A ship trajectory optimization method based on multi-station radar fusion signals
By screening and completing the frequency inconsistency problem in multi-station radar data, the data time and space consistency is ensured, and the problem of inaccurate ship trajectory in the existing technology is solved and navigation safety is improved.
Patent Information
- Application Number
- CN202410994936.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-24
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-07-24
AI Technical Summary
The prior art cannot effectively deal with the problem of inconsistent frequency of multi-station radar data acquisition, resulting in inaccurate trajectory position and trajectory of ships, which may cause navigation safety threats.
By comparing the data acquisition frequencies of different radar sites, filtering out data with inconsistent frequencies, and using time alignment algorithm and Kriging interpolation algorithm to complete data, ensuring data spatiotemporal consistency and trajectory integrity.
It effectively prevents time and space misalignment caused by different data acquisition frequency, improves the accuracy and stability of data fusion, and ensures the reliability and safety of ship trajectory optimization.
Smart Images

Figure CN118731939B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of ship trajectory optimization, and in particular to a ship trajectory optimization method based on multi-station radar fusion signals. Background Art
[0002] Ship trajectory optimization based on multi-station radar fusion signals is a method of using ship signal data collected by multiple radar sites, fusing these data for analysis and processing to optimize the ship's trajectory. First, the data collected by multiple radar sites can provide more comprehensive and accurate information on the ship's position and movement. By fusing these data, the limitations and errors that may exist in a single radar site can be eliminated, thereby more accurately determining the ship's position and trajectory. Secondly, by analyzing and processing the fused data, the ship's trajectory can be optimized, including path planning, navigation safety assessment, etc., to improve the ship's navigation efficiency and safety.
[0003] The prior art has the following deficiencies:
[0004] Before fusing the collected data, the existing technology cannot filter out the site information with inconsistent data collection frequency. If the data with inconsistent data collection frequency is fused, the fused data may be misaligned in time and space, which will seriously affect the accuracy of the ship's position and trajectory. Due to the different data collection frequencies of different sites, some data may be relatively lagging or ahead in time, so that they cannot be accurately aligned during the fusion process, and the fused data cannot reflect the true ship's position and trajectory information. This position offset and inaccurate trajectory overlap may pose a serious threat to the navigation safety of the ship, greatly increasing the risk of collision between the ship and other ships or obstacles, and may even cause major accidents;
[0005] When fusing the collected data, if data is lost at certain moments in the fusion process, it may cause the ship's trajectory to be interrupted or discontinuous, causing the ship's position information to appear faulty in the fused trajectory, making it difficult to accurately reflect the ship's actual movement path, which may also pose a serious threat to the ship's navigation safety.
[0006] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention
[0007] The purpose of the present invention is to provide a ship trajectory optimization method based on multi-station radar fusion signals. By comparing the data acquisition frequencies of different radar sites, the system can identify data with inconsistent frequencies and screen them out. This process involves precise matching of data timestamps to ensure that data from each site can be accurately aligned when fused, which can effectively prevent the time and space misalignment caused by different data acquisition frequencies, thereby ensuring the spatiotemporal consistency of the fused data, providing a reliable basis for subsequent trajectory optimization, and solving the problems in the above-mentioned background technology.
[0008] In order to achieve the above object, the present invention provides the following technical solution: a ship trajectory optimization method based on multi-station radar fusion signal, comprising the following steps:
[0009] Collect ship information from multiple radar sites. The data provided by different radar sites need to cover the target area and overlap in time and space.
[0010] Preprocess the collected data;
[0011] Before data fusion, conduct real-time analysis on whether the data collection frequencies of multiple sites are consistent, and filter out data with inconsistent data collection frequencies;
[0012] After filtering out data with inconsistent data collection frequencies, the data from different radar sites are fused to reflect the position and trajectory of the ship;
[0013] Monitor the fusion process of the screened data in real time to identify whether there is an abnormal risk of data fusion loss. When the abnormal risk is perceived, mark the fused data;
[0014] For the marked data, the Kriging interpolation algorithm is used to reconstruct the lost data and fill in the data at risk of data loss.
[0015] Preferably, the steps of performing real-time analysis on whether the data collection frequencies of multiple sites are consistent and filtering out data with inconsistent data collection frequencies are as follows:
[0016] For the data collected by each radar station, the time interval between adjacent data points is calculated, and then, based on the difference in timestamps, the time offset of each station relative to the reference station is calculated;
[0017] Select one of several radar sites as a reference site;
[0018] Use the time alignment algorithm to align the data of other sites to the time axis of the reference site based on the calculated time offset;
[0019] After time alignment, data points with inconsistent time intervals are marked as abnormal data points based on the timestamp of the data and the aligned time axis, and the marked data points are then filtered out.
[0020] Preferably, the specific steps of using the time alignment algorithm to align the data of other sites to the time axis of the reference site according to the calculated time offset are:
[0021] Step 1: Establish a least squares equation, taking the time offset as the independent variable, so that the residual sum of squares in the equation is minimized;
[0022] Step 2: Use the least square method to find the time offset that minimizes the residual.
[0023] Step 3: Apply the solved time offset to the data of other sites. For the data of each site, adjust each timestamp according to the corresponding time offset so that the data is aligned with the time axis of the reference site.
[0024] Preferably, the data accuracy loss rate information and clock drift information are obtained when the screened data is fused. The data accuracy loss rate information indicates the proportion of data accuracy loss in the data fusion process, and the clock drift information indicates the degree of deviation between the clocks of different radar sites and the reference clock. After acquisition, the data accuracy loss rate information and the clock drift information are analyzed for abnormalities to generate data accuracy loss rate abnormal incremental quantization values and clock drift quantization values.
[0025] Preferably, the logic for obtaining the abnormal incremental quantization value of the data precision loss rate is as follows:
[0026] Collect historical data from multiple sites, including how the rate of data accuracy loss changes over time;
[0027] Perform trend fitting on the change of data precision loss rate over time, and set the fitting function as , the fitting function is expressed as: ,in n is the order of the polynomial, are the coefficients of the fitting function, t Represents time, i is a power;
[0028] Calculate the residual at each time point, that is, the difference between the actual data precision loss rate and the fitting function. Indicates that the calculation formula is: ,in, For time t The actual data accuracy loss rate corresponding to the
[0029] The calculated residuals are compared with the preset residual reference thresholds, and the abnormal incremental quantitative value of the data precision loss rate is calculated by the residuals exceeding the residual reference threshold. , the calculation expression is:
[0030] ,in, N is the length of the time series, indicating the number of time points of the observed data, For time t The weight of the residual.
[0031] Preferably, the logic of obtaining the clock drift quantization value is as follows:
[0032] Get the signal data collected from different radar sites. For each pair of radar sites k and j , calculate the phase difference of the signals received by the two stations, assuming the phase difference is , express t Time Station k and Site j The phase difference;
[0033] Perform coherent integration on the phase difference and integrate it into a time window. Let the phase difference after integration be , Represents the phase difference within the time window, then the expression for coherent integration of the phase difference is:
[0034] ,in, w is the width of the time window, indicating the time span of the integration;
[0035] The phase difference obtained by phase integration is used to calculate the clock drift between the two radar sites. k and Site j The clock drift between , then the site k and Site j Clock drift between The calculation expression is:
[0036] ,in, is the center frequency of the radar signal;
[0037] Perform a weighted summation of the clock drifts between all sites and calculate the quantized value of the clock drift. Let the quantized value of the clock drift be , then the clock drift quantization value The calculation expression is:
[0038] ,in, is the weight between corresponding sites.
[0039] Preferably, the abnormal incremental quantitative value of the data precision loss rate generated when the screened data is fused is obtained. and clock drift quantization value After that, a data analysis model is established to generate the fusion loss risk assessment coefficient , by integrating the loss hidden danger assessment coefficient Intelligently perceive abnormal risks during data fusion.
[0040] Preferably, the fusion loss hidden danger assessment coefficient generated when the screened data is fused is compared and analyzed with a pre-set fusion loss hidden danger assessment coefficient reference threshold; if the fusion loss hidden danger assessment coefficient is greater than or equal to the fusion loss hidden danger assessment coefficient reference threshold, the data generated by the fusion at the corresponding moment will be marked; if the fusion loss hidden danger assessment coefficient is less than the fusion loss hidden danger assessment coefficient reference threshold, the data generated by the fusion at the corresponding moment will not be marked.
[0041] Preferably, for the marked data, the specific steps of filling the data with the risk of data loss are as follows:
[0042] Step 1: Determine the area to be interpolated and set the parameters required for the Kriging interpolation algorithm;
[0043] Step 2: Use the set parameters to execute the Kriging interpolation algorithm to interpolate the marked data points;
[0044] Step 3: Evaluate the interpolated data to check whether the interpolation results are reasonable;
[0045] Step 4: Fill the data processed by Kriging interpolation into the original data to replace the marked data.
[0046] In the above technical solution, the technical effects and advantages provided by the present invention are:
[0047] By comparing the data collection frequencies of different radar sites, the present invention enables the system to identify data with inconsistent frequencies and screen them out. This process involves precise matching of the data timestamps to ensure that the data of each site can be accurately aligned during fusion, which can effectively prevent the time and space misalignment caused by different data collection frequencies, thereby ensuring the spatiotemporal consistency of the fused data and providing a reliable basis for subsequent trajectory optimization.
[0048] The present invention continuously monitors whether there is any abnormal situation of data loss during the data fusion process. Once an abnormality is found, the system will immediately mark and record it. This real-time monitoring mechanism enables the system to promptly discover and respond to data loss, and then take remedial measures such as interpolation and completion to ensure the integrity and continuity of trajectory data. This intelligent monitoring and response mechanism effectively improves the reliability and stability of the data fusion process and provides reliable data support for ship trajectory optimization. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.
[0050] Figure 1 The present invention is a method flow chart of a ship trajectory optimization method based on multi-station radar fusion signals. DETAILED DESCRIPTION
[0051] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.
[0052] The present invention provides Figure 1 A ship trajectory optimization method based on multi-station radar fusion signals is shown, comprising the following steps:
[0053] Collect ship information from multiple radar sites. The data provided by different radar sites need to cover the target area and overlap in time and space.
[0054] The ship information collected from multiple radar sites usually includes the ship's position, speed, heading, speed, track angle, ship identification information, etc. This information can fully describe the ship's motion status and position at sea, help determine the ship's real-time position, heading trend and running speed, and thus provide an accurate data basis for trajectory optimization.
[0055] The data provided by different radar sites need to cover the target area and overlap in time and space in order to enhance the monitoring and analysis capabilities of ship trajectories. By covering the target area and ensuring the diversity of data sources, more comprehensive ship information can be obtained, and blind spots and missed reports can be reduced. The overlap in time and space can verify and calibrate the consistency of data from different radar sites, improve the reliability and accuracy of the data, and allow better tracking of ship movements, achieving more accurate ship trajectory reconstruction and optimization, thereby improving navigation safety and efficiency.
[0056] Preprocess the collected data;
[0057] Preprocessing includes operations such as data cleaning, denoising, and coordinate system conversion. Preprocessing the collected data can ensure high data quality and reduce errors and inaccuracies in the subsequent fusion process.
[0058] Before data fusion, conduct real-time analysis on whether the data collection frequencies of multiple sites are consistent, and filter out data with inconsistent data collection frequencies;
[0059] The steps to conduct real-time analysis on whether the data collection frequencies of multiple sites are consistent and filter out data with inconsistent data collection frequencies are as follows:
[0060] For the data collected by each radar station, the time interval between adjacent data points is calculated. Then, based on the difference in timestamps, the time offset of each station relative to the reference station (usually the station with the most complete and highest quality data) is calculated;
[0061] Select one of several radar sites as a reference site;
[0062] The selection of reference sites usually needs to consider the location layout and coverage of the sites, and select those that can cover the target area and have a certain overlap with other sites to ensure the comprehensiveness and reliability of the data.
[0063] Use the time alignment algorithm to align the data of other sites to the time axis of the reference site based on the calculated time offset;
[0064] Using the time alignment algorithm, the specific steps to align the data of other sites to the time axis of the reference site according to the calculated time offset are as follows:
[0065] Step 1: Establish a least squares equation, taking the time offset as the independent variable, so that the residual sum of squares in the equation is minimized;
[0066] The goal of the least squares method is to find the time offset that minimizes the residual, thereby achieving time alignment.
[0067] Step 2: Use the least square method to find the time offset that minimizes the residual.
[0068] The least squares solution method includes matrix inversion or gradient descent.
[0069] Step 3: Apply the solved time offset to the data of other sites. For the data of each site, adjust each timestamp according to the corresponding time offset so that the data is aligned with the time axis of the reference site.
[0070] After time alignment, data points with inconsistent time intervals are marked as abnormal data points based on the timestamp of the data and the aligned time axis, and the marked data points are then filtered out.
[0071] After filtering out data with inconsistent data collection frequencies, the data from different radar sites are fused to reflect the position and trajectory of the ship;
[0072] Fusion of data from different radar sites, which may involve time synchronization, spatial matching, data interpolation and other technologies, to ensure that the fused data can accurately reflect the position and trajectory of the ship. Fusion algorithms usually take into account factors such as performance differences, error characteristics and data quality of each radar site to maximize the accuracy and reliability of the data;
[0073] Fusion algorithms are a key step in integrating and fusing data from different radar sites to ensure that the fused data can accurately reflect the position and trajectory of the ship. The following are some common fusion algorithms:
[0074] Weighted average method: The weighted average method is a simple and effective fusion algorithm that performs weighted average on the data collected by different radar sites. The weight can be determined based on site performance, error characteristics or data quality. It is suitable for situations where the data quality is relatively consistent and the errors of each radar site are small.
[0075] Kalman filter: Kalman filter is a recursive filtering algorithm that fuses data from multiple radar stations through prediction and correction. It is suitable for situations where dynamic changes in data, noise characteristics, and system models need to be considered, such as ship motion state estimation and prediction.
[0076] Particle filter: Particle filter is a filtering algorithm based on the Monte Carlo method, which estimates the posterior distribution of the state through a series of particles. It is suitable for nonlinear systems and highly uncertain situations, such as ship trajectory tracking and target tracking.
[0077] Rule-based fusion algorithm: The rule-based fusion algorithm integrates and processes data from different sites through pre-defined rules or logic. It is suitable for simple scenarios and specific problem areas, such as navigation safety assessment and route planning.
[0078] Neural network fusion algorithm: The neural network fusion algorithm uses a neural network model to learn the complex relationship between data from different radar sites and perform data fusion and prediction. It is suitable for processing high-dimensional data and complex nonlinear relationships, such as ship dynamic modeling and prediction.
[0079] In summary, the choice of fusion algorithm depends on the specific application scenario, data characteristics and system requirements. Different algorithms have different advantages and applicability. The most appropriate algorithm can be selected according to the actual situation to achieve data fusion to improve the accuracy and reliability of ship trajectory optimization.
[0080] Monitor the fusion process of the screened data in real time to identify whether there is an abnormal risk of data fusion loss. When the abnormal risk is perceived, mark the fused data;
[0081] The data accuracy loss rate information and clock drift information are obtained when the screened data is fused. The data accuracy loss rate information indicates the proportion of data accuracy loss in the data fusion process, and the clock drift information indicates the degree of deviation between the clocks of different radar sites and the reference clock. After acquisition, the data accuracy loss rate information and clock drift information are analyzed for abnormalities to generate data accuracy loss rate abnormal incremental quantization values and clock drift quantization values.
[0082] Multi-station radar systems are usually equipped with sophisticated sensors that can measure information such as the ship's position, speed, and direction. By processing and analyzing the sensor measurement data, the data accuracy loss rate during the ship trajectory optimization process can be calculated.
[0083] Multi-station radar systems require time synchronization to ensure that the data from each radar site has a consistent time reference. Through time synchronization technology, clock drift information between different radar sites can be obtained.
[0084] The abnormal increase of data precision loss rate may cause data loss in data fusion. First, the abnormal increase of data precision loss rate means that the precision loss of data in transmission, processing or other links increases, which may reduce the quality of data. If the data quality drops to a certain extent, it may cause data loss, because low-quality data may be regarded as invalid or unreliable by the system and thus be discarded or ignored. Secondly, the abnormal increase of data precision loss rate may cause the fused data to become unreliable or inaccurate. When the fusion algorithm encounters data with large precision loss when processing data, it may cause the algorithm to produce wrong results or large deviation estimates, thereby affecting the accuracy of data fusion. If the quality of fused data is poor, it may cause data loss or errors in the fusion results. Finally, the abnormal increase of data precision loss rate may increase the difficulty of data processing. When the data precision loss rate is high, more complex data processing algorithms or more data correction measures may be required to process the data to minimize the precision loss of the data. This increases the complexity and computational complexity of data processing, which may increase the risk of data loss in the data fusion process.
[0085] The logic for obtaining the abnormal incremental quantization value of the data precision loss rate is as follows:
[0086] Collect historical data from multiple sites, including how the rate of data accuracy loss changes over time;
[0087] In the process of ship trajectory optimization, historical data from multiple radar sites need to be collected. These data include the change of data accuracy loss rate over time, that is, the value of data accuracy loss rate recorded at different time points for each site. The collected historical data can help analyze the change trend of ship trajectory, understand the data accuracy loss, and provide a reference for the optimization process.
[0088] Perform trend fitting on the change of data precision loss rate over time, and set the fitting function as , the fitting function is expressed as: ,in n is the order of the polynomial, are the coefficients of the fitting function, t Represents time, i is a power;
[0089] The polynomial fitting method approximates the relationship between data by fitting the data into the form of a polynomial function. The term represents time t of i The power means that the polynomial fitting function contains different powers of time. The coefficients of these terms are determined through the fitting process, thus describing the trend of data changing over time. Therefore, the fitting function It can show the trend of data accuracy loss rate over time.
[0090] Calculate the residual at each time point, that is, the difference between the actual data precision loss rate and the fitting function. It means that the calculation formula is: ,in, For time t The actual data accuracy loss rate corresponding to the
[0091] The calculated residuals are compared with the preset residual reference thresholds, and the abnormal incremental quantitative value of the data precision loss rate is calculated by the residuals exceeding the residual reference threshold. , the calculation expression is:
[0092] ,in, N is the length of the time series, indicating the number of time points of the observed data, For time t The weight of the residual at (the residual exceeding the residual reference threshold);
[0093] From the calculation expression of the abnormal incremental quantization value of the data precision loss rate, it can be seen that the larger the performance value of the abnormal incremental quantization value of the data precision loss rate generated when the screened data is fused, the greater the risk of data loss when the screened data is fused. Conversely, the smaller the risk of data loss when the screened data is fused.
[0094] Clock drift refers to the time error or offset generated by the clocks of different devices or systems during operation. In the process of ship trajectory optimization, when the screened data is fused, when there is clock drift between different radar sites, data fusion may cause data loss anomalies. First, when the clocks of different radar sites drift, the time recorded by each site will be inconsistent. In the process of data fusion, if the time synchronization problem is not effectively solved, the data collected by different sites will not be accurately aligned, which may cause data loss or misalignment. Secondly, clock drift may cause incoherence in the ship trajectory data collected by different radar sites. For example, if a ship is detected by a radar site at a certain point in time, there may be a time offset in the record of another radar site, resulting in inconsistent trajectory data, which in turn affects the accuracy of data fusion. Finally, clock drift may increase the risk of data loss. When the data fusion algorithm cannot correctly handle the clock drift between different radar sites, it may cause inconsistent data collection frequency, resulting in partial data loss. Especially in the case of high-speed movement, clock drift may cause data collection to be untimely or miss important ship information, thus affecting the quality of data fusion.
[0095] The logic for obtaining the clock drift quantization value is as follows:
[0096] Get the signal data collected from different radar sites. For each pair of radar sites k and j , calculate the phase difference of the signals received by the two stations, assuming the phase difference is , express t Time Station k and Site j The phase difference;
[0097] By calculating the phase difference between the signals received at the two sites, the clock differences between the different radar sites can be calculated.
[0098] Perform coherent integration on the phase difference and integrate it into a time window. Let the phase difference after integration be , Represents the phase difference within the time window, then the expression for coherent integration of the phase difference is:
[0099] ,in, w is the width of the time window, indicating the time span of the integration;
[0100] The phase difference obtained by phase integration is used to calculate the clock drift between the two radar sites. k and Site j The clock drift between , then the site k and Site j Clock drift between The calculation expression is:
[0101] ,in, is the center frequency of the radar signal;
[0102] Since the phase difference is in radians, we can calculate the phase difference by dividing by the center frequency. Converted to seconds, the center frequency of a radar signal refers to the location in the spectrum where the main frequency components or energy of the signal are concentrated. It is the reference frequency used to modulate and demodulate signals in radar systems, and determines the frequency range and characteristics of the signal. In radar systems, the center frequency is usually determined by the design and configuration of the radar equipment, and plays a key role in signal processing and data analysis, affecting the signal's resolution, distance measurement accuracy, and the ability to detect and identify targets.
[0103] Perform a weighted summation of the clock drifts between all sites and calculate the quantized value of the clock drift. Let the quantized value of the clock drift be , then the clock drift quantization value The calculation expression is:
[0104] ,in, is the weight between corresponding sites;
[0105] The weights between corresponding sites can be determined based on information such as the distance between the sites and signal strength. Generally, the closer the distance and the stronger the signal strength, the higher the weight of the site, and vice versa.
[0106] From the calculation expression of the clock drift quantization value, it can be seen that the larger the performance value of the clock drift quantization value generated when the screened data is fused, the greater the risk of data loss when the screened data is fused. Conversely, the smaller the risk of data loss when the screened data is fused.
[0107] The abnormal incremental quantitative value of the data precision loss rate generated when the screened data is obtained for fusion and clock drift quantization value After that, a data analysis model is established to generate the fusion loss risk assessment coefficient , by integrating the loss hidden danger assessment coefficient Intelligently perceive abnormal risks during data fusion.
[0108] The specific implementation method of the above data analysis model is not specifically limited here, and can achieve the quantification of abnormal increment of data precision loss rate and clock drift quantization value Conduct comprehensive analysis and generate fusion loss risk assessment coefficients In order to realize the technical solution of the present invention, the present invention provides a specific implementation method;
[0109] Fusion loss risk assessment coefficient The resulting calculation formula is: , where , They are the abnormal increment quantization values of data precision loss rate and clock drift quantization value The preset scaling factor of , Both are greater than 0.
[0110] It can be seen from the calculation formula that the larger the performance value of the abnormal incremental quantization value of the data accuracy loss rate generated when the screened data is fused, and the larger the performance value of the clock drift quantization value, the greater the risk of data loss when the screened data is fused. Conversely, the smaller the risk of data loss when the screened data is fused.
[0111] The fusion loss hidden danger assessment coefficient generated when the screened data is fused is compared and analyzed with the pre-set fusion loss hidden danger assessment coefficient reference threshold. If the fusion loss hidden danger assessment coefficient is greater than or equal to the fusion loss hidden danger assessment coefficient reference threshold, the data generated by the fusion at the corresponding moment is marked; if the fusion loss hidden danger assessment coefficient is less than the fusion loss hidden danger assessment coefficient reference threshold, the data generated by the fusion at the corresponding moment is not marked;
[0112] If the fusion loss hidden danger assessment coefficient is greater than or equal to the fusion loss hidden danger assessment coefficient reference threshold when the screened data is fused, it indicates that there is a risk of data loss when the screened data is fused. In this case, the data generated by the fusion at the corresponding moment will be marked to facilitate the distinction of the data with the risk of data loss. If the fusion loss hidden danger assessment coefficient is less than the fusion loss hidden danger assessment coefficient reference threshold when the screened data is fused, it indicates that the screened data can be efficiently fused.
[0113] For the marked data, the Kriging interpolation algorithm is used to reconstruct the lost data and fill in the data at risk of data loss;
[0114] For the marked data, the specific steps to fill the data at risk of data loss are as follows:
[0115] Step 1: Determine the area to be interpolated and set the parameters required for the Kriging interpolation algorithm;
[0116] Determining the area that needs to be interpolated refers to the area where the marked data points are located during the data fusion process. Before executing the interpolation algorithm, it is necessary to clarify the interpolation range, that is, to determine the spatial range that the interpolation operation should cover. This range usually consists of unknown data points around known data points to ensure the accuracy and rationality of the interpolation results.
[0117] Determining the parameters required for the Kriging interpolation algorithm means that when using the Kriging method for spatial interpolation, some parameters need to be set to control the interpolation results. These parameters include:
[0118] Range: Range refers to the typical distance range between adjacent data points in space, that is, the correlation distance of the semivariogram. It determines the extent to which the interpolation result is affected by spatial correlation. A larger range means that the spatial correlation between data points is stronger, and the interpolation result is affected by neighboring points more. The choice of range is usually based on the spatial variation of actual data and empirical adjustments.
[0119] Semivariogram type: The semivariogram describes the model in which the degree of variation between samples changes as the distance increases. Common semivariogram types include exponential, Gaussian, spherical, etc. Choosing the appropriate semivariogram type depends on the spatial characteristics of the sample data. Different types of semivariograms reflect different spatial correlation patterns.
[0120] Other parameters: In addition to the range and semivariogram type, you may also need to set other parameters, such as noise variance, spatial distribution of data points, etc. The setting of these parameters will also affect the accuracy and stability of the interpolation results.
[0121] Step 2: Use the set parameters to execute the Kriging interpolation algorithm to interpolate the marked data points;
[0122] The specific steps of executing the Kriging interpolation algorithm to interpolate the marked data points include: first, determine the range of the interpolation area, that is, take the data point to be interpolated as the center and determine the range of the interpolation area according to the set interpolation radius; second, calculate the spatial correlation between the data points according to the spatial distribution of the known data points in the interpolation area; then, estimate the value of the unknown data point according to the spatial correlation and the set semivariogram function type, and calculate its interpolation error; finally, interpolate the unknown data points according to the obtained interpolation results and errors to obtain the final interpolation results.
[0123] In the Kriging interpolation algorithm, calculating the spatial correlation between data points means evaluating the spatial variation or correlation between data points. This correlation tells us the spatial structure between data points, that is, the trend or correlation between data points. In spatial statistics, the semivariogram is usually used to describe the spatial correlation between data points. The semivariogram measures the variation of data values at different distances. The higher the correlation value, the more similar the spatial changes between data points, and vice versa. The changes are less correlated. In Kriging interpolation, the data values at unknown locations are determined by evaluating spatial correlation, so as to achieve interpolation estimation of the data.
[0124] Step 3: Evaluate the interpolated data to check whether the interpolation results are reasonable;
[0125] After interpolation, the interpolated data needs to be evaluated to ensure the accuracy and reliability of the interpolation results. There are two ways to evaluate the interpolation results: comparison with the original data and consideration of the physical meaning of the data. First, comparison with the original data refers to comparing the interpolated data with the original data to see whether they are consistent with the original data. If the interpolation results are consistent with the original data in terms of spatial distribution, trend, etc., and the difference is not large, the interpolation results are relatively reliable. Secondly, considering the physical meaning of the data means verifying the interpolation results in a physical sense based on the problem background and domain knowledge. For example, for trajectory data in the ocean field, the impact of environmental factors such as ocean currents and wind direction on the trajectory can be considered to check whether the interpolation results are consistent with these factors.
[0126] Evaluation through the above methods can help determine the accuracy and reliability of the interpolation results, thereby ensuring the effectiveness of subsequent data analysis and application.
[0127] Step 4: Fill the data processed by Kriging interpolation into the original data to replace the marked data;
[0128] This ensures that if there is lost data during the data fusion process, the lost data points can be reasonably filled, thereby reducing the impact of data loss on the trajectory optimization results.
[0129] By comparing the data collection frequencies of different radar sites, the present invention enables the system to identify data with inconsistent frequencies and screen them out. This process involves precise matching of the data timestamps to ensure that the data of each site can be accurately aligned during fusion, which can effectively prevent the time and space misalignment caused by different data collection frequencies, thereby ensuring the spatiotemporal consistency of the fused data and providing a reliable basis for subsequent trajectory optimization.
[0130] The present invention continuously monitors whether there is any abnormal situation of data loss during the data fusion process. Once an abnormality is found, the system will immediately mark and record it. This real-time monitoring mechanism enables the system to promptly discover and respond to data loss, and then take remedial measures such as interpolation and completion to ensure the integrity and continuity of trajectory data. This intelligent monitoring and response mechanism effectively improves the reliability and stability of the data fusion process and provides reliable data support for ship trajectory optimization.
[0131] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.
[0132] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which 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.
[0133] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.
Claims
1. A ship trajectory optimization method based on multi-station radar fusion signals, characterized in that: The following steps are involved: Collect ship information from multiple radar sites. The data provided by different radar sites need to cover the target area and overlap in time and space. Preprocess the collected data; Before data fusion, conduct real-time analysis on whether the data collection frequencies of multiple sites are consistent, and filter out data with inconsistent data collection frequencies; After filtering out data with inconsistent data collection frequencies, the data from different radar sites are fused to reflect the position and trajectory of the ship; Monitor the fusion process of the screened data in real time to identify whether there is an abnormal risk of data fusion loss. When the abnormal risk is perceived, mark the fused data; For the marked data, the Kriging interpolation algorithm is used to reconstruct the lost data and fill in the data at risk of data loss; The data precision loss rate information and clock drift information are obtained when the screened data is fused. The data precision loss rate information indicates the proportion of data precision loss in the data fusion process, and the clock drift information indicates the degree of deviation between the clocks of different radar sites and the reference clock. After obtaining the data precision loss rate information and the clock drift information, an abnormal analysis is performed on the data precision loss rate information and the clock drift information to generate a data precision loss rate abnormal incremental quantization value and a clock drift quantization value; The logic for obtaining the abnormal incremental quantization value of the data precision loss rate is as follows: Collect historical data from multiple sites, including how the rate of data accuracy loss changes over time; Perform trend fitting on the change of data precision loss rate over time, and set the fitting function as , the fitting function is expressed as: , where n is the order of the polynomial, are the coefficients of the fitting function, t represents time, and i is the power; Calculate the residual at each time point, that is, the difference between the actual data precision loss rate and the fitting function. It means that the calculation formula is: ,in, is the actual data accuracy loss rate corresponding to time t; The calculated residual is compared with the preset residual reference threshold, and the abnormal incremental quantitative value of the data precision loss rate is calculated by the residual exceeding the residual reference threshold. , the calculation expression is: , where N is the length of the time series, indicating the number of time points of the observed data, is the weight of the residual at time t.
2. A ship trajectory optimization method based on multi-station radar fusion signals according to claim 1, characterized in that: The steps to conduct real-time analysis on whether the data collection frequencies of multiple sites are consistent and filter out data with inconsistent data collection frequencies are as follows: For the data collected by each radar station, the time interval between adjacent data points is calculated, and then, based on the difference in timestamps, the time offset of each station relative to the reference station is calculated; Select one of several radar sites as a reference site; Use the time alignment algorithm to align the data of other sites to the time axis of the reference site based on the calculated time offset; After time alignment, data points with inconsistent time intervals are marked as abnormal data points based on the timestamp of the data and the aligned time axis, and the marked data points are then filtered out.
3. A ship trajectory optimization method based on multi-station radar fusion signals according to claim 2, characterized in that: Using the time alignment algorithm, the specific steps to align the data of other sites to the time axis of the reference site according to the calculated time offset are as follows: Step 1: Establish a least squares equation, taking the time offset as the independent variable, so that the residual sum of squares in the equation is minimized; Step 2: Use the least square method to find the time offset that minimizes the residual. Step 3: Apply the solved time offset to the data of other sites. For the data of each site, adjust each timestamp according to the corresponding time offset so that the data is aligned with the time axis of the reference site.
4. The ship trajectory optimization method based on multi-station radar fusion signals according to claim 1 is characterized in that: The logic for obtaining the clock drift quantization value is as follows: Get the signal data collected from different radar sites. For each pair of radar sites k and j, calculate the phase difference of the signals received by the two sites. Let the phase difference be , represents the phase difference between site k and site j at time t; Perform coherent integration on the phase difference and integrate it into a time window. Let the phase difference after integration be , Represents the phase difference within the time window, then the expression for coherent integration of the phase difference is: , where w is the width of the time window, indicating the time span of the integration; The phase difference obtained by phase integration is used to calculate the clock drift between the two radar sites. Suppose the clock drift between site k and site j is , then the clock drift between site k and site j is The calculation expression is: ,in, is the center frequency of the radar signal Perform a weighted summation of the clock drifts between all sites and calculate the quantized value of the clock drift. Let the quantized value of the clock drift be , then the clock drift quantization value The calculation expression is: ,in, is the weight between corresponding sites.
5. A ship trajectory optimization method based on multi-station radar fusion signals according to claim 4, characterized in that: The abnormal incremental quantitative value of the data precision loss rate generated when the screened data is obtained for fusion and clock drift quantization value After that, a data analysis model is established to generate the fusion loss risk assessment coefficient , by integrating the loss hidden danger assessment coefficient Intelligently perceive abnormal risks during data fusion.
6. A ship trajectory optimization method based on multi-station radar fusion signals according to claim 5, characterized in that: The fusion loss hidden danger assessment coefficient generated when the screened data is fused is compared and analyzed with the pre-set fusion loss hidden danger assessment coefficient reference threshold. If the fusion loss hidden danger assessment coefficient is greater than or equal to the fusion loss hidden danger assessment coefficient reference threshold, the data generated by the fusion at the corresponding moment will be marked; if the fusion loss hidden danger assessment coefficient is less than the fusion loss hidden danger assessment coefficient reference threshold, the data generated by the fusion at the corresponding moment will not be marked.
7. The ship trajectory optimization method based on multi-station radar fusion signals according to claim 1 is characterized in that: For the marked data, the specific steps to fill the data at risk of data loss are as follows: Step 1: Determine the area to be interpolated and set the parameters required for the Kriging interpolation algorithm; Step 2: Use the set parameters to execute the Kriging interpolation algorithm to interpolate the marked data points; Step 3: Evaluate the interpolated data to check whether the interpolation results are reasonable; Step 4: Fill the data processed by Kriging interpolation into the original data to replace the marked data.
Citation Information
Patent Citations
Intelligent ship-shore cooperative target tracking system and method based on shore-based radar
CN111157982A
Near-shore supervision system for illegal mining of sea sand and information fusion method thereof
CN117310689A
Ship dynamic trajectory prediction and tracking method based on multi-source data fusion
CN117556376A
Ship track excavation system and method based on AIS and radar transponder
CN117932514A
Track fusion method and device for unmanned surface vehicle
US20230104047A1
