Wave in-situ calibration method based on GNSS buoy

Abstract

The invention discloses a wave in-situ calibration method based on a GNSS buoy, and belongs to the field of wave in-situ calibration, and the method comprises the following steps: obtaining buoy deviation data based on the position and attitude data of the GNSS buoy under different sea conditions; constructing a deviation compensation model and a deviation sliding model based on the buoy deviation data; a deviation compensation result is obtained based on the deviation compensation model and the deviation sliding model, and the deviation compensation result comprises first deviation compensation and second deviation compensation; and fusing the deviation compensation result with the GNSS positioning data to obtain a wave parameter estimation result.

Description

Technical Field

[0001] The invention belongs to the technical field of wave in-situ calibration, and in particular relates to a wave in-situ calibration method based on a GNSS buoy. Background Art

[0002] When performing in-situ wave calibration based on GNSS buoys, the deviation of the buoy itself is a key technical issue. When the buoy floats on the sea, it is affected by complex sea conditions such as wind, waves, and ocean currents, and its position and attitude may deviate, resulting in the GNSS positioning data not being completely consistent with the actual wave conditions. This deviation is time-varying and nonlinear, and is closely related to the sea state. The deviation of the buoy can be decomposed into vector components in the horizontal and vertical directions, among which the deviation in the horizontal direction is more affected by the ocean current, while the deviation in the vertical direction has a stronger correlation with the waves. At the same time, the buoy deviation may also show a certain sliding characteristic, that is, the deviation shows a certain trend over time. This sliding characteristic is closely related to the changing process of the sea conditions and reflects the dynamic cumulative effect of the deviation. Therefore, how to accurately model and compensate for the buoy deviation is the key to improving the accuracy of wave in-situ calibration. Traditional deviation models often assume that the deviation is static or linear, which makes it difficult to fully characterize the buoy deviation characteristics under complex sea conditions, resulting in limited calibration accuracy. There is an urgent need to conduct in-depth research on the spatiotemporal characteristics and nonlinear mechanisms of buoy deviations, develop deviation modeling and compensation methods that adapt to dynamic sea conditions, and provide technical support for improving the accuracy of in-situ wave calibration. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a wave in-situ calibration method based on GNSS buoy to solve the problems existing in the above prior art.

[0004] To achieve the above object, the present invention provides a wave in-situ calibration method based on a GNSS buoy, comprising:

[0005] The buoy deviation data is obtained based on the position and attitude data of the GNSS buoy in different sea conditions;

[0006] Constructing a deviation compensation model and a deviation sliding model based on the buoy deviation data;

[0007] Obtaining a deviation compensation result based on the deviation compensation model and the deviation sliding model, wherein the deviation compensation result includes a first deviation compensation and a second deviation compensation;

[0008] The deviation compensation result is fused with the GNSS positioning data to obtain the wave parameter estimation result.

[0009] Optionally, a deviation compensation model is constructed based on the buoy deviation data to obtain a first deviation compensation, and the process of obtaining the first deviation compensation includes:

[0010] Obtain real-time location information and historical trajectory data of GNSS buoys;

[0011] Analyzing the horizontal displacement and vertical displacement of the GNSS buoy using a Kalman filter algorithm based on the real-time position information and historical trajectory data;

[0012] Calculate the horizontal component and the vertical component of the buoy deviation based on the horizontal displacement and the vertical displacement of the GNSS buoy;

[0013] Acquire multi-source ocean environment information, and perform feature extraction on the multi-source ocean environment information to obtain key environmental factors that affect buoy deviation;

[0014] Constructing a horizontal deviation compensation model for the horizontal component based on the ocean current factor among the key environmental factors;

[0015] Constructing a vertical deviation compensation model for the vertical component based on the wave factor in the key environmental factors;

[0016] A first deviation compensation is obtained based on the horizontal deviation compensation model and the vertical deviation compensation model.

[0017] Optionally, the process of constructing a bias sliding model includes:

[0018] Acquire buoy deviation data within a preset time range based on a sliding window;

[0019] The buoy deviation data is analyzed based on the feature extraction algorithm to obtain the feature vector of the buoy deviation sliding characteristics;

[0020] Input the characteristic vector into the deviation modeling module to obtain the cumulative change law and construct a deviation sliding model;

[0021] Calculating a deviation prediction value based on the cumulative change rule, and calculating a compensation value of the buoy deviation;

[0022] Based on the deviation sliding model, the compensation value and the current buoy deviation data are weightedly fused to obtain a second deviation compensation.

[0023] Optionally, the process of inputting the feature vector into the deviation modeling module to obtain the deviation sliding model further includes model optimization, and the process of model optimization includes:

[0024] Acquire historical deviation data under different sea conditions, analyze the historical deviation data, obtain the correlation strength between each sea condition factor and the deviation characteristic, and obtain a typical sea condition based on the correlation strength;

[0025] Constructing a deviation prediction model based on the typical sea conditions;

[0026] Calculate the comprehensive deviation prediction result under typical sea conditions based on the deviation prediction model;

[0027] The comprehensive deviation prediction result is introduced into the control system to optimize the second deviation compensation to obtain an optimized deviation sliding model.

[0028] Optionally, the process of fusing the deviation compensation result with the GNSS positioning data to obtain a wave parameter estimation result includes:

[0029] Obtain wave in-situ observation data and synchronized GNSS positioning data;

[0030] Constructing a wave in-situ calibration model based on the wave in-situ observation data;

[0031] Inputting the deviation compensation result into the wave in-situ calibration model to obtain the deviation compensation result of the wave in-situ observation data;

[0032] A particle filter algorithm is used to fuse the deviation compensation result of the wave in-situ observation data with the synchronized GNSS positioning data to obtain fused wave observation data;

[0033] The least square method is used to perform parameter estimation on the fused wave observation data to obtain a wave parameter estimation result.

[0034] Optionally, after performing parameter estimation on the fused wave observation data, the method further includes establishing a deviation compensation effect evaluation system by introducing external high-precision wave observation data;

[0035] The wave parameter estimation result is compared with the wave in-situ observation data to calculate the improvement of the wave in-situ calibration accuracy;

[0036] Setting a preset threshold based on the external high-precision wave observation data;

[0037] The wave in-situ calibration accuracy is determined by comparing the wave in-situ calibration accuracy improvement margin with the preset threshold.

[0038] Optionally, when the wave in-situ calibration accuracy improvement reaches a preset threshold, the wave parameter estimation result is output as the final wave in-situ calibration result;

[0039] When the improvement of the wave in-situ calibration accuracy does not reach a preset threshold, the wave in-situ observation data and GNSS positioning data are reacquired, and the wave in-situ calibration model is optimized until the preset threshold is met.

[0040] The present invention also provides a computer terminal device, comprising:

[0041] one or more processors;

[0042] A memory, coupled to the processor, for storing one or more programs;

[0043] When the one or more programs are executed by the one or more processors, the one or more processors implement a wave in-situ calibration method based on a GNSS buoy.

[0044] The present invention also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, a wave in-situ calibration method based on a GNSS buoy is implemented.

[0045] Compared with the prior art, the present invention has the following advantages and technical effects:

[0046] The present invention discloses a GNSS buoy deviation compensation method based on machine learning, which is used to improve the accuracy of wave in-situ calibration. The method analyzes the position and attitude data of the buoy under different sea conditions and establishes a nonlinear mapping model to describe the spatiotemporal characteristics of the buoy deviation. Deviation compensation models are constructed for the horizontal and vertical directions respectively, taking into account the influence of ocean currents and wave factors. Sliding window technology and adaptive weight adjustment strategy are used to improve the timeliness and robustness of deviation compensation. The prediction results of multiple algorithms are integrated through an integrated learning method to form a more reliable deviation compensation model. Finally, the deviation compensation results are fused with the GNSS positioning data to generate high-precision wave parameter estimates, and an evaluation system is established to continuously optimize the compensation effect, thereby significantly improving the accuracy of wave in-situ calibration. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0048] Figure 1 This is a flow chart of a wave in-situ calibration method based on a GNSS buoy according to an embodiment of the present invention;

[0049] Figure 2 A flowchart of predicting comprehensive deviation and compensating posture according to an embodiment of the present invention;

[0050] Figure 3 This is a flow chart of wave in-situ calibration optimization according to an embodiment of the present invention. DETAILED DESCRIPTION

[0051] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0052] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0053] Embodiment 1

[0054] like Figure 1-Figure 3 As shown, this embodiment provides a wave in-situ calibration method based on a GNSS buoy, comprising the following steps:

[0055] Based on the position and attitude data of GNSS buoys under different sea conditions, a nonlinear mapping model between buoy deviation and sea condition factors is established using machine learning algorithms, and a dynamic model describing the spatiotemporal characteristics of buoy deviation is obtained. This model can describe the nonlinear and time-varying characteristics of buoy deviation as sea conditions change.

[0056] Deviation compensation models are constructed for the horizontal and vertical components of the buoy deviation. The horizontal deviation compensation model focuses on the influence of ocean current factors, while the vertical deviation compensation model focuses on the influence of ocean waves. By integrating multi-source ocean environment information, refined compensation for the horizontal and vertical components of the buoy deviation is achieved.

[0057] The sliding window technology is used to analyze the buoy deviation data within a certain time range, extract the sliding characteristics of the deviation, and establish a deviation sliding model to describe and compensate for the dynamic cumulative effect of the buoy deviation, thereby improving the timeliness and continuity of deviation compensation.

[0058] In the deviation modeling process, an adaptive weight adjustment strategy is introduced to dynamically adjust the weight of each sea condition factor in the deviation model according to the significant differences in deviation characteristics under different sea conditions, thereby improving the adaptability and robustness of the deviation model.

[0059] Construct a wave in-situ calibration model based on deviation compensation, fuse the deviation compensation results with GNSS positioning data, generate high-precision wave parameter estimation results, and improve the accuracy of wave in-situ calibration;

[0060] Adopting ensemble learning method to integrate the prediction results of multiple machine learning algorithms, such as support vector machine and random forest, to form a more robust deviation compensation model and improve the reliability and accuracy of deviation compensation;

[0061] A deviation compensation effect evaluation system is established. By introducing external high-precision wave observation data, the wave in-situ calibration accuracy before and after deviation compensation is evaluated. Based on this, the deviation compensation model and strategy are iteratively optimized to achieve continuous improvement in the wave in-situ calibration accuracy.

[0062] As a specific implementation method of this embodiment, the following steps are included:

[0063] The buoy deviation data is obtained based on the position and attitude data of the GNSS buoy under different sea conditions. The following steps are included:

[0064] Based on the position and attitude data of GNSS buoy under different sea conditions, a nonlinear mapping model between buoy deviation and sea condition factors is established using a machine learning algorithm, and a dynamic model describing the spatiotemporal characteristics of buoy deviation is obtained. This model can characterize the nonlinear and time-varying characteristics of buoy deviation as sea conditions change.

[0065] Furthermore, the position and attitude data collected by the GNSS buoy under different sea conditions are obtained to construct a training data set and a test data set. The training data set is used to train the machine learning model, and the test data set is used to evaluate the model performance.

[0066] The support vector machine regression algorithm is used to train the nonlinear mapping model with sea condition parameters as input features and buoy deviation as output target.

[0067] During the training process, the hyperparameters of the nonlinear mapping model are optimized through cross-validation method to improve the generalization ability of the nonlinear mapping model;

[0068] The trained nonlinear mapping model is applied to the test data set to evaluate the accuracy of the nonlinear mapping model in predicting buoy deviation under different sea conditions.

[0069] If the prediction accuracy of the nonlinear mapping model meets the preset threshold, the nonlinear mapping model is determined as the optimal model;

[0070] If the prediction accuracy of the nonlinear mapping model does not meet the preset threshold, the structure or hyperparameters of the nonlinear mapping model are adjusted, and training and testing are performed again;

[0071] Using the optimal model and combining the sea condition parameters obtained in real time, the buoy deviation is dynamically predicted to change over time and space;

[0072] Based on the prediction results, a dynamic model describing the temporal and spatial characteristics of the buoy deviation is generated;

[0073] The dynamic model is analyzed to study the nonlinear and time-varying characteristics of the buoy deviation as the sea conditions change;

[0074] According to different sea conditions, the dynamic model is used to predict the changing trend of buoy deviation;

[0075] If the prediction results show that the buoy deviation exceeds the preset threshold, the warning mechanism will be triggered to prompt the user to take corresponding measures;

[0076] Continuously collect the position and attitude data of GNSS buoys under actual sea conditions, regularly update the training data set, and retrain the nonlinear mapping model. Through continuous iterative optimization, improve the dynamic model's ability to characterize the spatiotemporal characteristics of buoy deviations, and enhance the robustness and adaptability of the dynamic model.

[0077] Furthermore, in order to construct training and test data sets, 1000 sets of position and attitude data can be collected by GNSS buoys under different sea conditions, such as wave height of 5 meters, 1 meter, 5 meters, etc., wind speed of 5m / s, 10m / s, 15m / s, etc., of which 800 sets are used as training sets and 200 sets are used as test sets. When training the nonlinear mapping model, sea state parameters such as wave height and wind speed are used as input features, and buoy deviation distance is used as output target. The support vector machine regression algorithm is used to optimize the model hyperparameters, such as penalty coefficient C, kernel function type, etc., through 5-fold cross validation to improve the generalization ability of the model. The trained model is applied to the test set to evaluate the prediction accuracy under different sea conditions. If the mean absolute error is less than 1 meter, it is determined to be the optimal model; otherwise, the model structure is adjusted, such as changing the kernel function, increasing the number of samples, etc., and retraining and optimization is performed. Using the optimal model, combined with the sea state parameters obtained in real time, such as updating every 10 minutes, the change of buoy deviation in the next hour is predicted to generate a spatiotemporal dynamic model. The analysis found that the buoy deviation and the wave height show a nonlinear relationship. For example, when the wave height increases from 5 meters to 5 meters, the deviation distance increases from 8 meters to 5 meters, an increase of 215%. The deviation distance also has a time-varying characteristic. For example, when the wind speed is 10m / s, the deviation distance increases first and then decreases over time, showing a periodic fluctuation with a period of about 20 minutes. For the case where the wave height is greater than 5 meters, the buoy deviation is predicted to exceed 3 meters, triggering the early warning mechanism, prompting the user to take measures, such as increasing the length of the anchor chain, adjusting the buoy position, etc. 100 sets of new position and attitude data are collected every month, the training set is updated, and the optimization model is retrained to continuously improve the characterization ability and adaptability of the dynamic model.

[0078] Deviation compensation model and deviation sliding model are constructed based on buoy deviation data. The steps include:

[0079] Furthermore, a deviation compensation model is constructed based on the buoy deviation data to obtain a first deviation compensation, and the process of obtaining the first deviation compensation includes: obtaining real-time position information and historical trajectory data of the GNSS buoy; analyzing the horizontal displacement and vertical displacement of the GNSS buoy using a Kalman filter algorithm based on the real-time position information and the historical trajectory data; calculating the horizontal component and vertical component of the buoy deviation based on the horizontal displacement and vertical displacement of the GNSS buoy; obtaining multi-source ocean environment information, and performing feature extraction on the multi-source ocean environment information to obtain key environmental factors affecting the buoy deviation; constructing a horizontal deviation compensation model for the horizontal component based on the ocean current factors in the key environmental factors; constructing a vertical deviation compensation model for the vertical component based on the ocean wave factors in the key environmental factors; and obtaining the first deviation compensation based on the horizontal deviation compensation model and the vertical deviation compensation model.

[0080] Deviation compensation models are constructed for the horizontal and vertical components of the buoy deviation respectively. The horizontal deviation compensation model focuses on the influence of ocean current factors, while the vertical deviation compensation model focuses on the influence of ocean wave factors. By integrating multi-source ocean environment information, refined compensation for the horizontal and vertical components of the buoy deviation is achieved, namely, the first deviation compensation.

[0081] Furthermore, the real-time position information and historical trajectory data of the buoy are obtained, and the horizontal and vertical components of the buoy deviation are determined by analyzing the horizontal and vertical displacements of the buoy; multi-source ocean environmental information is obtained, and the multi-source ocean environmental information includes ocean current data, ocean wave data, wind speed data, and wind direction data, and the key environmental factors affecting the buoy deviation are obtained by data fusion and feature extraction of the multi-source ocean environmental information; for the horizontal component of the buoy deviation, a horizontal deviation compensation model combined with the influence of ocean current is constructed, and a machine learning algorithm is used to establish a nonlinear mapping relationship between the ocean current factor in the key environmental factors and the horizontal deviation; for the vertical component of the buoy deviation, a vertical deviation compensation model combined with the influence of ocean waves is constructed, and a time series analysis is used Methods The dynamic relationship between the wave factor and the vertical deviation in the key environmental factors is established; the adaptive adjustment mechanism is introduced into the horizontal deviation compensation model and the vertical deviation compensation model, and the model parameters of the horizontal deviation compensation model and the vertical deviation compensation model are adjusted in real time according to the change of the buoy position and the dynamic change of the key environmental factors to obtain the adjusted horizontal deviation compensation amount and the vertical deviation compensation amount; the adjusted horizontal deviation compensation amount and the vertical deviation compensation amount are fused to obtain the comprehensive compensation vector of the buoy deviation, and the comprehensive compensation vector is applied to the position information of the buoy to obtain the compensated buoy position information; the compensated buoy position information is transmitted to the monitoring center, and the buoy is finely tracked and managed according to the compensated buoy position information.

[0082] Furthermore, the real-time position information of the buoy is obtained through the GNSS positioning module, and the historical trajectory data is stored in the cloud database. The horizontal and vertical displacements of the buoy are estimated using the Kalman filter algorithm to obtain the horizontal and vertical components of the buoy deviation. The root mean square error of the horizontal displacement estimation is 5 meters, and the root mean square error of the vertical displacement estimation is 8 meters. Multi-source marine environmental information such as currents, waves, wind speed and wind direction is obtained from the marine environmental monitoring station, and the heterogeneous data is normalized using data fusion technology to extract the key environmental factors affecting the buoy deviation. Through the principal component analysis method, it is found that the current speed and wave height are the main factors affecting the horizontal deviation and vertical deviation of the buoy, and the cumulative contribution rate of the two is more than 85%. For the horizontal deviation of the buoy, a compensation model combined with the influence of the current is constructed, and the support vector machine regression algorithm is used to establish the nonlinear mapping relationship between the current speed and the horizontal deviation. The average absolute error of the 5-fold cross validation is 2 meters. For the vertical deviation of the buoy, a compensation model combined with the influence of waves was constructed. The ARIMA time series model was used to establish the dynamic relationship between wave height and vertical deviation. The root mean square error of the fitting residual was 6 meters. An adaptive adjustment mechanism was introduced into the compensation model. According to the changes in the buoy position and environmental factors, the model parameters were updated every 10 minutes to adjust the horizontal deviation compensation and vertical deviation compensation in real time. After 100 iterative updates, the absolute error of the horizontal deviation compensation model was reduced by 3 meters, and the root mean square error of the vertical deviation compensation model was reduced by 2 meters. Finally, the adjusted horizontal deviation compensation and vertical deviation compensation were vector-synthesized to obtain the comprehensive compensation vector of the buoy deviation, which was applied to the correction of the buoy position information. The corrected buoy position information was transmitted to the monitoring center in real time through the Beidou communication module. The monitoring center tracked and managed the buoy in detail according to its precise position. If the buoy position deviated from the preset route by more than 50 meters, the alarm mechanism was activated to notify the staff to inspect and maintain the buoy.

[0083] The sliding window technology is used to analyze the buoy deviation data within a certain time range and extract the sliding characteristics of the deviation. By establishing a deviation sliding model, the dynamic cumulative effect of the buoy deviation can be characterized and compensated, thereby improving the timeliness and continuity of deviation compensation.

[0084] Further, buoy deviation data within a preset time range is obtained, and the preset time range is determined by a sliding window technology;

[0085] The feature extraction algorithm is used to extract the features of the buoy deviation data, and the feature vector reflecting the sliding characteristics of the buoy deviation is obtained;

[0086] The feature vector is input into the deviation modeling module. The deviation modeling module is based on a machine learning algorithm and uses the feature vector as a training sample. The buoy deviation sliding model is obtained through training. The buoy deviation sliding model is used to describe the law of dynamic cumulative changes of the buoy deviation over time.

[0087] When new buoy deviation data is obtained, the new buoy deviation data is input into the buoy deviation sliding model, and the buoy deviation sliding model is used to predict the changing trend of the dynamic accumulation of buoy deviation in the future period of time;

[0088] Determine the compensation value of the buoy deviation according to the future change trend of the buoy deviation;

[0089] The compensation value is integrated with the current buoy deviation data to obtain the compensated buoy deviation data;

[0090] Continuously track the changes in the buoy deviation data. When the distribution characteristics of the buoy deviation data change significantly, trigger the retraining of the buoy deviation sliding model and use the latest buoy deviation data to update the parameters of the buoy deviation sliding model.

[0091] Furthermore, the buoy deviation data of the past 30 days in the preset time range is determined by sliding window technology. The wavelet transform algorithm is used to extract the features of the buoy deviation data, and a 10-dimensional feature vector reflecting the sliding characteristics of the buoy deviation is extracted, including statistical features such as mean, variance, skewness, and kurtosis. The extracted feature vector is input into the deviation modeling module based on the long short-term memory neural network (LSTM), and the feature vectors of the past 30 days are used as training samples. The buoy deviation sliding model is obtained through 100 iterations of training. The model is used to predict the trend of buoy deviation changes in the next 7 days, and the predicted mean absolute percentage error (MAPE) is 8%. According to the predicted future change trend, the exponential smoothing method is used to determine the compensation value of the buoy deviation, that is, the second deviation compensation. The calculation formula of the compensation value is: compensation value = current deviation value × (1 + predicted change rate). The calculated compensation value is weighted and fused with the current buoy deviation data, with a weight ratio of 6:4, to obtain the compensated buoy deviation data. We continuously track changes in buoy deviation data and conduct a Kolmogorov-Smirnov test on the data distribution every 7 days. When the P value of the test result is less than 05, we trigger retraining of the buoy deviation sliding model and use the buoy deviation data from the last 30 days to update the model parameters. The MAPE of the updated model on the test set is reduced to 6%.

[0092] Based on the deviation compensation model and the deviation sliding model, a deviation compensation result is obtained, wherein the deviation compensation result includes a first deviation compensation and a second deviation compensation. The following steps are included:

[0093] In the deviation modeling process, an adaptive weight adjustment strategy is introduced to dynamically adjust the weights of various sea condition factors in the deviation model according to the significant differences in deviation characteristics under different sea conditions, thereby improving the adaptability and robustness of the deviation model.

[0094] Furthermore, the process of inputting the characteristic vector into the deviation modeling module to obtain the deviation sliding model also includes model optimization, and the model optimization process includes: obtaining historical deviation data under different sea conditions, analyzing the historical deviation data, obtaining the correlation strength between each sea condition factor and the deviation characteristics, and obtaining the typical sea condition based on the correlation strength; constructing a deviation prediction model based on the typical sea condition; calculating the comprehensive deviation prediction result under the typical sea condition based on the deviation prediction model; introducing the comprehensive deviation prediction result into the control system to optimize the second deviation compensation to obtain the optimized deviation sliding model.

[0095] Furthermore, historical deviation data under different sea conditions are obtained, and the historical deviation data are analyzed to obtain the correlation strength between each sea condition factor and the deviation characteristics; for each sea condition, a corresponding deviation prediction model is constructed according to the correlation strength, and the initial weight of each sea condition factor is determined; current sea condition data is obtained to determine which typical sea condition type the current sea condition data belongs to; according to the typical sea condition type, the corresponding deviation prediction model is called, and the weight value of each sea condition factor in the deviation prediction model is dynamically adjusted according to the current sea condition data, wherein the weight of the sea condition factor with significant deviation characteristics is increased, and the weight of the sea condition factor with insignificant deviation characteristics is reduced; a comprehensive deviation prediction result is obtained by weighted combination of the deviation prediction values ​​of each sea condition factor; the comprehensive deviation prediction result is introduced into the control system to adjust and compensate the navigation attitude in real time; the deviation feedback data is continuously collected, and the deviation prediction model and weight value are regularly verified and updated.

[0096] Furthermore, the buoy deviation data under different sea conditions are obtained from the historical database, including wind speed, wind direction, current speed, current direction, wave height and other sea condition factors. The Pearson correlation coefficient method is used to analyze the correlation strength between each sea condition factor and the deviation characteristics. The results show that the correlation coefficient between wind speed and deviation amplitude is 8, the correlation coefficient between current speed and deviation direction is 7, and the correlation coefficient between wave height and deviation frequency is 6. For three typical sea conditions of strong wind, strong current and large waves, the deviation prediction model based on support vector machine (SVM) is constructed respectively, and the initial parameters of the model are determined by cross-validation. The current sea condition data is clustered and determined to be a strong current sea condition. The SVM deviation prediction model under strong current sea conditions is called, and the weight values ​​of the two features of current speed and current direction are dynamically adjusted through the sigmoid function according to the real-time current speed and direction data, with the weight range of 6 to 0. The predicted deviation values ​​of the three features of wind speed, current speed and wave height are weighted and combined, with a weight ratio of 3:5:2, to obtain a comprehensive deviation prediction result. The prediction result is transmitted to the autopilot control system, and the heading and speed are adjusted in real time through the PID control algorithm to compensate for the navigation deviation. After each voyage, GNSS positioning data and inertial navigation data are collected to calculate the root mean square error (RMSE) between the actual deviation value and the predicted deviation value. When the RMSE of three consecutive voyages exceeds 2 nautical miles, the retraining of the deviation prediction model is triggered, and the parameters of the SVM model are optimized and the weights are updated using the latest 100 historical deviation data samples to continuously improve the accuracy and adaptability of deviation prediction.

[0097] The deviation compensation result is integrated with the GNSS positioning data to obtain the wave parameter estimation result. The following steps are included:

[0098] A wave in-situ calibration model based on deviation compensation is constructed, and the deviation compensation results are fused with the GNSS positioning data to generate high-precision wave parameter estimation results, thereby improving the accuracy of the wave in-situ calibration.

[0099] Further, wave in-situ observation data and synchronized GNSS positioning data are obtained, and both the wave in-situ observation data and GNSS positioning data are time series data;

[0100] According to the wave in-situ observation data, a wave in-situ calibration model based on machine learning is constructed. The machine learning model selects at least one algorithm from support vector machine, neural network and random forest.

[0101] The wave in-situ calibration model is applied to the wave in-situ observation data to obtain the deviation compensation result of the wave in-situ observation data;

[0102] Using Kalman filtering or particle filtering algorithm, the deviation compensation result is fused with the GNSS positioning data to obtain the fused wave observation data;

[0103] According to the fused wave observation data, the wave parameters are estimated by using the least square method or the Bayesian estimation method, and the wave parameters include at least one of the wave height, the wave period and the wave direction, so as to obtain the wave parameter estimation result;

[0104] Compare the wave parameter estimation results with the wave in-situ observation data to calculate the improvement of the wave in-situ calibration accuracy;

[0105] If the wave in-situ calibration accuracy is improved to a preset threshold, the wave parameter estimation result is output as the final wave in-situ calibration result;

[0106] If the improvement of the wave in-situ calibration accuracy does not reach the preset threshold, return to the step of acquiring data, reacquire the wave in-situ observation data and GNSS positioning data, and optimize the wave in-situ calibration model until the accuracy requirements are met.

[0107] Furthermore, wave in-situ observation data and synchronous GNSS positioning data for one month were obtained from the wave buoy, with a data sampling frequency of 1Hz. For the wave in-situ observation data, the support vector regression (SVR) algorithm was used to construct the wave in-situ calibration model. The input features included six dimensions such as acceleration and angular velocity, and the output was the wave height deviation compensation value. The grid search method was used to optimize the hyperparameters of the SVR model, such as the kernel function type and penalty coefficient, and the optimal model was determined through 5-fold cross validation. The trained SVR model was applied to the wave in-situ observation data to obtain the deviation compensation result of the wave height. The Kalman filter algorithm was used, with GNSS positioning data as the observation quantity and the deviation compensation result as the control quantity, to establish a state space model, and recursively estimate the optimal value of the wave height through two steps of prediction and update, so as to realize the fusion of the wave in-situ observation data and GNSS positioning data. According to the fused wave observation data, the least squares method was used to estimate the wave parameters, including the significant wave height, the average wave period and the average wave direction. The estimated wave parameters were compared with the in-situ observations to calculate the improvement in wave height calibration accuracy. The results showed that the wave height error after calibration was reduced by 3m, meeting the preset threshold of 2m. The wave parameter estimation results were output as the final wave in-situ calibration results. To further improve the calibration accuracy, new observation data were continuously acquired, and the SVR model was dynamically updated using online learning to continuously optimize the wave in-situ calibration performance. After 3 months of iterative optimization, the wave height calibration accuracy was stabilized within 1m, meeting the requirements of engineering applications.

[0108] An ensemble learning method is used to integrate the prediction results of multiple machine learning algorithms, such as support vector machines and random forests, to form a more robust deviation compensation model and improve the reliability and accuracy of deviation compensation.

[0109] Furthermore, a historical data set is obtained, and the historical data set is preprocessed to obtain a training data set and a test data set; for the training data set, at least two machine learning algorithms are used for training to obtain at least two prediction models; the prediction results of the at least two prediction models are weighted averaged to obtain the prediction results of the integrated learning model; for the prediction results of the integrated learning model, a deviation compensation model is constructed, and the deviation compensation model is used to analyze the prediction error distribution characteristics of the integrated learning model and correct the prediction results of the integrated learning model; an adaptive weight adjustment strategy is used to dynamically adjust the weights of at least two prediction models in the integrated learning model according to their prediction performance; new data is obtained, and an online learning mechanism is used to incrementally update the integrated learning model according to the new data; the prediction results of the integrated learning model and the output of the deviation compensation model are combined to obtain the final prediction results.

[0110] Furthermore, data from the past year were obtained from the historical database, totaling 10,000 records. 80% were randomly selected as the training data set, and the remaining 20% ​​were used as the test data set. The data was preprocessed, including missing value filling, outlier processing, and data standardization. The KNN algorithm was used to fill the missing values, and the outliers exceeding 3 times the standard deviation were replaced with the average value. The Z-score method was used to standardize the data. For the training data set, the support vector machine (SVM) and random forest (Random Forest) algorithms were used for training, respectively. The model hyperparameters were optimized by the grid search method. The SVM model used the RBF kernel function, the penalty coefficient C = 10, the number of decision trees of the random forest model was 100, and the maximum tree depth was 5. The prediction results of the two models were weighted averaged according to the weight ratio of 2:1 to obtain the prediction results of the ensemble learning model. According to the prediction results of the ensemble learning model, a deviation compensation model is constructed. The local weighted linear regression (LWLR) algorithm is used to fit the compensation function in the local area according to the distribution characteristics of the prediction error, and the prediction results of the ensemble learning model are corrected. The compensation coefficient is adaptively adjusted according to the size of the prediction error, and the maximum compensation range does not exceed 20%. According to the prediction performance of the SVM and random forest models on the test data set, an adaptive weight adjustment strategy is adopted. When the prediction error of the SVM model is smaller than that of the random forest model, the weight of the SVM model is increased, otherwise the weight of the random forest model is increased, and the weight adjustment range is 5%. Whenever 100 new data are obtained, the online learning mechanism is used to add the new data to the training data set, and the SVM and random forest models are incrementally updated at the same time. The updated model continues to be used for ensemble learning. The prediction results of the ensemble learning model and the output of the deviation compensation model are combined to obtain the final prediction results. After evaluation, the prediction accuracy reaches more than 95%, which meets business needs.

[0111] A deviation compensation effect evaluation system is established. By introducing external high-precision wave observation data, the wave in-situ calibration accuracy before and after deviation compensation is evaluated. Based on this, the deviation compensation model and strategy are iteratively optimized to achieve continuous improvement in the wave in-situ calibration accuracy.

[0112] Furthermore, external high-precision wave observation data is obtained as a reference standard for evaluating the effect of deviation compensation; based on the external high-precision wave observation data, the wave in-situ calibration accuracy of the current deviation compensation model is evaluated; if the wave in-situ calibration accuracy of the current deviation compensation model does not reach the preset threshold, then according to the evaluation results, the parameters of the deviation compensation model are optimized using a machine learning algorithm; according to the optimized deviation compensation model, the wave observation data is re-biased and processed to obtain compensated wave observation data; the compensated wave observation data is input into the in-situ calibration model to obtain new calibration results; the new calibration results are compared with the external high-precision wave observation data to evaluate the improvement effect of the optimized deviation compensation model on the in-situ calibration accuracy; if the improvement effect reaches the preset threshold, the optimized deviation compensation model is applied to actual business; otherwise, the deviation compensation model is continuously iterated and optimized until the calibration accuracy meets the preset requirements.

[0113] Furthermore, high-precision wave observation data from the past year were obtained from the National Marine Environmental Monitoring Center, totaling 10,000 records, which were used as reference standards for evaluating the effect of bias compensation. The root mean square error (RMSE) indicator was used to evaluate the wave in-situ calibration accuracy of the current bias compensation model. If the RMSE was greater than 2m, it was considered that the calibration accuracy did not reach the preset threshold. The genetic algorithm was used to optimize the parameters of the bias compensation model, setting the population size to 50, the crossover probability to 8, the mutation probability to 1, the number of iterations to 100, and the objective function to minimize RMSE. The optimized bias compensation model was applied to the original wave observation data, and the bias compensation was performed to obtain the compensated wave observation data. The compensated data was input into the in-situ calibration model based on support vector regression (SVR), using the radial basis kernel function, the penalty coefficient C = 10, the number of training samples to 8000, and the number of test samples to 2000, to obtain new calibration results. The new calibration results were compared with the high-precision wave observation data, and the RMSE was calculated to evaluate the effect of the optimized bias compensation model on the improvement of the in-situ calibration accuracy. If the RMSE is reduced by more than 30%, it is considered that the improvement effect has reached the preset threshold, and the optimized deviation compensation model is applied to actual business; otherwise, the deviation compensation model is continuously optimized iteratively, and the parameter settings of the genetic algorithm are adjusted, such as increasing the population size, increasing the crossover probability, etc., until the calibration accuracy meets the requirement of RMSE less than 1m. Through the optimization of the deviation compensation model and the retraining of the in-situ calibration model, high-precision calibration of wave observation data is finally achieved. The consistency between the calibrated wave parameters and the high-precision observation data reaches more than 95%, which can meet the needs of various marine engineering and scientific research applications.

[0114] Embodiment 2

[0115] This embodiment also provides a computer terminal device, including:

[0116] one or more processors;

[0117] A memory, coupled to the processor, for storing one or more programs;

[0118] When the one or more programs are executed by the one or more processors, the one or more processors implement a wave in-situ calibration method based on a GNSS buoy.

[0119] Embodiment 3

[0120] This embodiment also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, a wave in-situ calibration method based on a GNSS buoy is implemented.

[0121] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A wave in-situ calibration method based on GNSS buoy, characterized in that: The following steps are involved: The buoy deviation data is obtained based on the position and attitude data of the GNSS buoy in different sea conditions; Constructing a deviation compensation model and a deviation sliding model based on the buoy deviation data; Obtaining a deviation compensation result based on the deviation compensation model and the deviation sliding model, wherein the deviation compensation result includes a first deviation compensation and a second deviation compensation; The deviation compensation result is fused with the GNSS positioning data to obtain the wave parameter estimation result.

2. The wave in-situ calibration method based on GNSS buoy according to claim 1, characterized in that: A deviation compensation model is constructed based on the buoy deviation data to obtain a first deviation compensation. The process of obtaining the first deviation compensation includes: Obtain real-time location information and historical trajectory data of GNSS buoys; Analyzing the horizontal displacement and vertical displacement of the GNSS buoy using a Kalman filter algorithm based on the real-time position information and historical trajectory data; Calculate the horizontal component and the vertical component of the buoy deviation based on the horizontal displacement and the vertical displacement of the GNSS buoy; Acquire multi-source ocean environment information, and perform feature extraction on the multi-source ocean environment information to obtain key environmental factors that affect buoy deviation; Constructing a horizontal deviation compensation model for the horizontal component based on the ocean current factor among the key environmental factors; Constructing a vertical deviation compensation model for the vertical component based on the wave factor in the key environmental factors; A first deviation compensation is obtained based on the horizontal deviation compensation model and the vertical deviation compensation model.

3. The wave in-situ calibration method based on GNSS buoy according to claim 2, characterized in that: The process of building a bias sliding model includes: Acquire buoy deviation data within a preset time range based on a sliding window; The buoy deviation data is analyzed based on the feature extraction algorithm to obtain the feature vector of the buoy deviation sliding characteristics; Input the characteristic vector into the deviation modeling module to obtain the cumulative change law and construct a deviation sliding model; Calculating a deviation prediction value based on the cumulative change rule, and calculating a compensation value of the buoy deviation; Based on the deviation sliding model, the compensation value and the current buoy deviation data are weightedly fused to obtain a second deviation compensation.

4. The wave in-situ calibration method based on GNSS buoy according to claim 3 is characterized in that: The process of inputting the feature vector into the deviation modeling module to obtain the deviation sliding model also includes model optimization, and the process of model optimization includes: Acquire historical deviation data under different sea conditions, analyze the historical deviation data, obtain the correlation strength between each sea condition factor and the deviation characteristic, and obtain a typical sea condition based on the correlation strength; Constructing a deviation prediction model based on the typical sea conditions; Calculate the comprehensive deviation prediction result under typical sea conditions based on the deviation prediction model; The comprehensive deviation prediction result is introduced into the control system to optimize the second deviation compensation to obtain an optimized deviation sliding model.

5. The wave in-situ calibration method based on GNSS buoy according to claim 3, characterized in that: The process of fusing the deviation compensation result with the GNSS positioning data to obtain the wave parameter estimation result includes: Obtain wave in-situ observation data and synchronized GNSS positioning data; Constructing a wave in-situ calibration model based on the wave in-situ observation data; Inputting the deviation compensation result into the wave in-situ calibration model to obtain the deviation compensation result of the wave in-situ observation data; A particle filter algorithm is used to fuse the deviation compensation result of the wave in-situ observation data with the synchronized GNSS positioning data to obtain fused wave observation data; The least square method is used to perform parameter estimation on the fused wave observation data to obtain a wave parameter estimation result.

6. The wave in-situ calibration method based on GNSS buoy according to claim 5, characterized in that: After the parameters of the fused wave observation data are estimated, the method also includes establishing a deviation compensation effect evaluation system by introducing external high-precision wave observation data; The wave parameter estimation result is compared with the wave in-situ observation data to calculate the improvement of the wave in-situ calibration accuracy; Setting a preset threshold based on the external high-precision wave observation data; The wave in-situ calibration accuracy is determined by comparing the wave in-situ calibration accuracy improvement margin with the preset threshold.

7. The wave in-situ calibration method based on GNSS buoy according to claim 6, characterized in that: When the wave in-situ calibration accuracy improvement reaches a preset threshold, the wave parameter estimation result is output as the final wave in-situ calibration result; When the improvement of the wave in-situ calibration accuracy does not reach a preset threshold, the wave in-situ observation data and GNSS positioning data are reacquired, and the wave in-situ calibration model is optimized until the preset threshold is met.

8. A computer terminal device, characterized in that: include: one or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the wave in-situ calibration method based on the GNSS buoy as described in any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the wave in-situ calibration method based on a GNSS buoy as described in any one of claims 1 to 7 is implemented.

Images