Polar navigation ship behavior analysis method taking environmental factors into consideration and electronic equipment

Abstract

The invention relates to a polar navigation ship behavior analysis method considering environmental elements and electronic equipment, and the method comprises the steps: collecting multi-source data including sea ice, wind speed and AIS data of a polar water area, and carrying out the preprocessing; wherein the sea ice data comprises sea ice concentration data and sea ice thickness data; performing space-time matching on the AIS data, sea ice and wind speed data to obtain a multi-modal data set containing ship identification information, position, speed and environmental factors, and solving the polar region coordinate system adaptation problem; in combination with the multi-modal data set, polar region water area ship navigation behavior characteristics are analyzed, and the polar region water area ship navigation behavior characteristics comprise the seasonal space distribution rule of ship tracks and the seasonal differentiation distribution characteristics of navigational speeds; and constructing a three-level quantitative model of collinearity diagnosis, principal component analysis dimensionality reduction and double regression model comparison, and realizing dynamic quantification of the influence of environmental elements on the navigational speed. Compared with the prior art, the method has the advantages of accurately measuring the dynamic influence of environmental elements on the polar region ship navigation behavior and the like.

Description

Technical Field

[0001] This invention relates to the fields of ship navigation safety and data processing technology, and in particular to a method and electronic device for analyzing the behavior of polar ships that takes environmental factors into account. Background Technology

[0002] As global warming accelerates the melting of polar sea ice, the commercial value of polar shipping routes is becoming increasingly apparent, leading to a gradual increase in the navigation activities of vessels such as LNG carriers in polar waters. Currently, the analysis of vessel navigation behavior in polar waters mainly relies on Automatic Identification System (AIS) data, which needs to be combined with environmental factors such as sea ice concentration, sea ice thickness, and wind speed to accurately reflect the interaction patterns between vessels and the polar environment.

[0003] A search revealed that Chinese invention patent application publication number CN117113659A discloses a convoy following simulation method that considers the navigation safety and energy consumption of ships in polar regions. This method considers the number of ships ahead, predictive utility, or ice floe density. Parameters are set in the model, and based on these parameters, a multi-ship following model assisted by ship networking technology is established, taking into account the future movement trends of multiple ships ahead and the impact of ice floes on ship navigation. The model updates ship status information and monitors ship following status. However, the existing patent application uses mostly single parameters for sea ice environment data, neglecting the coupling effects of multiple factors. Therefore, it fails to accurately reflect the interaction patterns between ships and the polar environment.

[0004] A search revealed that Chinese invention patent application publication number CN119207162A discloses a generative method, device and medium for predicting daily navigation risks in polar waterways. The AIS data processing simply removes outliers and does not optimize the algorithm for issues such as satellite signal attenuation and positioning errors in high-latitude polar environments.

[0005] A search revealed that Chinese invention patent application publication number CN118194066A discloses a method and system for processing ice condition data of polar navigation vessels. The fusion of environmental data and AIS data often adopts simple spatiotemporal matching, which does not solve the problem of adapting polar stereo projection and WGS84 coordinate system, resulting in poor data correlation.

[0006] In addition, existing navigation behavior analysis mostly uses linear regression models, which cannot capture the nonlinear response of ship behavior under extreme conditions such as sudden increases in sea ice thickness and sudden changes in wind speed, and it is difficult to quantify the dynamic impact of environmental factors on speed and trajectory.

[0007] Understanding the characteristics of ship navigation behavior and speed change mechanisms under complex sea ice conditions in polar routes to support the safe and efficient operation of polar vessels has become a technical problem that needs to be solved. Summary of the Invention

[0008] The purpose of this invention is to overcome the shortcomings of the prior art by providing a method and electronic device for analyzing the behavior of polar navigation vessels that takes environmental factors into account.

[0009] The objective of this invention can be achieved through the following technical solutions: According to one aspect of the present invention, a method for analyzing the behavior of polar navigation vessels considering environmental factors is provided, the method comprising: Multi-source data acquisition and preprocessing: Multi-source data including sea ice, wind speed and AIS data from polar waters were acquired and preprocessed; the sea ice data included sea ice concentration data and sea ice thickness data. Multimodal data fusion: Spatiotemporal matching of AIS data with sea ice and wind speed data to obtain a multimodal dataset containing ship identification information, location, speed and environmental elements, solving the polar coordinate system adaptation problem; Navigation behavior characteristic analysis: Combining multimodal datasets, we analyze the characteristics of ship navigation behavior in polar waters, including the seasonal spatial distribution pattern of ship trajectories and the seasonal variation distribution characteristics of ship speed. Environmental impact quantification analysis: Construct a three-level quantitative model that includes collinearity diagnosis, principal component analysis dimensionality reduction, and biregression model comparison to capture the linear and nonlinear effects of environmental factors on speed, and achieve dynamic quantification of the impact of environmental factors on speed.

[0010] As a preferred technical solution, the spatiotemporal matching process includes: Coordinate system transformation: The Proj4 library is used to implement the coordinate system transformation between AIS data's WGS84 latitude and longitude and polar stereo projection; Spatiotemporal matching: A spatiotemporal matching of AIS data with sea ice and wind speed data is achieved through time alignment and spatial interpolation.

[0011] As a preferred technical solution, the Proj4 string for the coordinate system transformation is: + proj = stere + lat 0 = 90 + lat ts = 70 + lon 0 = -45 + x 0 = 0 + y 0 = 0 + datum = WGS 84+ units = m + no defs , Among them, + proj = stere The projection type is stereoscopic projection; lat 0 = 90: Latitude of the projection origin; lat ts =70: Standard latitude of the parallel; lon 0 = -45: Longitude of the central meridian; x 0 = 0, y 0 = 0: Offset of projection origin coordinates; datum = WGS 84: Reference ellipsoid; units = m The coordinate unit after projection is meters.

[0012] As a preferred technical solution, the coordinate system transformation maps AIS data from spherical coordinates to planar coordinates, and the specific transformation formula is as follows: , The mapping coefficient R from spherical coordinates to planar coordinates is calculated by the following formula: , Where (x, y) are planar coordinates; λ Longitude; λ 0 = -45° is the longitude of the central meridian; R 0 represents the Earth's radius; φ Latitude φ ts The latitude is the standard parallel.

[0013] As a preferred technical solution, the spatiotemporal matching includes: Time matching: AIS data and wind speed data are matched at hourly resolution; AIS data and sea ice data are matched at daily resolution. Spatial matching: Linear interpolation is used to map the transformed AIS planar coordinates to the grid of sea ice data to obtain the sea ice data of the corresponding grid; based on the AIS WGS84 latitude and longitude, the wind speed data of the corresponding grid is obtained through nearest neighbor interpolation.

[0014] As a preferred technical solution, based on the multimodal dataset, the ship's trajectory throughout the year is plotted, and the seasonal spatial distribution pattern of the ship's trajectory is statistically analyzed; based on the multimodal dataset, the ship speed statistics are plotted, and the seasonal differences in the distribution characteristics of the speed are statistically analyzed.

[0015] As a preferred technical solution, the collinearity diagnosis includes: using the variance inflation factor to test the collinearity of three environmental variables; if the calculated collinearity value is less than 5, it is determined that there is no obvious collinearity; otherwise, it is determined that collinearity exists. The three environmental variables mentioned above include sea ice concentration, sea ice thickness, and wind speed.

[0016] As a preferred technical solution, the dual regression model includes an OLS linear regression model and a random forest nonlinear regression model; If there is no obvious collinearity, the original environmental variables are used directly to construct an OLS linear regression model; If collinearity exists, use principal component analysis to extract principal components for dimensionality reduction and construct an OLS linear regression model; By comparing the results of the two regression models, the model that is more suitable for predicting the speed of ships in polar waterways is selected as the final regression model.

[0017] As a preferred technical solution, the method further includes: before multimodal data fusion, performing data cleaning, cubic spline interpolation of natural boundary conditions, and data thinning on the AIS data.

[0018] According to another aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the program to implement the method described thereon.

[0019] Compared with the prior art, the present invention has the following beneficial effects: 1) This invention is specifically adapted to polar navigation scenarios. By adapting to the polar coordinate system and accurately fusing multimodal data, it eliminates the deviation of data spatiotemporal correlation. The navigation behavior patterns mined by combining polar characteristics are more in line with the actual scenario. The three-level quantification model of collinearity diagnosis, dimensionality reduction and dual regression model comparison not only avoids the collinearity interference of environmental variables, but also captures the linear and nonlinear effects at the same time, which greatly improves the accuracy and dynamic adaptability of quantifying the impact of polar environmental factors on navigation speed.

[0020] 2) This invention clarifies the coordinate system transformation process in polar scenarios (including Proj4 parameters and mapping formulas), effectively solving the coordinate system incompatibility problem between AIS data and sea ice data, and ensuring the coordinate consistency of multi-source data. At the same time, its categorized temporal resolution matching (hourly wind speed, daily sea ice) and corresponding interpolation spatial matching eliminate the correlation deviation caused by the difference in spatiotemporal resolution of different data, improving the accuracy of multimodal data fusion. The coordinate system transformation and spatiotemporal matching process ensures the reproducibility of the data fusion process, providing reliable high-quality dataset support for subsequent navigation behavior analysis and environmental impact quantification.

[0021] 3) This invention uses the variance inflation factor to test for collinearity of environmental variables, accurately identifying and avoiding the bias of collinearity in the estimation of regression model parameters, thus ensuring the basic reliability of the model. At the same time, the design of selecting OLS models according to different scenarios and comparing dual regression models not only preserves the explanatory power of the original variables when there is no collinearity, but also eliminates interference through principal component dimensionality reduction when there is collinearity. Furthermore, it can select the model that is more suitable for polar speed prediction by comparing the results of the two models, which significantly improves the accuracy of quantifying the impact of environmental factors on speed and its practical application value. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating the method for analyzing ship navigation behavior in polar waters in this invention. Figure 2 This is a schematic diagram of the AIS data processing flow in this invention; Figure 3 This is a schematic diagram of the multimodal data fusion process in this invention; Figure 4 This is a flowchart illustrating the process of analyzing the relationship between ship trajectory and speed in this invention. Figure 5 This is a schematic diagram showing the results of the analysis of the ship trajectory and speed patterns in this invention; Figure 6 This is a schematic diagram of the process for quantitative analysis of environmental impact in this invention; Figure 7 This is a schematic diagram illustrating the relationship between actual speed and predicted speed in this invention; Figure 8 This is a schematic diagram illustrating the relationship between SIC and speed in this invention; Figure 9 This is a schematic diagram illustrating the relationship between SIT and speed in this invention; Figure 10 This is a schematic diagram illustrating the relationship between I10wg and speed in this invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0024] This invention addresses the problems of low AIS processing accuracy, insufficient data fusion, and the inability of ordinary models to capture nonlinear relationships and narrow applicability in existing polar water vessel navigation behavior analysis technologies. It provides a polar navigation vessel behavior analysis method that considers environmental factors. Through multi-source data acquisition and preprocessing, AIS data processing, multimodal data fusion, navigation behavior feature analysis, and environmental impact quantification analysis, it achieves high-precision analysis of polar vessel navigation trajectories and speed characteristics, quantifies the dynamic impact of environmental factors such as sea ice and wind speed on vessel navigation behavior, and provides a scientific basis for polar route planning and vessel ice class selection.

[0025] This embodiment relates to a method for analyzing the behavior of ships navigating in polar regions that takes into account environmental factors, in order to comprehensively understand the characteristics of ship navigation behavior and the mechanism of speed change under complex sea ice conditions in polar routes.

[0026] This method, based on sea ice concentration and thickness products from the University of Bremen, ERA-5 reanalysis wind field data, and AIS trajectory data of LNG carriers transiting polar waters in 2024, employs data cleaning, cubic spline interpolation, Douglas-Peucker track thinning, multimodal spatiotemporal matching, variance inflation factor test, principal component analysis, ordinary least squares regression, and random forest regression. It systematically reveals the route selection and speed response mechanisms of ships in polar waters under different seasons, ice conditions, and wind field combinations, from multiple dimensions including sea ice concentration, sea ice thickness, and gust wind speed. This method can provide scientific basis and decision support for polar regulatory agencies, shipping companies, and energy enterprises to formulate seasonal operating plans for polar routes, route and vessel configuration schemes, and speed control strategies under complex sea ice and strong wind conditions. In the long term, this research is expected to support the safe and efficient operation of the polar LNG transportation network, promote the smooth transition of polar channels from experimental development to large-scale commercial operation, and is of great significance for improving the resilience and sustainable development level of the global energy transportation system.

[0027] The technical solution of the present invention is achieved through the following steps, corresponding to the appendix. Figure 1 (Overall method flowchart), the detailed logic of each step and the associated diagrams are as follows: S100: Multi-source data acquisition and preprocessing, acquiring sea ice, wind speed, and AIS data in polar waters, defining data format, resolution, and screening criteria: S110: Sea ice data acquisition. Sea ice concentration (SIC) data was downloaded from the University of Bremen's sea ice database in Germany. The data format was NetCDF4, the variable name was "sea_ice_concentration", the spatial resolution was 6.25×6.25km, the temporal resolution was 1 day, and the data range was 66°34′-90°N and -180°~180°L. Valid data with 0%≤SIC≤100% were selected. Sea ice thickness (SIT) data was also collected in NetCDF4 format. The variable "combined_thickness" (SMOS and SMAP sensor fusion value) was selected, the spatial resolution was 40×40km, the temporal resolution was 1 day, and the data range was the same as SIC. Valid data with 0cm≤SIT≤50cm were selected.

[0028] S120: Wind speed data acquisition. Download instantaneous 10m windgust (I10wg) data from the ERA-5 dataset. The data format is NetCDF, the variable name is "I10wg", the spatial resolution is 0.25°×0.25°, the time resolution is 1 hour, the data range is 66°34′~90° north latitude and -180°~180° longitude, and the unit is m / s. Valid data with I10wg≥0m / s are filtered.

[0029] S130: AIS data collection, obtained and downloaded from databases such as Shipxy.com, Lloyd's List Intelligence (LLI), and Clarkson. Static data includes: IMO number (International Maritime Organization number), flag state, ice class, length, beam, deadweight tonnage, etc.; dynamic data includes: MMSI (Maritime Mobile Services Identifier), speed, longitude, latitude, time, ship status, etc.

[0030] S200: AIS data processing. Addressing the high-missibility and high-interference nature of AIS data in polar regions, it outputs high-quality track data through data cleaning, interpolation, and thinning. (Corresponding appendix...) Figure 2 The AIS data preprocessing flowchart includes: S210: Data cleaning, implemented using Python, consists of three sub-steps: 1) Duplicate value deletion: Time repetition: Only one center record is retained per minute for the same vessel; Location duplication: If the difference in latitude and longitude between two adjacent records is ≤0.0001°, it is determined to be a location duplication, and the latter record is deleted; Duplicate records across all fields: Delete records with duplicate hash values ​​by comparing MMSI+Speed+Lon_d+Lat_d+Receivedtime; 2) Outlier filtering: Speed ​​anomalies: Calculate the 0.1 and 99.9 quantiles of speed for the entire dataset, and delete records below the 0.1 quantile and above the 99.9 quantile; Geographical anomalies: Delete records where Lat_d < 66.34° or Lat_d > 90°, Lon_d < -180° or Lon_d > 180°; Abnormal status: Only retain "In Navigation" status data, and delete status records such as "Anchored" and "Boated"; 3) Missing value deletion: Records with a missing key field rate greater than 5% will be deleted directly.

[0031] S220: Data interpolation, using cubic spline interpolation with natural boundary conditions for the missing segments of 10-60 minutes in the cleaned data: 1) Definition of interpolation model: Let the AIS data time series be t 0, t 1, ..., t n (Unit: minutes), corresponding location coordinates are ( x 0, y 0),( x 1, y 1), ..., ( x n , y n )( x =Lon_d, y =Lat_d). If t k+1 - t k If ∈[10,60], then construct a cubic spline function. S ( x ): , in, S k ( x ) = a k x 3 + b k x 2 + c kx + d k ( k = 0, 1, ..., n -1), satisfying: Continuity: S ( x k ) = y k , S ’ ( x k ) = y ’ k , S ’’ ( x k ) = y ’’ k ( k = 0, 1, ..., n ); Natural boundary conditions: S ’’ ( x 0) = 0, S ’’ ( x n ) = 0; 2) Interpolation parameter settings: Number of adjacent points: Select two valid points before and after the missing segment to construct the model; Interpolation step size: 1 minute / point, that is, if 30 minutes are missing, 30 interpolation points will be added.

[0032] S230: Data thinning. To address the issue of excessively large data volume after interpolation, the Douglas-Peucker (DP) algorithm is used for compression, preserving key feature points. The core formula of the DP algorithm: Let the sequence of waypoints be... P 0, P 1, ..., P m Connect the starting point P 0 and the endpoint P m Forming line segments P 0 P m Calculate the remaining points P i to line segment P 0 P m vertical distance di : , like d i >ϵ ( ϵ If the thinning threshold is used, then retain... P i and with P i Use this as the dividing point for recursive processing. P 0- P i and P i - P m ;like d i ≤ ϵ , then delete P i ; Threshold determination: Based on polar flight path characteristics, set... ϵ = 0.01°.

[0033] S300: Multimodal data fusion resolves coordinate system differences between AIS and sea ice / wind speed data, achieving high-precision spatiotemporal matching, corresponding to the attached... Figure 3 ,include: S310: Coordinate system transformation. The coordinate system transformation of AIS data is implemented using the Proj4 library. The projection parameters are consistent with those of the sea ice data. The formula and parameters are as follows: Projection definition: Polar stereo projection, applicable to high-latitude polar regions. The Proj4 string is: + proj = stere + lat 0 = 90 + lat ts = 70 + lon 0 = -45 + x 0 = 0 + y 0 = 0 + datum = WGS 84+ units = m + no defs , in: + proj = stere The projection type is stereoscopic projection; lat 0 = 90: Latitude of the projection origin (polar point); lat ts= 70: Standard parallel of latitude (parallel of latitude parallel to the equator, with no length distortion); lon 0 = -45: Central meridian (the meridian of the projection center); x 0 = 0, y 0 = 0: Offset of projection origin coordinates; datum = WGS 84: Reference ellipsoid; units = m The coordinate unit after projection is meters.

[0034] Conversion Formula: Converts AIS data's WGS84 latitude and longitude to planar coordinates (x, y). The conversion formula is based on the mapping from spherical coordinates to planar coordinates: , in: , R 0 represents the Earth's radius, taken as 6,378,137 meters; φ Latitude φ ts = 70° is the standard latitude of a parallel; λ Longitude λ 0 = -45° is the longitude of the central meridian, and R is the mapping coefficient from spherical coordinates to planar coordinates.

[0035] S320: Spatiotemporal matching, achieving temporal and spatial alignment between AIS and sea ice and wind speed data, including: 1) Time matching: AIS data time (Receivedtime) is rounded to the nearest hour to match the hour resolution of I10wg; AIS data date is matched with the day resolution of SIC / SIT to ensure that data can be correlated within the same time dimension.

[0036] 2) Spatial matching: Sea ice data (SIC, SIT): Linear interpolation is used to map the transformed AIS planar coordinates to the SIC and SIT grids to obtain the corresponding SIC and SIT values. Wind speed data (I10wg): Based on AIS WGS84 latitude and longitude, I10wg values ​​of 0.25°×0.25° grid are obtained by nearest neighbor interpolation. Fusion result: A multimodal dataset of “IMO number-time-location-velocity-SIC-SIT-I10wg” is generated.

[0037] S400: Navigation behavior characteristic analysis, combining multimodal datasets to analyze the seasonal and spatial patterns of ship trajectories and speeds. See the appendix for the detailed process. Figure 4 The analysis of navigation behavior characteristics is shown in the attached diagram. Figure 5 : S410: Trajectory feature analysis and spatial distribution patterns, drawing ship trajectories throughout the year, and statistically analyzing navigation areas in different seasons: ①Summer (July-October): Navigation area: The entire Northeast Passage, covering the Barents Sea, Kara Sea, and Laptev Sea; some ships can reach around 80° North latitude. Route selection: 85% of ships choose high-latitude routes, which shorten the voyage by about 800-1200 kilometers compared to traditional low-latitude routes.

[0038] ②Winter (November to June of the following year): Navigation area: No tracks were found only along the coast of the Barents Sea; no tracks were found in the Kara Sea and areas to the east. Limiting factors: From January to May, if SIC ≥ 90% and SIT ≥ 50cm, ships cannot sail eastward.

[0039] S420: Speed ​​characteristic analysis, drawing speed statistics charts, and statistically analyzing speed distribution across different seasons and months, including: 1) Annual speed distribution: The main range: sections 14-20, accounting for 71.3%; Low-speed range (≤8 knots): accounting for 4.7%, mainly concentrated in September and October; High-speed sections (>20 knots): accounted for 1.9%, occurring only in open waters during the summer.

[0040] 2) Seasonal differences: Summer (July-October): The distribution shows a bimodal pattern, with the low-speed peak (≤8 sections) accounting for 8.2% and the medium-to-high-speed peak (16-18 sections) accounting for 32.5%. Winter (November to June of the following year): Unimodal distribution, with 42.3% of the nodes being between 14 and 16, and 1.5% being between 14 and 16.

[0041] S500: Environmental impact quantification analysis, constructing a three-level quantification model including collinearity diagnosis, principal component analysis dimensionality reduction, and biregression model comparison, quantifying the impact of SIC, SIT, and I10wg on flight speed. The corresponding flowchart for environmental impact quantification analysis is attached. Figure 6 Environmental impact quantification models, such as Figures 7-10 The red line represents the average predicted speed by the random forest model when the environmental variables on the horizontal axis take specific values, while keeping other environmental variables constant.

[0042] S510: Diagnosis of multicollinearity: The variance inflation factor (VIF) test is used to determine the multicollinearity of three environmental variables. The formula and judgment criteria are as follows: VIF calculation formula: For each independent variable X j ( j = 1: SIC , 2: SIT , 3: I 10 wg ), constructing an auxiliary regression model Calculate the coefficient of determination R 2 j ,but The smaller the VIF value, the weaker the collinearity; VIF < 5 indicates no significant collinearity; 5 ≤ VIF ≤ 10 indicates moderate collinearity; VIF > 10 indicates severe collinearity.

[0043] S520: Principal Component Analysis (PCA) Dimensionality Reduction: For months exhibiting collinearity (VIF ≥ 5), PCA is used to extract principal components, resolving the collinearity between SIC and SIT. The formula and steps are as follows: Data standardization: SIC, SIT, and I10wg are standardized to eliminate the influence of dimensions and obtain standardized variables. , in, μ j for X j The mean, σ j for X j The standard deviation.

[0044] Covariance matrix calculation: Constructing a 3×3 covariance matrix for standardized variables R： , Among them, elements r ij for Z i and Z j The correlation coefficient.

[0045] S530: Regression Model Construction and Comparison: OLS linear regression and random forest nonlinear regression models were constructed respectively to quantify the impact of environmental factors on speed. Ordinary Least Squares (OLS) Linear Regression: If the multicollinearity problem is not significant, a linear regression model can be constructed directly using the original environmental variables: , If multicollinearity exists, use principal components after dimensionality reduction via PCA to construct a linear regression model: , in, ε The random error is represented by the following variables: speed (ship speed), SIC (sea ice concentration), SIT (sea ice thickness), and I10wg (wind field intensity at a height of 10 meters). PC1, PC2, and PC3 are principal components obtained by dimensionality reduction through PCA (principal component analysis), which are linear combinations of SIC, SIT, and I10wg. Each principal component represents the main direction of variation of the original variable. It is a constant, representing the baseline value of the speed when the independent variable is 0; , , These are the regression coefficients, corresponding to the weights of each principal component.

[0046] The linear regression model estimates the regression coefficients using the least squares method, minimizing the sum of squared residuals: ,in Let i be the actual speed value of the i-th sample. Let be the predicted speed value of the i-th sample, and n be the total number of samples. is the regression coefficient.

[0047] After the model is built, through t The significance of each regression coefficient was evaluated by testing. F The significance of the overall model is evaluated, and the adjusted goodness of fit is calculated. R 2 The goodness of fit of the model was evaluated using values. Furthermore, the basic assumptions of the model were tested through residual analysis, including the normality, homoscedasticity, and independence of the residuals.

[0048] Random Forest (RF) Nonlinear Regression: Set model parameters, including the number of decision trees (n_estimators), the maximum depth of each tree (max_depth), and the minimum number of samples for node splits (min_samples_split). Perform bootstrap sampling from the original training set to generate n_estimators subsets. Construct a decision tree for each subset, randomly selecting some features at each node split, and choosing the optimal feature for splitting. For a new prediction sample, input it into all decision trees and take the average as the final prediction result, as shown below: , in, B It is the number of decision trees. T b (x ) is the first b Decision trees for samples x The predicted values ​​of the random forest model for the input samples x The final predicted value.

[0049] By comparing a biregression model using linear (OLS) and nonlinear (random forest) regression, we can verify whether the impact of environmental factors on ship speed is linear or nonlinear. We then compare the fitting performance and explanatory power of the two models to select the model more suitable for ship speed prediction in polar waterways.

[0050] This invention collects and extracts ship navigation data, sea ice data, and meteorological data in polar waters based on ship AIS data, sea ice databases, and climate databases. It formulates polar-specific AIS cleaning rules, employs cubic spline interpolation to ensure track smoothness, and achieves a DP thinning compression rate of approximately 95%. Based on the timestamps of AIS and environmental data, it solves the coordinate system adaptation problem through Proj4 projection, achieving multi-source fusion of SIC, SIT, I10wg, and AIS data. Combining OLS and random forest models, it quantifies the impact of environmental factors on ship speed, capturing nonlinear characteristics such as sudden speed drops and the U-shaped effect of winter winds. This method enables in-depth research into ship navigation behavior in polar waters, helping crew members take targeted navigation measures in different scenarios, improving navigation safety, and providing direct support for LNG transportation in polar regions such as the Yamal LNG project. It also supports the addition of environmental factors such as sea ice temperature and visibility, and the model modules can be independently replaced to adapt to the future commercial development needs of polar routes.

[0051] This embodiment also relates to a method for analyzing the behavior of polar vessels considering environmental factors. The overall process includes multi-source data acquisition and preprocessing, AIS data processing, multimodal data fusion, navigation behavior characteristic analysis, and environmental impact quantification analysis. Each step is interconnected, forming a complete technology chain from data to conclusions. The implementation process of each step is described in detail below with reference to specific embodiments: Step 1: Data Acquisition and Preprocessing (corresponding to S100-S200) Taking the navigation behavior analysis of 22 LNG carriers in polar waters in 2024 as an example, the full-process implementation details of the present invention are explained in detail, including the specific amount of data and processing results (unit: data) for each ship, as detailed in Table 1.

[0052] Table 1 Example 2: Multimodal data fusion (corresponding to S300) By verifying the fused data and cross-validating it with image data such as polar sea ice maps, the spatiotemporal consistency of the fused results was ensured. The fused database covers AIS trajectory data and environmental data of 22 LNG carriers, as detailed in Table 2.

[0053] Table 2 Step 3: Navigation behavior characteristic analysis: Track characteristics: Summer (July-October): The tracks of LNG carriers cover the Kara Sea, Barents Sea and Laptev Sea, with some ships even attempting to shorten the sailing distance by approaching high latitude routes at 80°N; Winter (November-June of the following year): The number of eastward-bound tracks of LNG carriers decreases rapidly, and they are mainly concentrated west of the Kara Sea.

[0054] Speed ​​Characteristics: Based on the seasonal characteristics of polar regions, the year can be divided into two main periods, summer and winter, for analysis, as detailed in Table 3. The speed distribution of LNG carriers in winter months exhibits significant seasonal characteristics; during the summer navigation window, the speed distribution of LNG carriers shows a more dispersed and diverse pattern, exhibiting a bimodal characteristic.

[0055] Table 3 Step 4: Quantifying Environmental Impact VIF test and PCA dimensionality reduction: To assess the degree of multicollinearity among the three environmental variables SIC, SIT and I10wg, VIF tests were performed on the data from January to December. The monthly VIF values ​​of the environmental variables are detailed in Table 4.

[0056] Table 4 To eliminate potential multicollinearity and simplify the model structure, principal component analysis (PCA) was performed on the three environmental variables for all months. Principal components were extracted and their variance contribution rates were calculated. The variance explained by the three principal components and the cumulative explained by the first two principal components for each month, as well as the monthly PCA explanation rates for environmental variables, are detailed in Table 5.

[0057] Table 5 OLS Linear Regression: Based on the principal components after PCA dimensionality reduction, an OLS linear regression model for 12 months was constructed to analyze the linear relationship between the principal components and the speed of LNG carriers. Table 6 shows the results of the linear regression model for each month, including... And the regression coefficients and significance of each principal component. p value.

[0058] Table 6 Random Forest (RF) Nonlinear Regression: The results of the RF nonlinear regression are shown in Table 7, and the importance of environmental factors on speed in the RF model is shown in Table 8.

[0059] Table 7 Table 8 The electronic device of this invention includes a central processing unit (CPU), which can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) or loaded from storage units into random access memory (RAM). The RAM may also store various programs and data required for device operation. The CPU, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0060] Multiple components in the device are connected to the I / O interface, including: input units such as keyboards and mice; output units such as various types of displays and speakers; storage units such as disks and optical discs; and communication units such as network interface cards (NICs), modems, and wireless transceivers. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0061] The processing unit performs the various methods and processes described above. For example, in some embodiments, the methods may be implemented as computer software programs tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on the device via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by the CPU, one or more steps of the methods described above may be performed. Alternatively, in other embodiments, the CPU may be configured to execute the methods by any other suitable means (e.g., by means of firmware).

[0062] 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-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0063] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0064] In the context of this invention, 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. Machine-readable media can include, but are 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 fibers, portable compact disk read-only memory, optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0065] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for analyzing the behavior of polar navigation vessels considering environmental factors, characterized in that, The method includes: Multi-source data acquisition and preprocessing: Multi-source data including sea ice, wind speed and AIS data from polar waters were acquired and preprocessed; the sea ice data included sea ice concentration data and sea ice thickness data. Multimodal data fusion: Spatiotemporal matching of AIS data with sea ice and wind speed data to obtain a multimodal dataset containing ship identification information, location, speed and environmental elements, solving the polar coordinate system adaptation problem; Navigation behavior characteristic analysis: Combining multimodal datasets, we analyze the characteristics of ship navigation behavior in polar waters, including the seasonal spatial distribution pattern of ship trajectories and the seasonal variation distribution characteristics of ship speed. Environmental impact quantification analysis: Construct a three-level quantitative model that includes collinearity diagnosis, principal component analysis dimensionality reduction, and biregression model comparison to capture the linear and nonlinear effects of environmental factors on speed, and achieve dynamic quantification of the impact of environmental factors on speed.

2. The method for analyzing the behavior of polar navigation vessels considering environmental factors according to claim 1, characterized in that, The spatiotemporal matching process includes: Coordinate system transformation: The Proj4 library is used to implement the coordinate system transformation between AIS data's WGS84 latitude and longitude and polar stereo projection; Spatiotemporal matching: A spatiotemporal matching of AIS data with sea ice and wind speed data is achieved through time alignment and spatial interpolation.

3. The method for analyzing the behavior of polar navigation vessels considering environmental factors according to claim 2, characterized in that, The Proj4 string for the coordinate system transformation is: + proj = stere + lat 0 = 90 + lat ts = 70 + lon 0 = -45 + x 0 = 0 + y 0= 0 + datum = WGS 84 + units = m + no defs ， Among them, + proj = stere The projection type is stereoscopic projection; lat 0 = 90: Latitude of the projection origin; lat ts = 70: Standard latitude; lon 0 = -45: Longitude of the central meridian; x 0 = 0, y 0 = 0: Offset of projection origin coordinates; datum = WGS 84: Reference ellipsoid; units = m The coordinate unit after projection is meters.

4. The method for analyzing the behavior of polar navigation vessels considering environmental factors according to claim 2, characterized in that, The coordinate system transformation maps AIS data from spherical coordinates to planar coordinates. The specific transformation formula is as follows: ， The mapping coefficient R from spherical coordinates to planar coordinates is calculated by the following formula: ， Where (x, y) are planar coordinates; λ Longitude; λ 0 = -45° is the longitude of the central meridian; R 0 represents the Earth's radius; φ Latitude φ ts The latitude is the standard parallel.

5. The method for analyzing the behavior of polar navigation vessels considering environmental factors according to claim 2, characterized in that, The spatiotemporal matching mentioned above includes: Time matching: AIS data and wind speed data are matched at hourly resolution; AIS data and sea ice data are matched at daily resolution. Spatial matching: Linear interpolation is used to map the transformed AIS planar coordinates to the grid of sea ice data to obtain the sea ice data of the corresponding grid; based on the AIS WGS84 latitude and longitude, the wind speed data of the corresponding grid is obtained through nearest neighbor interpolation.

6. The method for analyzing the behavior of polar navigation vessels considering environmental factors according to claim 1, characterized in that, Based on the multimodal dataset, we plotted the ship trajectories throughout the year and statistically analyzed the seasonal spatial distribution patterns of the ship trajectories; based on the multimodal dataset, we plotted the ship speed statistics and statistically analyzed the seasonal differences in the distribution characteristics of the speed.

7. The method for analyzing the behavior of polar navigation vessels considering environmental factors according to claim 1, characterized in that, In the process of quantitative analysis of environmental impact, the collinearity diagnosis includes: using the variance inflation factor to test the collinearity of the three environmental variables. If the calculated collinearity value is less than 5, it is judged that there is no obvious collinearity; otherwise, it is judged that collinearity exists. The three environmental variables mentioned above include sea ice concentration, sea ice thickness, and wind speed.

8. The method for analyzing the behavior of polar navigation vessels considering environmental factors according to claim 7, characterized in that, The dual regression model includes the OLS linear regression model and the random forest nonlinear regression model; If there is no obvious collinearity, the original environmental variables are used directly to construct an OLS linear regression model; If collinearity exists, use principal component analysis to extract principal components for dimensionality reduction and construct an OLS linear regression model; By comparing the results of the two regression models, the model that is more suitable for predicting the speed of ships in polar waterways is selected as the final regression model.

9. The method for analyzing the behavior of polar navigation vessels considering environmental factors according to claim 1, characterized in that, The method further includes: before multimodal data fusion, performing data cleaning, cubic spline interpolation of natural boundary conditions, and data thinning on the AIS data.

10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 9.

Images