Drift trajectory prediction method, device and equipment of floating object and storage medium

Abstract

Embodiments of the present disclosure relate to a method, device, equipment and storage medium for predicting a drift trajectory of a floating object. The method comprises: performing numerical simulation based on a sea state simulation model to determine sea state characteristic parameters of each preset grid point at each preset time; determining an estimated sea state characteristic parameter of a position of the floating object at a jth preset time based on the position of the floating object at the jth preset time, positions of each preset grid point and the sea state characteristic parameters of each preset grid point at the jth preset time, where j=1, 2, …, N-1; calculating a predicted drift speed of the floating object based on the estimated sea state characteristic parameter corresponding to the jth preset time, and calculating a position of the floating object at a (j+1)th preset time; and determining the drift trajectory of the floating object according to the position of the floating object at each preset time. The embodiments can more accurately predict the drift trajectory of the floating object, thereby reducing the search area of the floating object.

Description

Technical Field

[0001] This disclosure relates to the field of maritime technology, and in particular to a method, apparatus, device, and storage medium for predicting the drift trajectory of floating objects. Background Technology

[0002] Maritime accidents can generate various types of floating debris. For example, an oil spill may produce oil slicks; while a ship collision may produce debris such as collision fragments, life rafts, and wrecked vessels.

[0003] To quickly salvage and clear various floating objects after a maritime accident, it is necessary to determine the approximate and accurate area where the floating objects are located, that is, to determine their drift trajectory. Existing drift trajectory prediction methods only predict the general drift direction based on sea state characteristics, and cannot accurately estimate the drift trajectory of floating objects, and therefore cannot accurately determine the floating area of ​​the objects. Summary of the Invention

[0004] To solve the above-mentioned technical problems, or at least partially solve them, this disclosure provides a method, apparatus, device, and storage medium for predicting the drift trajectory of floating objects.

[0005] In a first aspect, embodiments of this disclosure provide a method for predicting the drift trajectory of a floating object, including:

[0006] The sea state observation parameters within the assimilation time window are assimilated to obtain the assimilated sea state simulation model.

[0007] Numerical simulation is performed based on the sea state simulation model to determine the sea state characteristic parameters of each preset grid point at the i-th preset time in the period to be predicted, i = 1, 2, ..., N;

[0008] Based on the location of the floating object to be predicted at the j-th preset time, the location of each preset grid point, and the sea state characteristic parameters at the j-th preset time, the estimated sea state characteristic parameters of the location of the floating object to be predicted at the j-th preset time are determined, j = 1, 2, ..., N-1, where the location corresponding to j = 1 is known;

[0009] The predicted drift velocity of the floating object to be predicted is calculated based on the estimated sea state characteristic parameters corresponding to the j-th preset time, and the position of the floating object to be predicted at the (j+1)-th preset time is calculated based on the predicted drift velocity.

[0010] The drift trajectory of the floating object is determined based on its position at each preset time.

[0011] Optionally, the estimated sea state characteristic parameters include surface current characteristic parameters, wind characteristic parameters, and wave characteristic parameters;

[0012] The predicted drift velocity of the object to be predicted is calculated based on the estimated sea state characteristic parameters corresponding to the j-th preset time, including:

[0013] The flow-induced drift velocity is determined based on the surface water flow characteristic parameters, the wind-induced drift velocity is determined based on the wind characteristic parameters, and the wave-induced drift velocity is determined based on the wave characteristic parameters.

[0014] The predicted drift velocity is determined based on the flow-induced drift velocity, the wind-induced drift velocity, and the wave-induced drift velocity.

[0015] Optionally, determining the wave-induced drift velocity based on wave characteristic parameters includes:

[0016] The wave feature parameters are input into a pre-trained wave-induced drift deep learning model to obtain the wave-induced drift velocity.

[0017] The wave-induced drift deep learning model is trained based on the first sample drift data, which includes the sample drift velocity observation value of the sample floating object and the corresponding sample surface water flow characteristic parameters, sample wind characteristic parameters and sample wave characteristic parameters.

[0018] Optionally, determining the flow-induced drift velocity based on the surface water flow characteristic parameters includes: calculating the flow-induced drift velocity according to the surface water flow characteristic parameters and the flow-induced drift coefficient;

[0019] Determining the wind-induced drift velocity based on the wind characteristic parameters includes: calculating the wind-induced drift velocity according to the wind characteristic parameters and the wind-induced drift coefficient;

[0020] The flow-induced drift coefficient and the wind-induced drift coefficient are obtained during the training of the wave-induced drift deep learning model.

[0021] Optionally, before calculating the flow-induced drift velocity based on the surface water flow characteristic parameters and the flow-induced drift coefficient, the method further includes: adding a random perturbation to the flow-induced drift coefficient to obtain the perturbed flow-induced drift coefficient;

[0022] Before calculating the wind-induced drift velocity based on the wind characteristic parameters and the wind-induced drift coefficient, the method further includes: adding a random perturbation to the wind-induced drift coefficient to obtain the perturbed wind-induced drift coefficient.

[0023] Optionally, before calculating the wind-induced drift velocity based on the wind characteristic parameters and the wind-induced drift coefficient, the method further includes:

[0024] The predicted drift bias is determined based on the estimated sea state characteristic parameters corresponding to the floating object at the j-th preset time.

[0025] The wind-induced drift coefficient is selected based on the predicted drift bias.

[0026] Optionally, determining the predicted drift bias based on the estimated sea state characteristic parameters corresponding to the object to be predicted at the j-th preset time includes:

[0027] The estimated sea state characteristic parameters are input into the pre-trained drift bias deep learning model to determine the predicted drift bias of the floating object.

[0028] The biased deep learning model is trained based on the second sample drift data, which includes the drift biased observations of the sample floating objects and the sample sea state characteristic parameters.

[0029] In a second aspect, embodiments of this disclosure provide a device for predicting the drift trajectory of a floating object, comprising:

[0030] The simulated sea state model building unit is used to assimilate the sea state observation parameters within the assimilation time window to obtain the assimilated sea state simulation model.

[0031] The numerical simulation unit is used to perform numerical simulation based on the sea state simulation model to determine the sea state characteristic parameters of each preset grid point at the i-th preset time in the period to be predicted, i = 1, 2, ..., N;

[0032] The estimated sea state characteristic parameter determination unit is used to determine the estimated sea state characteristic parameters of the floating object to be predicted at the j-th preset time based on the position of the floating object to be predicted at the j-th preset time, the position of each preset grid point and the sea state characteristic parameters at the j-th preset time, j = 1, 2, ..., N-1, where the position corresponding to j = 1 is known;

[0033] The position determination unit is used to calculate the predicted drift velocity of the floating object to be predicted based on the estimated sea state characteristic parameters corresponding to the j-th preset time, and to calculate the position of the floating object to be predicted at the (j+1)-th preset time based on the predicted drift velocity.

[0034] The trajectory determination unit is used to determine the drift trajectory of the floating object based on its position at each preset time.

[0035] Thirdly, embodiments of this disclosure provide a terminal device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, it can implement the method described in the first aspect.

[0036] Fourthly, embodiments of this disclosure provide a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the method described in the first aspect.

[0037] The technical solution provided in this disclosure has the following advantages compared with the prior art:

[0038] The solution provided in this disclosure first assimilates the observed sea state parameters within the assimilation time window to obtain an assimilated sea state simulation model. Then, based on the assimilated sea state simulation model, the sea state characteristic parameters for each preset time point are determined. Next, the sea state characteristic parameters for the corresponding position of the floating object at the j-th preset time point are determined iteratively, and the predicted drift velocity at that position is determined. The position at the (j+1)-th preset time point is calculated based on the predicted drift velocity and the current position, until the positions for all preset time points are calculated. After determining the positions at the preset time points, the drift trajectory of the floating object can be determined. On the one hand, the estimated sea state characteristic parameters are obtained based on the assimilated sea state simulation model, which is more accurate. On the other hand, when calculating the drift trajectory, the uncertainties of the wind-induced drift coefficient and the current-induced drift coefficient, as well as the wind-induced drift bias, are fully considered, making the drift trajectory calculation results more accurate. Attached Figure Description

[0039] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0040] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart of the method for predicting the drift trajectory of floating objects provided in the embodiments of this disclosure;

[0042] Figure 2 This is a flowchart of a method for training a wave-induced drift deep learning model based on first sample data according to an embodiment of this disclosure;

[0043] Figure 3This is a schematic diagram of the structure of the drift trajectory prediction device for floating objects provided in the embodiments of this disclosure;

[0044] Figure 4 This is a schematic diagram of the structure of the terminal device provided in the embodiments of this disclosure. Detailed Implementation

[0045] To better understand the above-mentioned objectives, features, and advantages of this disclosure, the solutions disclosed herein will be further described below. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0046] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0047] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below. It should be noted that the concepts of "first", "second", etc., used in this disclosure are only used to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0048] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0049] This disclosure provides a method for predicting the drift velocity of floating objects on water. In practical applications, the method provided can be used in the maritime field to predict the drift trajectories of floating objects at sea. Of course, the method can also be applied to predict the drift trajectories of floating objects in inland rivers and lakes.

[0050] Figure 1 This is a flowchart of a method for predicting the drift trajectory of a floating object provided in an embodiment of this disclosure. Figure 1As shown, the drift trajectory prediction method for floating objects provided in this embodiment may include steps S110-S150.

[0051] It should be noted that the drift trajectory prediction method for floating objects provided in this embodiment is executed by a terminal device. The terminal device can be a server dedicated to data processing, or a computing device such as a laptop computer, a shipborne terminal, a personal digital assistant (PDA), or a wearable device for rescue personnel.

[0052] Step S110: Assimilate the sea state observation parameters within the assimilation time window to obtain the assimilated sea state simulation model.

[0053] In this embodiment of the disclosure, the time window to be assimilated is a time window preceding the time period to be predicted. In specific applications, the time window to be assimilated can be a time window of a relatively recent period preceding the time period to be predicted.

[0054] The sea state observation parameters within the assimilation time window can include atmospheric characteristic observation parameters, wave characteristic observation parameters, and water flow characteristic observation parameters.

[0055] Atmospheric characteristic observation parameters are observational parameters used to characterize the atmospheric features above the surface of a water body. These parameters may include at least one of atmospheric wind characteristic parameters, temperature characteristic parameters, and cloud cover characteristic parameters. In specific embodiments, atmospheric characteristic observation parameters that directly or indirectly characterize atmospheric features can be obtained by using observation methods such as atmospheric observation stations in the sea area to be predicted, high-frequency ground-wave radar, and remote sensing images acquired by remote sensing equipment (e.g., satellite cloud images acquired by remote sensing satellites).

[0056] Wave characteristic observation parameters are observation parameters used to characterize the wave characteristics of a water body. Wave characteristic parameters may include at least one of wave height, wave period, and wave direction. In this embodiment of the disclosure, wave characteristic observation parameters characterizing wave characteristics can be obtained using buoy or underway observation methods.

[0057] Water flow characteristic parameters are observational parameters used to characterize the flow characteristics of a water body. These parameters may include surface water flow velocity, direction, and turbulence characteristics. In this embodiment, buoys, ground-wave radar, and mobile observation methods can be used to acquire these water flow characteristic observational parameters.

[0058] In this embodiment of the disclosure, the sea state simulation model may include an atmospheric dynamic sub-model, a hydrodynamic sub-model, and a wave sub-model. The atmospheric dynamic sub-model may be a model constructed using WRF (Weather Research and Forecasting), the hydrodynamic sub-model may be a model constructed based on FVCOM (Finite-Volume Coastal Ocean Model) or ROMS (Regional Ocean modeling system), and the wave sub-model may be a model constructed based on SWAN (Simulating Waves Nearshore) or WaveWatch (WAVE-height, Water Depth and Current Hindcasting).

[0059] Data assimilation of sea state observation parameters within the assimilation time window involves correcting the sea state characteristics of grid points in the sea state simulation model based on the sea state observation parameters. This makes the sea state characteristics of each grid point more consistent with the actual situation, thereby enabling the sea state simulation model to more accurately predict the sea state characteristic parameters of the period to be predicted.

[0060] In practice, computing devices can employ various data assimilation methods to assimilate sea state observation parameters and obtain an assimilated sea state simulation model. For example, computing devices can use three-dimensional variational assimilation methods, ensemble Kalman filter assimilation methods, and adaptive optimal interpolation assimilation methods to assimilate sea state observation parameters and obtain a sea state simulation model.

[0061] In some embodiments of this disclosure, the method for determining the sea state simulation model using a three-dimensional variational assimilation method includes steps S1111-S1115.

[0062] Step S1111: Divide the time window to be assimilated into N sub-assimilation time windows, each of which includes sea state observation parameters.

[0063] In some embodiments of this disclosure, the computing device can perform sliding window segmentation of the assimilation window based on a predetermined sub-assimilation time window to obtain N sub-assimilation time windows. Among the N sub-assimilation time windows, adjacent sub-assimilation time windows may or may not overlap in time, and this disclosure does not impose any particular limitation.

[0064] In other embodiments of this disclosure, the computing device may also employ a random partitioning method to divide the assimilation window into N sub-assimilation time windows.

[0065] It should be noted that each sub-assimilation time window should include sea state observation parameters. In practical applications, if a sub-assimilation time window does not have sea state observation parameters, it can be merged with the adjacent sub-assimilation time window.

[0066] Step S1112: Use the three-dimensional variational assimilation method to assimilate the sea state observation parameters of the i-th sub-assimilation time window to obtain the simulated sea state at the initial moment of the i-th sub-assimilation time window.

[0067] In this embodiment of the disclosure, i = 0, 1, 2, ..., N-1. In the specific implementation process, the computing device first sets i to 0, that is, starting from the 0th sub-assimilation time window, it uses the three-dimensional variational assimilation method to assimilate the sea state observation parameters included therein, so as to obtain the simulated sea state at the initial moment of the 0th sub-assimilation time window.

[0068] Similarly, for each subsequent sub-assimilation time window, the same three-dimensional variational assimilation method is used to assimilate the sea state observation parameters included in this sub-assimilation time window to obtain the simulated sea state at the initial moment of each sub-assimilation time window.

[0069] When performing three-dimensional variational assimilation for a specific sub-assimilation time window, the computing device minimizes the objective function corresponding to this sub-assimilation time window, thereby correcting the sea state simulation model to obtain the corrected model. The objective function is J = J0 B +J O , , Where x is the prediction parameter matrix constructed based on the sea state prediction parameters determined by the sea state simulation model before correction. b Let B be the background parameter matrix constructed based on the sea state prediction parameters, and O be the observation error covariance matrix. o H represents the observation parameter matrix constructed based on sea state observation parameters, and H is the observation operator.

[0070] Step S1113: Perform numerical simulation based on the sea state simulation state at the initial moment of the i-th sub-assimilation time window to obtain the sea state simulation state at the beginning moment of the (i+1)-th sub-assimilation time window.

[0071] After obtaining the sea state simulation state at the initial moment of the i-th sub-assimilation time window, numerical simulation is performed based on the sea state simulation state at the initial moment to obtain the sea state simulation state at the beginning moment of the (i+1)-th sub-assimilation time window.

[0072] In practical applications, if the i-th sub-assimilation time window and the (i+1)-th sub-assimilation time window overlap, the numerical simulation can be performed up to the start of the overlap, which will give us the simulated sea state at the start of the (i+1)-th sub-assimilation time window.

[0073] If the i-th sub-assimilation time window does not overlap with the (i+1)-th sub-assimilation time window, then after the numerical simulation is completed in the i-th sub-assimilation time window, the numerical simulation continues until the start of the (i+1)-th sub-assimilation time window in order to obtain the simulated sea state at the start of the (i+1)-th sub-assimilation time window.

[0074] Step S1114: Determine if i is less than N-1; if not, proceed to step S1115; if yes, make i = i+1, and re-execute steps S1112-S1113.

[0075] Step S1115: Calculate the sea state simulation state at the end of the N-1th sub-assimilation time window based on the sea state simulation state at the start of the N-1th sub-assimilation time window, and determine the assimilated sea state simulation model based on the sea state simulation state at the end of the N-1th sub-assimilation time window.

[0076] By employing the aforementioned steps S1111-S1115, the assimilation time window is divided into N sub-assimilation time windows, and the sea state simulation model is corrected for the sea state observation parameters in each of the N sub-assimilation time windows, thus progressively refining the sea state simulation model. Because progressively correcting the sea state simulation model using N sub-assimilation time windows ensures that the corrected simulation model within each sub-assimilation time window better matches the characteristics of the sea state observation parameters, the final corrected sea state simulation model is more accurate. Furthermore, the assimilation method provided in this embodiment effectively eliminates the spin-up phenomenon.

[0077] In some embodiments of this disclosure, the use of ensemble Kalman filtering assimilation method to assimilate the sea state observation parameters of the time window to be assimilated to obtain the sea state observation parameters of the time window to be assimilated may include steps S1121-S1124.

[0078] Step S1121: Based on the sea state simulation model before assimilation, randomly determine the sea state simulation state at multiple randomly selected time points within the current assimilation window, and obtain the sea state climate state at the corresponding time points.

[0079] In this embodiment of the disclosure, the current assimilation time window is ( After determining the current assimilation time window, N time points are randomly selected within the current assimilation time window, namely t1, t2, ..., t... N Subsequently, based on the initial time of the current assimilation time window, that is... Numerical simulations of the simulated sea state at various times yield simulated sea state states M1, M2, L, M at each time point. N .

[0080] Simultaneously, the sea state and climate state C1, C2, L, C corresponding to each time point can be obtained. N .

[0081] Step S1122: Determine the corresponding disturbance matrix based on multiple time points, the corresponding simulated sea state, and the sea state climatology.

[0082] In some embodiments of this disclosure, determining the corresponding disturbance matrix based on multiple time points, corresponding simulated sea state and sea state climatology may include steps S11221-S11222.

[0083] Step S11221: Calculate the time interval from each time point to the initial time of the current assimilation time window, and the state difference between the simulated sea state and the sea state-climate state at each time point.

[0084] Specifically, the time interval from each time point to the initial time of the current assimilation window can be calculated using... Calculated.

[0085] The difference between the simulated sea state and the sea state climatological state at various time points can be calculated using ΔM. n =M n -C n n = 1, 2, ..., N.

[0086] Step S322: Determine the weighted average matrix of state differences based on the time interval, the corresponding state differences, and the length of the current assimilation time window;

[0087] In this embodiment of the disclosure, the following methods can be used: Determine the weighted average matrix of state differences

[0088] Step S11222: Determine the corresponding perturbation matrix based on the weighted average matrix and the differences between each state.

[0089] In this embodiment of the disclosure, when determining the weighted average matrix After that, it can be adopted Determine the perturbation matrix corresponding to each state difference, where K is the weighting coefficient.

[0090] Step S1123: Determine the simulated sea state at the end of the current assimilation time window based on the simulated sea state model before assimilation, and obtain the observed sea state parameters and observation error covariance matrix at the end of the current assimilation time window.

[0091] Step S1124: Based on the sea state simulation state at the end of the current assimilation time window, the disturbance matrix, and the observation error covariance matrix of the observed sea state parameters at the end of the current assimilation window, determine the filtered sea state simulation state, and determine the assimilated sea state simulation model based on the filtered sea state simulation state.

[0092] Based on the simulated sea state before assimilation, the simulated sea state at the end of the current assimilation time window is determined to be A. j+1,n The observed sea state parameter at the end of the current assimilation window is O. j+1 The observation error covariance matrix is ​​R j+1 Then the simulated sea state at the end of the current assimilation time window after filtering is determined as follows: in H represents the filtered simulated sea state, and H is the observation transformation matrix.

[0093] The assimilation method provided in this disclosure can be used to assimilate the set of state vectors that estimate the background error covariance from the numerical simulation results of a single sea state simulation and the sea state climatology as background error, thereby reducing the amount of data processing and ensuring good simulation results.

[0094] The method of using adaptive optimal interpolation assimilation to assimilate sea state observation parameters within the assimilation time window includes steps S1131-S1136.

[0095] Step S1131: Calculate the sea state prediction parameters for each point to be predicted at the j-th observation time within the assimilation time window, j = 1, 2, ..., M, where M is the number of observation time points included in the assimilation time window.

[0096] Step S1132: Obtain the sea state observation parameters of each observation point at the j-th observation time within the time window to be assimilated, and determine the sea state observation parameters of the preset number of observation points closest to the point to be predicted.

[0097] In this embodiment of the disclosure, the computing device can employ an adaptive query method to determine a preset number of nearest observation points and obtain the sea state observation parameters of the preset number of observation points at the corresponding time. Specifically, the computing device can determine the preset number of nearest grid points by setting a virtual circle, gradually expanding the radius of the virtual circle, and performing a search.

[0098] Step S1133: Calculate the forecast error covariance matrix and the observation error covariance matrix based on the sea state prediction parameters of each point to be predicted and the sea state observation parameters of the most recent preset number of observation points.

[0099] Step S1134: Calculate the weight matrix based on the forecast error covariance matrix and the observation error covariance matrix.

[0100] Step S1135: Using the sea state prediction parameters, sea state observation parameters, and weight matrix, calculate the sea state simulation state corrected at the j-th observation time.

[0101] In this embodiment of the disclosure, the prediction error covariance matrix adopts P b Let R be the observation error covariance matrix, and P be the analysis error covariance matrix. a It means that P a =P b -KHP b .

[0102] After obtaining the prediction error covariance matrix, P is used. b After obtaining the observation error covariance matrix R, the weight matrix K = P can be obtained. b H T HP b H T +R) -1 Then we can obtain X. a -X b =K(Y o -HX b ), where X a The corrected simulated sea state at time i.

[0103] Step S1136: Determine if j is less than M. If yes, set j = j + 1 and re-execute step S1132. If no, execute step S1137.

[0104] Step S1137: Construct the assimilated sea state simulation model based on the sea state simulation state obtained at the last observation time.

[0105] By employing the aforementioned steps S1131-S1137, the simulated sea state model is corrected based on the sea state observation parameters of a preset number of nearby observation points, ensuring that the sea state simulation model can better simulate the actual sea state and guaranteeing the accuracy of the prediction.

[0106] After obtaining the sea state simulation model, subsequent steps can be performed.

[0107] Step S120: Perform numerical simulation based on the sea state simulation model to determine the sea state characteristic parameters of each preset grid point at the i-th preset time in the period to be predicted, i = 1, 2, ..., N.

[0108] The predicted grid points are determined grids with relatively fixed locations within the sea area to be predicted. These grid points are combined to form a grid, preferably a triangular grid. In other words, the predicted grid points are preferentially triangular grid points.

[0109] In this embodiment of the disclosure, numerical simulation is performed based on a sea state simulation model to determine the sea state characteristic parameters of each grid point at each preset time point in the period to be predicted. This is done by determining the sea state characteristic parameters of each grid point at the initial time of the period to be predicted based on the sea state simulation model, and then performing simulation according to a preset dynamic simulation method based on the sea state characteristic parameters at the initial time to obtain the sea state prediction parameters at each subsequent time point in the period to be predicted.

[0110] Step S130: Based on the location of the floating object to be predicted at the j-th preset time, the location of each preset grid point, and the sea state characteristic parameters at the j-th preset time, determine the estimated sea state characteristic parameters of the location of the floating object to be predicted at the j-th preset time, j = 1, 2, ..., N-1.

[0111] In this embodiment of the disclosure, the position corresponding to j=1 is known.

[0112] Specifically, based on the location of the floating object to be predicted at the j-th preset time, the estimated sea state characteristic parameters for the j-th preset time are determined by identifying neighboring grid points among all preset grid points based on the location at the j-th preset time, and then finding the sea state characteristic parameters of these neighboring grid points at the j-th preset time. Subsequently, based on the locations of the neighboring grid points, the sea state characteristic parameters at the j-th preset time, and the location of the floating object to be predicted at the j-th preset time, the estimated sea state characteristic parameters for the floating object at its current location are determined.

[0113] Assuming the preset grid points are triangular grid points, and the location coordinates of the floating object to be predicted are (x, y), and the location coordinates of the three adjacent grid points are (x1, y1), (x2, y2), and (x3, y3), then based on the coordinates of the four points, the weight coefficients for the corresponding three grid points can be a, b, and c, respectively. The estimated sea state characteristic parameters at the location of the floating object to be predicted are z = a × z1 + b × z2 + c × z3, where z1, z2 and z3 are the sea state characteristic parameters of the three grid points mentioned above.

[0114] Step S140: Calculate the predicted drift velocity of the floating object to be predicted based on the estimated sea state characteristic parameters corresponding to the j-th preset time, and calculate the position of the floating object to be predicted at the (j+1)-th preset time based on the predicted drift velocity and the position at the j-th prediction time.

[0115] After determining the estimated sea state parameters corresponding to the j-th preset time, the predicted drift velocity of the floating object to be predicted can be calculated according to the estimated sea state parameters and the preset calculation method.

[0116] In some embodiments of this disclosure, the estimated sea state characteristic parameters include surface current characteristic parameters, wind characteristic parameters, and wave characteristic parameters.

[0117] Correspondingly, step S140, which calculates the predicted drift velocity of the floating object based on the estimated sea state parameters, includes steps S141-S142.

[0118] Step S141: Determine the flow-induced drift velocity based on surface water flow characteristic parameters, determine the wind-induced drift velocity based on wind characteristic parameters, and determine the wave-induced drift velocity based on wave characteristic parameters.

[0119] Surface flow characteristic parameters are parameters that characterize the fluidity of surface water. These parameters can include the flow velocity and direction. The characteristic parameters representing the location of the floating object to be predicted are obtained through numerical simulation using a hydrodynamic sub-model.

[0120] Flow-induced drift velocity is the velocity acquired by the floating object due to the action of surface water flow. The direction of the flow-induced drift velocity is the same as the direction of the surface water flow, and the magnitude of the flow-induced drift velocity is directly proportional to the velocity of the surface water flow.

[0121] In this embodiment of the disclosure, the computing device can calculate the flow-induced drift velocity using a pre-set flow-induced drift coefficient and the surface water velocity. Specifically, it can use... The flow-induced drift velocity was calculated. Where λ c The flow-induced drift coefficient, This represents the velocity of the surface water flow. The flow-induced drift coefficient λ c The selection is based on the characteristics of the floating object to be predicted and the water flow characteristics of the water area.

[0122] Wind characteristic parameters are used to characterize the wind features above the water surface. These parameters can include wind speed and direction. In practical applications, wind characteristic parameters can be the wind speed and direction at a height of 10m above the water surface. The wind characteristic parameters at the location of the floating object to be predicted are obtained through numerical simulation using an atmospheric dynamic submodel.

[0123] Wind-induced drift velocity is the speed at which a floating object, to be predicted, drifts due to the action of wind on the water surface. Wind-induced drift velocity includes drift velocity along the wind direction and drift velocity perpendicular to the wind direction. In this embodiment, the wind-induced drift velocity can be calculated based on the wind-induced drift coefficient and wind characteristic parameters. The wind-induced drift coefficient is selected based on the characteristics of the floating object and the wind characteristics of the water area.

[0124] Wave characteristic parameters are parameters used to characterize the wave characteristics of a water body surface. Wave characteristic parameters can include at least one of the wave height, wave period, and wave direction. Wave characteristic parameters are obtained through numerical simulation using a wave sub-model.

[0125] Wave-induced drift speed is the drift speed acquired by a floating object due to the action of ocean waves. In some embodiments of this disclosure, wave-induced drift speed is calculated based on wave feature parameters. This can be achieved by inputting at least one parameter from the estimated sea state feature parameters into a pre-trained deep learning model of wave-induced drift to obtain the wave-induced drift speed.

[0126] The wave-induced drift deep learning model is a deep learning model used to predict wave-induced drift velocity based on estimated sea state characteristic parameters. In this embodiment, the wave-induced drift deep learning model is trained based on first sample drift data. The first sample drift data includes the sample drift velocity observations of sample floating objects and the corresponding sample surface current characteristic parameters, sample wind characteristic parameters, and sample wave characteristic parameters. How the wave-induced drift deep learning model is trained based on the first sample drift data will be analyzed later.

[0127] Step S142: Determine the predicted drift velocity based on the flow-induced drift velocity, wind-induced drift velocity, and wave-induced drift velocity.

[0128] After obtaining the flow-induced drift velocity, wind-induced drift velocity, and wave-induced drift velocity, the flow-induced drift velocity, wind-induced drift velocity, and wave-induced drift velocity are added together to calculate the predicted drift velocity of the floating object to be predicted.

[0129] In this embodiment of the disclosure, after obtaining the predicted drift velocity, an integral budget can be performed based on the predicted drift velocity to determine the drift distance from the j-th prediction time to the (j+1)-th prediction time. After determining the drift distance, combined with the position of the object to be predicted at the j-th prediction time, the position of the object to be predicted at the (j+1)-th prediction time can be obtained. Specifically, the position at the (j+1)-th time is s. j+1 =s j +v·(t s+1 -t s ), where s j Let v be the position of the floating object to be predicted at the j-th prediction time, v be the predicted drift velocity, and t be the position of the floating object to be predicted at the j-th prediction time.s+1 For the (j+1)th prediction time, t s Let j be the predicted time.

[0130] In practical applications of the embodiments of this disclosure, the aforementioned steps S130 and S140 are executed cyclically until the location of all preset floating objects to be predicted is determined.

[0131] Step S150: Determine the drift trajectory of the object to be predicted based on its position at each preset time.

[0132] After determining the positions of the floating objects to be predicted at all preset times, a line is then drawn sequentially based on the positions at each preset time to determine the drift trajectory of the floating objects. In this embodiment of the disclosure, the connecting lines are used as the drift trajectory of the floating objects.

[0133] The method provided in this disclosure first assimilates the observed sea state parameters within the assimilation time window to obtain an assimilated sea state simulation model. Then, based on the assimilated sea state simulation model, the sea state characteristic parameters for each preset time point are determined. Next, the sea state characteristic parameters for the corresponding position of the floating object to be predicted at the j-th preset time point are determined iteratively, and the predicted drift velocity at that position is determined. The position at the (j+1)-th preset time point is then calculated based on the predicted drift velocity and the current position, until the positions for all preset time points are calculated. After determining the positions at the preset time points, the drift trajectory of the floating object to be predicted can be determined.

[0134] Using the method provided in this disclosure, the predicted drift position and location of the floating object to be predicted can be determined by iterative calculation, and then the drift trajectory of the floating object to be predicted can be calculated, thereby achieving a more accurate prediction of the drift trajectory.

[0135] As mentioned earlier, in some embodiments of this disclosure, the wave-induced drift deep learning model is trained based on the first sample data.

[0136] Figure 2 This is a flowchart illustrating a method for training a wave-induced drift deep learning model based on first sample data, as described in this disclosure. Figure 2 As shown, the method for training a wave-induced drift deep learning model based on the first sample data includes steps S210-S250.

[0137] Step S210: Obtain the flow-induced drift coefficient and the wind-induced drift coefficient.

[0138] In some embodiments of this disclosure, the computing device may directly use the flow-induced drift coefficient and wind-induced drift coefficient provided by various other technical documents or related software products as the flow-induced drift coefficient and wind-induced drift coefficient.

[0139] Step S220: Based on the surface water flow characteristic parameters and flow-induced drift coefficient of the sample, calculate the corresponding estimated value of the sample flow-induced drift velocity.

[0140] Step S230: Calculate the corresponding estimated value of sample wind-induced drift velocity based on sample wind characteristic parameters and wind-induced drift coefficient.

[0141] The estimated sample flow-induced drift velocity is calculated by multiplying the flow-induced drift coefficient and the sample surface flow velocity based on the sample surface flow characteristic parameters and the flow-induced drift coefficient. Similarly, the estimated sample wind-induced drift velocity is calculated by multiplying the wind-induced drift coefficient and the sample wind characteristic parameters.

[0142] Step S240: Based on the sample drift velocity observation, the corresponding sample flow-induced drift velocity estimate and the sample wind-induced drift velocity estimate, calculate the corresponding sample wave-induced drift velocity estimate.

[0143] After obtaining the estimated values ​​of the sample flow-induced drift velocity and the sample wind-induced drift velocity, the estimated values ​​of the sample flow-induced drift velocity and the estimated values ​​of the sample wind-induced drift velocity are subtracted from the sample drift velocity observation values ​​to obtain the estimated value of the sample wave-induced drift velocity.

[0144] Step S250: Train a deep learning model for wave-induced drift using sample wave feature parameters and corresponding sample wave-induced drift velocity estimates.

[0145] After obtaining the estimated sample wave-induced drift velocity, the sample wave parameters and the corresponding estimated sample wave-induced drift velocity are then used as training samples to train the wave-induced drift velocity, thus obtaining the wave-induced drift deep learning model.

[0146] The wave drift deep learning model can be any of the possible models known in the art, such as a BP neural network model, a recurrent neural network model, etc., and the embodiments disclosed herein are not particularly limited.

[0147] Specifically, in order to train the final optimal wave-drift deep learning model, multiple pre-built deep learning models can be established in this embodiment of the disclosure, and each neural network model can be trained separately. The quality of the training results determines which model to sample as the final wave-drift deep learning model.

[0148] In the previous embodiments, the wind-induced drift coefficient and the flow-induced drift coefficient were directly adopted from the coefficients provided by various other technical documents or related software products.

[0149] In some embodiments of this disclosure, the wind-induced drift coefficient and the flow-induced drift coefficient can also be determined during the training of the wave-induced drift deep learning model. Determining the wind-induced drift coefficient and the flow-induced drift coefficient during the training of the wave-induced drift deep learning model involves fitting the observed sample drift velocities as approximations of the sum of the flow-induced drift velocities and the wind-induced drift velocities to the flow-induced drift coefficients and the wind-induced drift coefficients.

[0150] Specifically, formulas can be constructed. An approximate relationship is established between the observed sample drift velocity, the sum of the flow-induced drift velocity and the wind-induced drift velocity. Wherein, V ox-s and V oy-s These are the components of the sample drift velocity observations along the x and y axes (the x-axis can be the latitude direction, and the y-axis can be the longitude direction), V F-surface-current-sx and V F-surface-current-sy L represents the components of the flow-induced drift velocity of the sample floating object on the x and y axes. sx and L sy The wind-induced drift velocity of the sample floating object is represented by the components on the x and y axes.

[0151] Based on the relationship between flow-induced drift velocity and surface water flow characteristic parameters, the formula can be used. Determine V F-surface-current-sx and V F-surface-current-sy , where λ c V is the flow-induced drift coefficient. csx and V csy These are the components of the surface flow velocity on the x-axis and y-axis, respectively.

[0152] Based on the relationship between wind-induced drift velocity and wind characteristic parameters, the formula can be used. Calculate and determine L sx and L sy ,in Among them W 10mwind-s For sample wind speed, The wind-induced drift speed along the wind direction, a is the wind-induced drift velocity perpendicular to the wind direction. d b d a c b c ε is the wind-induced drift coefficient. d Let ε be the wind-induced random fluctuation coefficient along the wind direction. c This represents the random fluctuation coefficient when the wind direction is deflected to the left and right.

[0153] By fitting the aforementioned formula to the first sample of drift data, the flow-induced drift coefficient λ can be determined. c And wind-induced drift coefficient a d b d ac b c ε d and ε c Specifically, the least squares method can be used for parameter fitting to obtain the aforementioned flow-induced drift coefficient and wind-induced drift coefficient.

[0154] In specific implementations of this disclosure, multiple sets of first sample drift data may be included. Based on each set of first sample drift data, a set of flow-induced drift coefficients λ can be obtained. c And wind-induced drift coefficient a d b d a c b c ε d and ε c After obtaining multiple sets of the aforementioned drift coefficients, the mean and standard deviation of each drift coefficient are calculated, and data meeting preset conditions are selected based on the mean and standard deviation. Finally, the flow-induced drift coefficient λ is calculated based on the data meeting the preset conditions. c And wind-induced drift coefficient a d b d a c b c ε d and ε c .

[0155] In some embodiments of this disclosure, in order to ensure that the calculated flow-induced drift velocity and wind-induced drift velocity reflect the random characteristics of force and velocity changes during the actual drift of the object to be predicted, the following steps can be performed before calculating the flow-induced drift velocity based on the aforementioned flow-induced drift coefficient and surface water velocity: adding a random perturbation to the flow-induced drift coefficient to obtain a perturbed flow-induced drift coefficient. By adding a random perturbation to the flow-induced drift coefficient, the calculated flow-induced drift velocity also becomes random. In specific implementations, the perturbation range for adding random perturbation to the flow-induced drift coefficient is determined in advance, specifically during the calculation of the flow-induced drift coefficient.

[0156] Similarly, before performing the aforementioned calculation of wind-induced drift velocity based on wind characteristic parameters and wind-induced drift coefficient, the following steps can be performed: Add a random perturbation to the wind-induced drift coefficient to obtain a perturbed wind-induced drift coefficient. By adding a random perturbation to the wind-induced drift coefficient, the calculated wind-induced drift velocity becomes random. In specific implementation, the range of adding a random perturbation to the wind-induced drift coefficient can be determined according to ε. d and ε c Sure.

[0157] Optionally, in some embodiments of this disclosure, steps S310-S320 may be performed before calculating the wind-induced drift velocity based on wind characteristic parameters and wind-induced drift coefficient.

[0158] Step S310: Determine the predicted drift bias based on the estimated sea state characteristic parameters of the floating object at the location of the object to be predicted at the time of prediction.

[0159] In some embodiments of this disclosure, determining the drift bias of a floating object to be predicted may involve inputting estimated sea state characteristic parameters into a pre-trained deep learning model of drift bias to determine the predicted drift bias.

[0160] The aforementioned drift bias deep learning model is trained based on second sample drift data, which includes drift bias observations of sample floating objects and corresponding estimated sea state characteristic parameters.

[0161] In practical applications, various types of floating objects can be sampled in specific aquatic environments to conduct floating object drift experiments. The drift trajectory of the floating objects can be tracked using tracking observation vessels and positioning buoys. The corresponding surface water flow characteristic parameters, wind characteristic parameters, and wave characteristic parameters of the sample can be obtained using equipment such as acoustic Doppler current profilers, meteorological sensors, and wave sensors carried on the tracking observation vessels.

[0162] After obtaining a large amount of sample floating object drift experimental data, the data can be analyzed to determine the observed drift bias of the sample floating objects. In practical applications, the observed drift bias of the sample floating objects can be determined by subtracting the surface water velocity from the observed drift velocity of the sample floating objects to roughly obtain the wind-induced drift velocity. Then, the drift bias of the sample floating objects can be determined using the direction of the wind-induced drift velocity and the wind direction. Subsequently, a second sample drift data set can be constructed using the observed drift bias of the sample floating objects, along with the corresponding surface water flow characteristic parameters, wind characteristic parameters, and wave characteristic parameters.

[0163] After obtaining the second sample drift data, the pre-built biased deep learning model can be trained using the second sample drift data to obtain a trained biased deep learning model. In the application of this disclosure embodiment, the biased deep learning model can be various possible models, such as widely used neural network models like the BP neural network model.

[0164] Step S320: Select the corresponding wind-induced drift coefficient based on the predicted drift bias.

[0165] In this embodiment of the disclosure, different predicted drift biases correspond to different wind-induced drift coefficients, specifically different wind-induced drift coefficients perpendicular to the wind direction. For example, when the bias is left-leaning, the aforementioned wind-induced drift coefficient a... c b c ε c a c+ b c+ ε c+ However, when the bias is right-leaning, the aforementioned wind-induced drift coefficient a c b c ε c a c- b c- and ε c- .

[0166] It should be noted that when selecting the corresponding wind-induced drift coefficient based on the predicted drift bias, the aforementioned formula is applied to fit the flow-induced drift coefficient and the wind-induced drift coefficient. Specifically Where δ is 1, it indicates that the drift deflection is perpendicular to the wind direction and deflects to the left, that is, when the drift deflection is to the left... To be used for fitting calculation a c+ b c+ ε c+ When δ is 0, it indicates that the drift deflection is to the right of the vertical wind direction. Used for fitting calculation a c- b c- and ε c- .

[0167] In addition to providing the aforementioned method for predicting the drift trajectory of floating objects, this disclosure also provides a device 300 for predicting the drift trajectory of floating objects.

[0168] Figure 3 This is a schematic diagram of the structure of the drift trajectory prediction device for floating objects provided in an embodiment of this disclosure. Figure 3 As shown, the drift trajectory prediction device 300 for floating objects includes a simulated sea state model construction unit 301, a numerical simulation unit 302, a sea state characteristic parameter estimation unit 303, a position determination unit 304, and a trajectory determination unit 305.

[0169] The simulated sea state model building unit 301 is used to assimilate the sea state observation parameters within the assimilation time window to obtain the assimilated sea state simulation model.

[0170] The numerical simulation unit 302 is used to perform numerical simulation based on the sea state simulation model to determine the sea state characteristic parameters of each preset grid point at the i-th preset time in the period to be predicted, i = 1, 2, ..., N.

[0171] The sea state characteristic parameter determination unit 303 is used to determine the estimated sea state characteristic parameters of the floating object to be predicted at the j-th preset time based on the position of the floating object to be predicted at the j-th preset time, the position of each preset grid point and the sea state characteristic parameters at the j-th preset time, j = 1, 2, ..., N-1, where the position corresponding to j = 1 is known.

[0172] The position determination unit 304 is used to calculate the predicted drift velocity of the floating object to be predicted based on the estimated sea state characteristic parameters corresponding to the j-th preset time, and to calculate the position of the floating object to be predicted at the (j+1)-th preset time based on the predicted drift velocity.

[0173] The trajectory determination unit 305 is used to determine the drift trajectory of the floating object based on its position at each preset time.

[0174] In some embodiments of this disclosure, the estimated sea state characteristic parameters include surface current characteristic parameters, wind characteristic parameters, and wave characteristic parameters. Correspondingly, the position calculation unit includes a sub-velocity calculation sub-unit and a velocity and calculation sub-unit. The sub-velocity calculation sub-unit is used to determine the current-induced drift velocity based on the surface current characteristic parameters, the wind-induced drift velocity based on the wind characteristic parameters, and the wave-induced drift velocity based on the wave characteristic parameters. The velocity and calculation sub-unit is used to determine the predicted drift velocity based on the current-induced drift velocity, the wind-induced drift velocity, and the wave-induced drift velocity.

[0175] In some embodiments of this disclosure, the sub-velocity calculation sub-unit inputs wave feature parameters into a pre-trained wave-induced drift deep learning model to obtain the wave-induced drift velocity; wherein, the wave-induced drift deep learning model is trained based on first sample drift data, the first sample drift data including sample drift velocity observations of sample floating objects and corresponding sample surface water flow feature parameters, sample wind feature parameters and sample wave feature parameters.

[0176] In some embodiments of this disclosure, the sub-velocity calculation sub-unit calculates the flow-induced drift velocity based on surface water flow characteristic parameters and flow-induced drift coefficient, and calculates the wind-induced drift velocity based on wind characteristic parameters and wind-induced drift coefficient; wherein the flow-induced drift coefficient and wind-induced drift coefficient are obtained during the training of the wave-induced drift deep learning model.

[0177] In some embodiments of this disclosure, the drift trajectory prediction device 300 for floating objects further includes a disturbance addition unit, which is used to add random disturbances to the flow-induced drift coefficient to obtain a disturbed flow-induced drift coefficient; and to add random disturbances to the wind-induced drift coefficient to obtain a disturbed wind-induced drift coefficient.

[0178] In some embodiments of this disclosure, the drift trajectory prediction device 300 for floating objects further includes a wind-induced drift coefficient determination unit. The wind-induced drift coefficient determination unit is used to determine the predicted drift bias based on the estimated sea state characteristic parameters corresponding to the floating object at a preset time j; and to select the corresponding wind-induced drift coefficient based on the predicted drift bias.

[0179] In some embodiments of this disclosure, the wind-induced drift coefficient determination unit inputs estimated sea state characteristic parameters into a pre-trained drift bias deep learning model to determine the predicted drift bias of the floating object; wherein, the bias deep learning model is trained based on second sample drift data, the second sample drift data including the drift bias observations of the sample floating object and the sample sea state characteristic parameters.

[0180] Figure 4 This is a schematic diagram of the structure of the terminal device provided in the embodiments of this disclosure. For example... Figure 4 As shown, the computing device 400 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 401, which can perform various appropriate actions and processes based on a program stored in a read-only memory (ROM) 402 or a program loaded from a storage device 408 into a random access memory (RAM) 403. The RAM 403 also stores various programs and data required for the operation of the computing device 400. The processing unit 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0181] Typically, the following devices can be connected to I / O interface 405: input devices 406 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 407 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 408 including, for example, magnetic tapes, hard disks, etc.; and communication devices 409. Communication device 409 allows computing device 400 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 4 A computing device 400 with various devices is shown, but it should be understood that it is not required to implement or have all of the devices shown. More or fewer devices may be implemented or have alternatively.

[0182] In particular, based on embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 409, or installed from storage device 408, or installed from ROM 402. When the computer program is executed by processing device 401, it performs the functions defined in the methods of embodiments of this disclosure.

[0183] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0184] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0185] The aforementioned computer-readable medium may be included in the aforementioned computing device; or it may exist independently and not assembled into the computing device.

[0186] The aforementioned computer-readable medium carries one or more programs. When the aforementioned one or more programs are executed by the computing device, the computing device causes the following: to assimilate the sea state observation parameters within the assimilation time window to obtain an assimilated sea state simulation model; to perform numerical simulation based on the sea state simulation model to determine the sea state characteristic parameters of each preset grid point at the i-th preset time within the prediction period, i = 1, 2, ..., N; to determine the estimated sea state characteristic parameters of the position of the floating object to be predicted at the j-th preset time based on the position of the floating object to be predicted at the j-th preset time, the positions of each preset grid point, and the sea state characteristic parameters at the j-th preset time, j = 1, 2, ..., N-1, where the position corresponding to j = 1 is known; to calculate the predicted drift velocity of the floating object to be predicted based on the estimated sea state characteristic parameters corresponding to the j-th preset time, and to calculate the position of the floating object to be predicted at the (j+1)-th preset time based on the predicted drift velocity and the position at the j-th prediction time; and to determine the drift trajectory of the floating object to be predicted based on the position of the floating object at each preset time.

[0187] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including but not limited to object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0188] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0189] The units described in the embodiments of this disclosure can be implemented in software or hardware. The names of the units are not, in some cases, intended to limit the specific unit.

[0190] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0191] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0192] This disclosure also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can perform the above-described functions. Figures 1-4 The methods in any of the embodiments are similar in execution and beneficial effects, and will not be described again here.

[0193] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0194] The above are merely specific embodiments of this disclosure, enabling those skilled in the art to understand or implement this disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not to be limited to these embodiments, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting the drift trajectory of a floating object, characterized in that, include: The sea state observation parameters within the assimilation time window are assimilated to obtain the assimilated sea state simulation model. Numerical simulations were performed based on the aforementioned sea state simulation model to determine the position of each preset grid point within the predicted time period. i Sea state characteristic parameters at a preset time, i=1,2,……,N ; Based on the floating object to be predicted in the first j The location at the preset time, the location of each preset grid point, and the position at the preset moment. j The sea state characteristic parameters at a preset time are used to determine the floating object to be predicted at the [number]th [time]. j Estimated sea state characteristic parameters for a given location at a preset time. j=1,2,...,N-1 ,in j=1 The location corresponding to the time is known; Based on the first j The predicted drift velocity of the object to be predicted is calculated based on the estimated sea state characteristic parameters corresponding to a preset time, and the predicted drift velocity is calculated based on the predicted drift velocity and the first preset time. j The position of the object to be predicted at the predicted time is calculated. j+1 The position at a preset time; The drift trajectory of the object to be predicted is determined based on its position at each preset time. The data assimilation process employs either ensemble Kalman filtering assimilation or adaptive optimal interpolation assimilation. The ensemble Kalman filter assimilation includes: randomly determining the sea state simulation state at multiple randomly selected time points within the current assimilation window based on the sea state simulation model before assimilation, and obtaining the sea state climatological state at the corresponding time points; The adaptive optimal interpolation assimilation includes: determining a preset number of nearest observation points through adaptive query, obtaining the sea state observation parameters of the preset number of observation points at the corresponding time; setting a virtual circle, gradually expanding the radius of the virtual circle and performing a search to determine the preset number of nearest grid points; The step of performing numerical simulation based on the sea state simulation model to determine the sea state characteristic parameters of each preset grid point at the i-th preset time in the period to be predicted involves determining the sea state characteristic parameters of each preset grid point at the initial time of the period to be predicted based on the sea state simulation model, and then performing simulation according to a preset dynamic simulation method based on the sea state characteristic parameters at the initial time to obtain the sea state prediction parameters for each subsequent time point in the period to be predicted. The steps of determining the estimated sea state characteristic parameters of the floating object at the j-th preset time based on the position of the floating object to be predicted at the j-th preset time, the positions of each preset grid point, and the sea state characteristic parameters at the j-th preset time, and the steps of calculating the predicted drift velocity of the floating object based on the estimated sea state characteristic parameters corresponding to the j-th preset time, and calculating the position of the floating object at the (j+1)-th preset time based on the predicted drift velocity and the position at the j-th preset time, are performed cyclically until the positions of the floating objects to be predicted at all preset times are determined; The estimated sea state characteristic parameters include wind characteristic parameters, and the calculation of the predicted drift velocity of the object to be predicted based on the estimated sea state characteristic parameters corresponding to the j-th preset time includes: According to the predicted floating object in the first j The estimated sea state characteristic parameters corresponding to each preset time point determine the predicted drift bias. Select the corresponding wind-induced drift coefficient based on the predicted drift bias; The wind-induced drift velocity is calculated based on the wind characteristic parameters and the wind-induced drift coefficient; wherein, the wind-induced drift velocity is used to calculate the predicted drift velocity of the floating object to be predicted; The prediction of the floating object in the first j The estimated sea state characteristic parameters corresponding to each preset time point determine the predicted drift bias, including: The estimated sea state characteristic parameters are input into a pre-trained drift bias deep learning model to determine the predicted drift bias of the floating object. The biased deep learning model is trained based on the second sample drift data, which includes the drift biased observations of the sample floating objects and the sample sea state characteristic parameters.

2. The method according to claim 1, characterized in that, The estimated sea state characteristic parameters also include surface current characteristic parameters and wave characteristic parameters; Based on the first j The calculation of the predicted drift velocity of the floating object to be predicted, based on the estimated sea state characteristic parameters corresponding to a preset time, further includes: The flow-induced drift velocity is determined based on the surface water flow characteristic parameters, and the wave-induced drift velocity is determined based on the wave characteristic parameters. The predicted drift velocity is determined based on the flow-induced drift velocity, the wind-induced drift velocity, and the wave-induced drift velocity.

3. The method according to claim 2, characterized in that, The determination of wave-induced drift velocity based on wave characteristic parameters includes: The wave feature parameters are input into a pre-trained wave-induced drift deep learning model to obtain the wave-induced drift velocity. The wave-induced drift deep learning model is trained based on the first sample drift data, which includes the sample drift velocity observation value of the sample floating object and the corresponding sample surface water flow characteristic parameters, sample wind characteristic parameters and sample wave characteristic parameters.

4. The method according to claim 3, characterized in that, The determination of the flow-induced drift velocity based on the surface water flow characteristic parameters includes: calculating the flow-induced drift velocity according to the surface water flow characteristic parameters and the flow-induced drift coefficient; The flow-induced drift coefficient and the wind-induced drift coefficient are obtained during the training of the wave-induced drift deep learning model.

5. The method according to claim 4, characterized in that, Before calculating the flow-induced drift velocity based on the surface water flow characteristic parameters and the flow-induced drift coefficient, the method further includes: adding a random perturbation to the flow-induced drift coefficient to obtain the perturbed flow-induced drift coefficient; Before calculating the wind-induced drift velocity based on the wind characteristic parameters and the wind-induced drift coefficient, the method further includes: adding a random perturbation to the wind-induced drift coefficient to obtain the perturbed wind-induced drift coefficient.

6. A device for predicting the drift trajectory of a floating object, characterized in that, include: The simulated sea state model building unit is used to assimilate the sea state observation parameters within the assimilation time window to obtain the assimilated sea state simulation model. The numerical simulation unit is used to perform numerical simulations based on the sea state simulation model to determine the position of each preset grid point within the predicted time period. i Sea state characteristic parameters at a preset time, i=1,2,……,N ; The estimated sea state characteristic parameter determination unit is used to determine the floating object to be predicted based on the first sea state characteristic parameter in the first sea state. j The location at the preset time, the location of each preset grid point, and the position at the preset moment. j The sea state characteristic parameters at a preset time are used to determine the floating object to be predicted at the [number]th [time]. j Estimated sea state characteristic parameters for a given location at a preset time. j=1,2,...,N-1 ,in j=1 The location corresponding to the time is known; Position determination unit, used for determining position based on the first j The predicted drift velocity of the object to be predicted is calculated based on the estimated sea state characteristic parameters corresponding to a preset time, and the predicted drift velocity of the object to be predicted is calculated based on the predicted drift velocity at the [presumably a specific time point]. j+1 The position at a preset time; The trajectory determination unit is used to determine the drift trajectory of the floating object based on its position at each preset time. The data assimilation process employs either ensemble Kalman filtering assimilation or adaptive optimal interpolation assimilation. The ensemble Kalman filter assimilation includes: randomly determining the sea state simulation state at multiple randomly selected time points within the current assimilation window based on the sea state simulation model before assimilation, and obtaining the sea state climatological state at the corresponding time points; The adaptive optimal interpolation assimilation includes: determining a preset number of nearest observation points through adaptive query, obtaining the sea state observation parameters of the preset number of observation points at the corresponding time; setting a virtual circle, gradually expanding the radius of the virtual circle and performing a search to determine the preset number of nearest grid points; The step of performing numerical simulation based on the sea state simulation model to determine the sea state characteristic parameters of each preset grid point at the i-th preset time in the period to be predicted involves determining the sea state characteristic parameters of each preset grid point at the initial time of the period to be predicted based on the sea state simulation model, and then performing simulation according to a preset dynamic simulation method based on the sea state characteristic parameters at the initial time to obtain the sea state prediction parameters for each subsequent time point in the period to be predicted. The steps of determining the estimated sea state characteristic parameters of the floating object at the j-th preset time based on the position of the floating object to be predicted at the j-th preset time, the positions of each preset grid point, and the sea state characteristic parameters at the j-th preset time, and the steps of calculating the predicted drift velocity of the floating object based on the estimated sea state characteristic parameters corresponding to the j-th preset time, and calculating the position of the floating object at the (j+1)-th preset time based on the predicted drift velocity and the position at the j-th preset time, are performed cyclically until the positions of the floating objects to be predicted at all preset times are determined; The estimated sea state characteristic parameters include wind characteristic parameters, and the calculation of the predicted drift velocity of the object to be predicted based on the estimated sea state characteristic parameters corresponding to the j-th preset time includes: According to the predicted floating object in the first j The estimated sea state characteristic parameters corresponding to each preset time point determine the predicted drift bias. Select the corresponding wind-induced drift coefficient based on the predicted drift bias; The wind-induced drift velocity is calculated based on the wind characteristic parameters and the wind-induced drift coefficient; wherein, the wind-induced drift velocity is used to calculate the predicted drift velocity of the floating object to be predicted; The prediction of the floating object in the first j The estimated sea state characteristic parameters corresponding to each preset time point determine the predicted drift bias, including: The estimated sea state characteristic parameters are input into a pre-trained drift bias deep learning model to determine the predicted drift bias of the floating object. The biased deep learning model is trained based on the second sample drift data, which includes the drift biased observations of the sample floating objects and the sample sea state characteristic parameters.

7. A terminal device, characterized in that, include: A memory and a processor, wherein the memory stores a computer program that, when executed by the processor, implements the method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The storage medium stores a computer program that, when executed by a processor, implements the method as described in any one of claims 1-5.

Images