitable for scenarios where the stability requirements for switching priorities in ship operations are not high. A comprehensive decision can be made based on the actual application scenario, hardware computing power, and cost budget, prioritizing the performance requirements of core businesses (primary businesses), and determining the elevation angle switching threshold calculation scheme by considering accuracy, efficiency, and cost.

[0139] By using a dynamic switching threshold adaptive algorithm, which integrates multiple influencing factors to adjust the switching timing in real time, the method adapts to different flight speeds, business types, and constellation coverage densities, thereby enhancing scenario compatibility. The scenario adaptability of the simulation method is improved by more than 60% compared to the traditional fixed threshold scheme.

[0140] Step S4: At the link switching time, the target ship switches the communication link and establishes a communication connection with the candidate satellite.

[0141] At the link switching moment, the target ship switches the communication link and establishes a communication connection with the candidate satellite, specifically including steps S41-S43.

[0142] Step S41: Calculate the switching synchronization delay based on the ship data and the constellation parameters, and determine whether the switching synchronization delay is greater than the switching delay threshold.

[0143] Based on ship data and constellation parameters, the handover synchronization delay is calculated, and it is determined whether the handover synchronization delay exceeds the handover delay threshold. In this embodiment, the handover synchronization delay includes carrier synchronization delay, frame synchronization delay, and dynamic compensation delay. The handover delay threshold is set to 0.5s to meet the 1s service disconnection threshold requirement for Level 1 services.

[0144] Among them, the carrier synchronization delay is calculated based on a phase-locked loop (PLL) model simulation, adapting to the Doppler frequency offset caused by the relative motion between the satellite and the ship. For example, the carrier synchronization delay... In the formula, This represents the standard deviation of Doppler frequency offset, in Hz, for example, in the LEO sparse constellation. , For phase-locked loop gain, it is typically between 0.1 and 0.5. The phase-locked loop bandwidth is expressed in Hz, and the adaptive frequency offset range is set to... .in, The standard deviation of the Doppler frequency offset can be determined based on historical measured data. The Doppler frequency offset can be calculated using the relative velocity of the starship. ( For the transmission carrier frequency, At the speed of light, (The relative velocity of the starship) is used to statistically determine the standard deviation of the Doppler frequency offset.

[0145] Frame synchronization delay, simulated based on a synchronization code detection model, is calculated using a fixed synchronization header (e.g., an 8-byte PN code). For example, the frame synchronization delay... In the formula, Synchronization code length, in bits. The symbol period, measured in seconds, is determined by the transmission rate; for example, a transmission rate of 1 Mbps corresponds to... , In this embodiment, the synchronization code detection probability is set as follows: .

[0146] Dynamic compensation delay is used to compensate for synchronization deviations caused by simulated satellite-ship dynamics (changes in ship heading, satellite orbital motion). Dynamic compensation delay is a compensation value dynamically adjusted based on real-time link status (such as distance changes, Doppler shift). In this embodiment, dynamic compensation delay includes propagation-dynamic compensation delay and frequency offset-dynamic compensation delay, used to compensate for changes in propagation delay (distance changes caused by changes in satellite / ship position), compensate for phase / time offsets caused by Doppler shift, and dynamically adjust to adapt to fixed backoff delay. For example, propagation delay... In the formula, At the speed of light, for The distance to the starship at any given time is used as the dynamic compensation delay, which is the difference between the current propagation delay and the propagation delay at the reference time. In the formula For propagation - dynamic compensation of delay, The propagation delay is a reference time (e.g., the propagation delay at the communication establishment time); Doppler frequency offset causes the phase to change linearly with time, and the corresponding frequency offset-dynamic compensation delay is... , For the transmission carrier frequency, for Timing phase shift. In this embodiment, dynamic compensation delay is reduced by real-time prediction of motion parameters to compensate in advance; calculated based on the maximum motion speed, it typically does not exceed 0.1s. In other embodiments, considering the propagation-dynamic compensation delay and frequency offset-dynamic compensation delay under ideal conditions, dynamic compensation is applied to the total propagation delay. Furthermore, by incorporating the fixed backoff delay, the corrected dynamic compensation delay is determined. , The device processing delay can be set based on the device information; This is the retreat coefficient, which can be set based on experience.

[0147] In this embodiment, a link synchronization simulation model is built using Matlab / Simulink, integrating the SGP4 satellite orbit model and the ship maneuver model to recreate the synchronization process under the dynamic scenario of the starship. The synchronization module uses a standard digital phase-locked loop and a PN code synchronization detector, with parameters adapted to the actual communication protocol.

[0148] Step S42: If the value is not greater than the target satellite's value, the target ship switches its communication link at the link switching time and establishes a communication connection with the candidate satellite.

[0149] If the switching synchronization delay is not greater than the switching delay threshold, it is determined that the target ship and the candidate satellite can establish a communication link in a timely manner, the target ship's communication link switching can be successful, and the target ship will perform the link switching at the time of link switching to establish a communication connection with the candidate satellite.

[0150] Furthermore, step S4 also includes: Step S44: If the value is greater than the threshold, update the ship data and return to step S1 until the switching synchronization delay is not greater than the switching delay threshold. Then, redetermine the link switching time. At the new link switching time, the target ship performs communication link switching and establishes a communication connection with the candidate satellite.

[0151] If the switching synchronization delay is greater than the switching delay threshold, it is expected that the target ship and the candidate satellite will not be able to establish a communication link in time. In this case, the ship data is updated, for example, the ship speed is adjusted, and the process returns to step S1. Steps S1-S4 are then re-executed until the switching synchronization delay is no greater than the switching delay threshold. The new link switching time and candidate satellite determined by re-executing steps S1-S4 are then obtained. At the new link switching time, the target ship performs a communication link switch and establishes a communication connection with the new candidate satellite.

[0152] Furthermore, when using this method to simulate ship communication link switching in a constellation planning scenario, the constellation parameters can be updated. Based on the updated ship data and constellation parameters, steps S1-S4 are re-executed until the switching synchronization delay is no greater than the switching delay threshold. The link switching result, the recalculated switching synchronization delay, the updated ship data, and the constellation parameters are then used as the simulation result.

[0153] In other embodiments, the synchronization algorithm parameters can be adjusted (such as increasing the phase-locked loop gain or optimizing the synchronization code length), or candidate satellites can be reselected from the candidate switching satellite list to ensure that the total switching delay (resource allocation delay and switching synchronization delay) does not exceed the service disconnection threshold.

[0154] Furthermore, corresponding thresholds can be set for carrier synchronization delay and frame synchronization delay, such as a carrier synchronization delay threshold (e.g., 0.2s) and a frame synchronization delay threshold (e.g., 0.2s). When the carrier synchronization delay and / or frame synchronization delay do not meet the corresponding threshold requirements, the ship data and / or constellation parameters are updated.

[0155] By quantifying the handover synchronization delay, a decision reference is provided for the communication link handover execution of the target vessel. Through the handover delay threshold and the update mechanism of vessel data and constellation parameters, the handover synchronization delay is guaranteed not to exceed the handover delay threshold, thereby improving the prediction accuracy and ensuring that the link handover meets the business needs of the target vessel.

[0156] Furthermore, to reduce link switching errors, three key verification indicators are monitored: synchronization delay (resource allocation delay, switching synchronization delay), synchronization error, and bit error rate. This ensures that within 100ms after link switching, the synchronization error is ≤1% of the symbol period, and the bit error rate is ≤10%. -6 This meets the requirements for seamless integration of primary business functions.

[0157] This invention provides a method for predicting ship communication link handover for LEO sparse constellations, such as... Figure 2 As shown, it includes: The data acquisition module is used to acquire the constellation parameters of the LEO constellation and the ship data of the target vessel; the LEO constellation includes multiple satellites; The constellation-ship coupling modeling module is used to perform constellation-ship coupling modeling based on the constellation parameters and the ship data, and to determine candidate satellites. The dynamic switching decision module is used to calculate the elevation angle switching threshold between the target ship and the current satellite based on the constellation parameters and the ship data, and to determine the link switching time corresponding to the target ship; wherein, the current satellite is the satellite with which the target ship has a communication link; The dynamic switching decision module is also used to enable the target ship to switch communication links and establish a communication connection with the candidate satellite at the link switching time.

[0158] The above-described method and system embodiments are based on the same principles, and their related aspects can be referenced from each other to achieve the same technical effects. For specific implementation processes, please refer to the foregoing embodiments, which will not be repeated here.

[0159] In summary, the ship communication link handover prediction method and system for LEO sparse constellations according to embodiments of the present invention have at least one of the following beneficial effects: 1. Calculate the elevation angle switching threshold based on constellation parameters and ship data to ensure dynamic adaptation between the elevation angle switching threshold and constellation status, ship dynamic characteristics and business requirements, thereby improving multi-scenario adaptability; determine the link switching time based on the elevation angle switching threshold to accurately predict coverage gaps, proactively prepare for switching, improve communication continuity, and enhance the ability to guarantee critical services.

[0160] 2. Based on the constellation parameters, a sparse constellation topology model is constructed to determine the constellation topology data. At the same time, based on ship data, a ship maneuvering model is used to predict the ship's navigation trajectory and determine the feasibility of link communication between the target ship and each visible satellite. Then, candidate satellites are determined. Combined with the prediction of the ship's future trajectory, the current link coverage gap is determined, and then candidate satellites are determined. This achieves accurate prediction of the current link coverage gap, ensures the feasibility of link switching to candidate satellites, and guarantees the quality of communication links.

[0161] 3. Based on ship data and the target ephemeris data and target orbital period parameters of the current satellite, calculate the satellite-ship elevation angle change rate. Combined with the ship's speed on the trajectory, ship service priority, and candidate satellite link quality, calculate the elevation angle switching threshold. This enables dynamic determination of the elevation angle switching threshold for different scenarios, improving multi-scenario adaptability. Furthermore, the link switching time is determined based on the elevation angle change rate, accurately predicting the link switching execution time and providing a time benchmark for resource preparation and scheduling before the switch.

[0162] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0163] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for predicting ship communication link handover for LEO sparse constellations, characterized in that, include: Obtain the constellation parameters of the LEO constellation and the ship data of the target vessel; wherein, the LEO constellation includes multiple satellites; the ship data packet includes ship speed and ship service priority; Based on the constellation parameters and the ship data, constellation-ship coupling modeling is performed to determine candidate satellites; wherein, the candidate satellites are those that can stably cover the candidate period of link switching and whose link quality meets the standards; Based on the constellation parameters and the ship data, the elevation angle switching threshold between the target ship and the current satellite is calculated, and the link switching time corresponding to the target ship is determined; wherein, the current satellite is the satellite with which the target ship has a communication link; At the link switching time, the target ship switches its communication link and establishes a communication connection with the candidate satellite; The elevation angle switching threshold is calculated using the following steps: Based on the ship data and the constellation parameters, extract the target ephemeris data and target orbital period parameters of the current satellite, as well as the candidate satellite link quality corresponding to the candidate satellite; Based on the ship data, the target ephemeris data, and the target orbital period parameters, calculate the rate of change of the target ship's elevation angle relative to the current satellite; The elevation angle switching threshold is calculated based on the ship's speed, the ship's service priority, the quality of the candidate satellite link, and the rate of change of the satellite's elevation angle.

2. The method according to claim 1, characterized in that, The constellation parameters include constellation deployment progress, ephemeris data of each satellite, and orbital period parameters; Based on the constellation parameters and the ship data, constellation-ship coupling modeling is performed to determine candidate satellites, including: Based on the constellation deployment progress, the ephemeris data, and the orbital period parameters, the visible satellites and their ECEF coordinates at multiple time steps are determined according to a preset time step; based on the ship data, the ship maneuvering model is used to predict the ship's navigation trajectory. Based on the ship data, the ship's navigation trajectory, and the ECEF coordinates of each of the satellites, at each of the time steps, the feasibility of link communication between the target ship and each of the visible satellites is determined; The candidate satellites are determined based on the ship data, the constellation parameters, the ship's navigation trajectory, and the link communication feasibility between the target ship and each visible satellite at each time step.

3. The method according to claim 2, characterized in that, The ship data also includes ship antenna parameters; Based on the ship data, the ship's navigation trajectory, and the ECEF coordinates of each of the satellites, at each of the time steps, the feasibility of link communication between the target ship and each of the visible satellites is determined, including: Based on the ship's navigation trajectory and the ECEF coordinates of each of the satellites, at each of the time steps, the distance between the target ship and each of the visible satellites, the elevation angle of the satellite, and the azimuth angle of the satellite are calculated. At each time step, the geometric feasibility of the link between the target vessel and each visible satellite is determined based on the distance between each of the satellites and the elevation angle of the satellites. Based on the elevation angle, azimuth angle, and antenna parameters of each satellite, determine the feasibility of link pointing between the target ship and each visible satellite; Based on the geometric feasibility and directional feasibility of each link, the feasibility of link communication between the target ship and each of the visible satellites is determined.

4. The method according to claim 2, characterized in that, The constellation parameters also include the satellite link quality of each of the satellites; Based on the ship data, the constellation parameters, the ship's navigation trajectory, and the link communication feasibility between the target ship and each visible satellite at each time step, the candidate satellites are determined, including: The ship's navigation trajectory, each of the orbital period parameters, and the link communication feasibility between the target ship and each of the visible satellites corresponding to each time step are fitted to obtain the link feasibility probability curve between the target ship and each of the visible satellites; Based on the feasibility probability curves of each link, determine the current link coverage gap; Based on the current link coverage gap, determine the link switching candidate time period, and calculate the candidate visibility duration of each visible satellite during the link switching candidate time period; The candidate satellites are determined based on the candidate visibility duration and the satellite link quality corresponding to each of the visible satellites.

5. The method according to claim 4, characterized in that, The feasibility of link communication between the target ship and each visible satellite at each time step is fitted to the ship's navigation trajectory, each orbital period parameter, and the link communication feasibility probability curve between the target ship and each visible satellite at each time step, including: The ship's navigation trajectory, each of the orbital period parameters, and each of the link communication feasibility are standardized to construct a link feasibility time series dataset; Feature extraction is performed on the aforementioned link feasibility time-series dataset to obtain a feasibility time-series feature matrix; The feasibility time-series feature matrix is ​​fitted, and the feasibility probability corresponding to each timestamp is calculated. All the timestamps and their corresponding feasibility probabilities constitute the link feasibility probability curve.

6. The method according to claim 4, characterized in that, Based on the current link coverage gap, candidate time periods for link switching are determined, and the candidate visibility duration of each visible satellite during the candidate time period for link switching is calculated, including: Based on the current link coverage gap and in conjunction with the preset warning gap, candidate time periods for link handover are determined; During the link switching candidate period, each of the visible satellites is screened to determine the recommended switching satellite; Based on the feasibility probability curves of each link, the candidate visibility duration of the recommended switching satellite is calculated.

7. The method according to claim 1, characterized in that, Based on the constellation parameters and the ship data, calculate the elevation angle switching threshold between the target ship and the current satellite, and determine the link switching time of the target ship, including: Based on the starship elevation angle change rate, the moment when the current starship elevation angle of the target ship and the current satellite reaches the elevation angle switching threshold is calculated and used as the link switching moment.

8. The method according to claim 2, characterized in that, Based on the constellation deployment schedule, the ephemeris data, and the orbital period parameters, the visible satellites and their ECEF coordinates at multiple time steps are determined, including: Quantify the deployment progress of the aforementioned constellations; Based on the quantified constellation deployment progress, the ephemeris data, and the orbital period parameters, a sparse constellation topology model is constructed, and constellation topology data for multiple time steps are obtained according to the preset time step. The sparse constellation topology model includes a data access layer, a physical computing layer, a topology construction layer, and an output service layer. The constellation topology data includes the visible satellites and the ECEF coordinates of the satellites.

9. The method according to claim 2, characterized in that, The ship data also includes ship AIS data; Based on the ship data, the ship's navigation trajectory is predicted using a ship maneuvering model, including: Obtain the current state vector from the ship's AIS data; The current state vector is input into the ship maneuvering model, and the predicted state vectors for multiple time steps are obtained according to the preset time step. The ship's navigation trajectory is generated based on the current state vector and each predicted state vector.

10. A ship communication link handover prediction system for LEO sparse constellations, characterized in that, include: The data acquisition module is used to acquire constellation parameters of the LEO constellation and ship data of the target vessel; wherein, the LEO constellation includes multiple satellites; the ship data packet includes ship speed and ship service priority; The constellation-ship coupling modeling module is used to perform constellation-ship coupling modeling based on the constellation parameters and the ship data, and to determine candidate satellites; wherein, the candidate satellites are those that can stably cover the link switching candidate period and whose link quality meets the standards; The dynamic switching decision module is used to calculate the elevation angle switching threshold between the target ship and the current satellite based on the constellation parameters and the ship data, and to determine the link switching time corresponding to the target ship; wherein, the current satellite is the satellite with which the target ship has a communication link; The dynamic switching decision module is also used to enable the target ship to switch communication links and establish a communication connection with the candidate satellite at the link switching time. The dynamic handover decision module calculates the elevation angle handover threshold through the following steps: Based on the ship data and the constellation parameters, extract the target ephemeris data and target orbital period parameters of the current satellite, as well as the candidate satellite link quality corresponding to the candidate satellite; Based on the ship data, the target ephemeris data, and the target orbital period parameters, calculate the rate of change of the target ship's elevation angle relative to the current satellite; The elevation angle switching threshold is calculated based on the ship's speed, the ship's service priority, the quality of the candidate satellite link, and the rate of change of the satellite's elevation angle.

Images