Offshore data monitoring method based on wireless communication system

Abstract

The invention relates to the technical field of offshore data monitoring, and discloses an offshore data monitoring method based on a wireless communication system, and the method comprises the following steps: S1, data collection: obtaining offshore monitoring data of a target; s2, data preprocessing: performing format conversion, duplicate removal, time alignment and data interpolation processing on the data acquired by the data acquisition; s3, data fusion: matching and fusing the processed data based on a time interleaving algorithm and a spatial clustering algorithm; s4, performing target tracking, and performing target trajectory prediction by adopting a Kalman filtering method under the condition of target data missing; and S5, monitoring output, generating a fusion data set, and outputting dynamic information of the target. Through combination of deep learning target identification, LSTM network state prediction and an efficient wireless communication technology, more accurate marine target monitoring, stable real-time data transmission and efficient target state tracking are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of marine data monitoring, and in particular to a marine data monitoring method based on a wireless communication system. Background Art

[0002] At present, maritime data monitoring mainly relies on radar, AIS (Automatic Identification System) and satellite remote sensing technology. These technologies have good performance in target detection, basic trajectory tracking and communication coverage. Radar monitoring can continuously track targets in bad weather, the AIS system can provide the identity and location information of the ship, and satellite remote sensing can cover a large area of ​​sea to provide support for long-distance target monitoring. In addition, some monitoring systems have also introduced filtering algorithms for target state estimation, combined with certain rule algorithms, to improve the efficiency of data processing. Overall, existing technologies can already meet some maritime monitoring needs and achieve basic target recognition and state tracking in specific environments.

[0003] Although the existing technology has made certain progress, it still has some shortcomings in the face of a more complex maritime environment. First, the recognition accuracy of traditional target recognition methods is significantly reduced in complex environments, especially when the sea conditions change drastically or the target behavior is abnormal, the misjudgment rate is high, affecting situational awareness. Second, conventional target trajectory prediction methods are unstable when dealing with non-uniform and highly maneuverable targets, and the trajectory deviation is large, resulting in tracking inaccuracy. Third, the wireless communication system lacks transmission stability in the maritime environment, and the data is easily affected by signal attenuation or interference, resulting in delays and loss of monitoring information. Finally, the monitoring system is still highly dependent on manual labor, with a limited level of intelligence, and there are still shortcomings in accident handling and autonomous decision-making. Summary of the invention

[0004] In view of the deficiencies of the prior art, the present invention provides a marine data monitoring method based on a wireless communication system, which solves the problems of low marine target recognition accuracy, inaccurate target state prediction and unstable data transmission in the prior art.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: A marine data monitoring method based on a wireless communication system comprises the following steps: S1. Data collection, obtaining the target’s maritime monitoring data; S2, data preprocessing, performing format conversion, deduplication, time alignment and data interpolation processing on the data obtained by the data collection; S3, data fusion, matching and fusion of processed data based on time interleaving algorithm and spatial clustering algorithm; S4, target tracking, in the absence of target data, the Kalman filter method is used to predict the target trajectory; S5. Monitor output, generate fused data set, and output dynamic information of the target.

[0006] Preferably, the data collection includes: Receive ship automatic identification system data through the AIS data receiving module and analyze the static and dynamic information of the ship; The radar detection data receiving module is used to obtain the radar detection data of the marine target and extract the position information and motion parameters of the target; Acquire the data of the mobile terminal device through the mobile communication terminal collection and positioning module, and analyze the terminal's location information and unique identification; The data storage module is used to cache the collected data to ensure data integrity.

[0007] Preferably, the data preprocessing includes: The time window method is used to deduplicate AIS data, radar data and mobile communication terminal data; Adopting the geodetic coordinate conversion model, the location information of different data sources is standardized by unified geographic coordinates; Linear interpolation method is used to fill in missing data and ensure the continuity of data time series; The filtering algorithm is used to denoise the original data to reduce the impact of data outliers.

[0008] Preferably, the data fusion includes: A time interleaving algorithm is used to align the timestamps of different data sources and remove data whose time error exceeds the threshold. A spatial clustering algorithm is used to match the target location information from different data sources, and the matching threshold is set according to the spatial distance; The target association method is used to match AIS data, radar data and mobile communication terminal data, and a multi-source fusion target data set is established; The data consistency detection method is used to optimize the fused data to improve the reliability of the data.

[0009] Preferably, the target tracking includes: The Kalman filter method is used to estimate the target state and predict the target position at the next moment based on the state transition equation; In the absence of AIS data, radar data and mobile communication terminal data are used to predict target trajectory and dynamically adjust state estimation parameters; In the absence of radar data, AIS data and mobile communication terminal data are used to predict target trajectory and compensate for trajectory deviation; In the absence of mobile communication terminal data, AIS data and radar data are used to predict target trajectory and correct target speed changes; The data confidence assessment method is used to judge the data accuracy of the tracking target and adjust the tracking parameters.

[0010] Preferably, the monitoring output includes: Generate a fused data set based on the target’s time synchronization information and store the target’s historical data; Output the fused target dynamic information, including the target's position, speed, heading and timestamp; According to the source information of the target data, identify the data source category and adjust the data source weight; Adopt data storage and management module to classify and store fusion data and provide data calling interface; The data visualization module is used to convert the fused data into a graphical display and support real-time monitoring query.

[0011] Preferably, the time interleaving algorithm comprises: Calculate the time difference between different data sources and set the time matching window to ensure data time synchronization; Use interpolation calculation method to correct the timestamps of different data sources and perform time alignment processing; Eliminate data whose time error exceeds the set threshold, and retain valid data for subsequent fusion; A time synchronization mechanism is used to time align data from different data sources and ensure the time consistency of the data.

[0012] Preferably, the spatial clustering algorithm includes: Calculate the spatial distance between AIS data, radar data and mobile communication terminal data, and set the distance matching threshold; The target clustering method is used to spatially cluster targets from different data sources and merge data belonging to the same target; Use data association analysis methods to compare target trajectories and ensure the stability of target matching; The trajectory matching method is used to perform correlation analysis on the historical trajectory and real-time trajectory of the target, and optimize the matching degree of the target position.

[0013] Preferably, the Kalman filtering method comprises: The state transfer equation is used to predict the state of the target at the next moment and calculate the target motion parameters; The target state is corrected according to the observation equation and the Kalman gain matrix is ​​updated to adjust the target prediction result; Adopt the target state estimation algorithm to optimize the target trajectory and improve the accuracy of target positioning; The error correction method is used to correct the target trajectory prediction error and optimize the target's motion trajectory.

[0014] Preferably, the source information of the target data includes: When the target relies only on AIS data, the data source is identified as AIS, and the track information of the AIS target is stored; When the target relies only on radar data, the data source is identified as radar, and the detection data of the radar target is stored; When the target relies only on mobile communication terminal data, the data source is identified as the mobile communication terminal, and the location information of the mobile terminal is stored; When the target fuses multiple data sources, the data source is identified as fused data, and the weight of the fused data is calculated based on the confidence of the data source; A data classification management method is adopted to classify and store target data from different data sources, and a multi-level data management interface is provided.

[0015] The present invention provides a method for monitoring marine data based on a wireless communication system. It has the following beneficial effects: 1. The present invention improves the accuracy of target classification by combining deep learning and multi-dimensional target feature analysis, and solves the problem that traditional classification technology cannot accurately respond to complex environmental changes. Compared with the simple rule classification in the prior art, the present invention can adapt to different scenarios more flexibly.

[0016] 2. Using LSTM network to accurately predict the target motion not only improves the estimation accuracy of the target state, but also can correct the tracking results in real time. In this way, the system responds faster to dynamically changing targets and is much more stable than traditional methods.

[0017] 3. With the help of wireless communication technology, the present invention ensures efficient and real-time data transmission, reduces delays and packet loss. Compared with the shortcomings of existing monitoring systems, the optimization of the present invention makes real-time feedback more stable and improves the reliability of marine monitoring.

[0018] 4. The target recognition and decision support mechanism of the present invention realizes intelligent marine monitoring. This not only reduces the burden of manual intervention, but also improves the overall monitoring efficiency. Compared with traditional labor-intensive operations, this level of automation allows the system to respond more quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Please see attached Figure 1 The embodiment of the present invention provides a method for monitoring marine data based on a wireless communication system, comprising the following steps: S1. Data collection, obtaining the target’s maritime monitoring data; Generally, the monitoring of maritime targets relies on multiple data sources, each of which has different measurement methods, time synchronization mechanisms, and data formats. Therefore, this embodiment provides a multi-source data collection method to ensure that the monitoring information of maritime targets is comprehensive and stable, and can effectively support the operation of subsequent data processing modules.

[0022] In this embodiment, data collection mainly includes AIS data reception, radar detection data reception, mobile communication terminal data collection and data storage caching mechanism, ensuring that information from different data sources can be stored according to a unified standard, providing a basis for subsequent fusion processing.

[0023] AIS (Automatic Identification System) is an automatic broadcast system based on VHF (very high frequency radio) communication. Ships periodically send their status information, including static and dynamic information.

[0024] AIS static data is mainly used to identify the ship, including: Unique vessel identification number ( ): Corresponding to MMSI (Maritime Mobile Service Identity, maritime mobile service identity).

[0025] Ship name ( ): Vessel registration name.

[0026] Vessel Type ( ): such as cargo ships, tankers, fishing boats, etc.

[0027] Size information ( ): The length and breadth of the ship.

[0028] Draft ( ): Maximum draft of the ship.

[0029] AIS static data is usually manually input by the ship and stored in the shipboard AIS terminal equipment, and generally does not change frequently.

[0030] AIS dynamic data is automatically updated by the ship's onboard GPS or navigation control system, including: Current location ( ): Latitude and longitude coordinates expressed in the WGS-84 coordinate system.

[0031] Speed ​​( ): The unit is knots, which is 1 nautical mile per hour.

[0032] course( ): 0-360 degrees, indicating the movement angle relative to the north direction.

[0033] Send timestamp ( ): Record the UTC time when the current data is sent.

[0034] AIS data is usually broadcast every 2-10 seconds, and the faster the ship is, the more frequent the updates.

[0035] Radar (Radio Detection and Ranging) uses the principle of electromagnetic wave ranging to calculate the distance, direction and motion status of the target by receiving the echo signal reflected by the target.

[0036] The basic measurement parameters of radar include: Target distance ( ): The distance between the radar and the target, calculated by the formula: ; in: is the propagation speed of electromagnetic waves in air; It is the time interval between radar transmitting and receiving signals (unit: seconds).

[0037] Target latitude and longitude ( , ) calculation, it is necessary to combine the radar site location ( , ) to perform coordinate transformation: ; ; in: , are the latitude and longitude coordinates of the radar station; is the target distance (unit: meter); is the azimuth of the target (unit: radians); is the latitude of the target detected by the radar, in degrees (°); is the longitude of the target detected by the radar.

[0038] Some high-end maritime radars have a target size estimation function. The radar can calculate the size of the target based on the echo strength: ; in: is the equivalent radar cross section (RCS) of the target; is the received echo power; is the radar transmitting power; is the system calibration factor.

[0039] Generally, RCS values ​​can be used to classify targets, such as distinguishing large ships, small boats or offshore buoys.

[0040] Mobile communication terminal data is located based on wireless signaling. Generally, the positioning methods include multi-base station triangulation and wireless fingerprint matching.

[0041] By measuring the signal transmission time from the target device to different base stations, its location can be calculated: ; in: Target to base station Distance (unit: meter); is the base station signal transmission time (unit: seconds); is the signal propagation speed.

[0042] The latitude and longitude of the target can be solved by the least squares method, the formula is as follows: ; in: is the latitude of the mobile terminal target, in degrees (°); is the longitude of the mobile terminal target, in degrees (°); is the base station coordinates; is the distance from the target to each base station.

[0043] In areas with sparse base stations, the signal fingerprint database matching method can be used to infer the target location based on historical signal strength data.

[0044] In order to ensure the time synchronization between different data sources, this embodiment adopts a short-term cache mechanism. The data is stored in the memory so that time alignment can be performed during subsequent data fusion.

[0045] ; in: The timestamp of the latest AIS data; The latest AIS data timestamp in seconds (s); The timestamp of the latest radar data, in seconds (s); The latest mobile communication terminal data timestamp in seconds (s).

[0046] In general, The range can be set to 10-30 seconds to ensure that the system can integrate the most recent data.

[0047] Through different measurement principles and calculation methods, the accuracy and timeliness of monitoring data are ensured. At the same time, through the data caching mechanism, the time alignment of multi-source data is guaranteed, providing a solid data foundation for subsequent data fusion, tracking analysis and other steps.

[0048] S2, data preprocessing, format conversion, deduplication, time alignment and data interpolation processing of the data obtained by data collection; Directly using unprocessed data for fusion may lead to target matching errors, trajectory breaks or abnormal predictions. Therefore, before entering data fusion, a series of data preprocessing operations are required to unify the data format, remove redundant data, adjust timestamp synchronization, fill in missing information and optimize data quality.

[0049] In this embodiment, data preprocessing mainly involves format conversion, data deduplication, time alignment, data interpolation and completion, and abnormal data processing, ensuring that each data source is fused and calculated in the same time reference system and spatial coordinate system, thereby improving monitoring accuracy.

[0050] Different data sources may use different coordinate systems and time standards. In order to ensure the uniformity of subsequent calculations, format conversion is required.

[0051] Generally, AIS data uses the WGS-84 coordinate system, i.e., longitude and latitude coordinates, while radar data usually uses the polar coordinate system, and mobile communication data may use a projected coordinate system (such as the UTM coordinate system). In order to ensure calculation consistency, all data in this embodiment are converted to the Earth-Centered, Earth-Fixed coordinate system (ECEF).

[0052] In one possible implementation, the formula for converting WGS-84 longitude and latitude coordinates to ECEF coordinates is as follows: ; ; ; in: , , is the three-dimensional coordinate in the ECEF coordinate system, in meters (m); , is the latitude and longitude of the target, in degrees (°); is the target altitude, in meters (m); For the semi-major axis of the Earth, take 6378137 meters; It is the first eccentricity of the Earth; is the radius of curvature of the convex circle.

[0053] For radar data, since it is expressed in polar coordinates, it needs to be converted to local Cartesian coordinates first, and then converted to the ECEF coordinate system. In this embodiment, the formula for radar polar coordinate conversion is as follows: ; ; in: , : The two-dimensional coordinates of the target detected by the radar in the radar station coordinate system, in meters (m); is the distance from the target to the radar station, in meters (m); It is the target azimuth measured by the radar, in radians (rad).

[0054] Since AIS, radar and mobile communication terminals may collect information of the same target multiple times in a short period of time, data deduplication is required. In this embodiment, a time window method and a spatial threshold method are used to filter duplicate data.

[0055] As an option, during a set time window Check whether there is information about the same target in multiple data sources: ; in: , The timestamp of the target data, in seconds (s); The time window threshold is in seconds (s), and is usually 1-10s.

[0056] If the time interval meets the above conditions, the spatial distance between the targets is further calculated: ; in: The target data point and The spatial distance between them is in meters (m); , , and , , is the ECEF coordinate of the target in meters (m).

[0057] The data timestamps of different data sources may be inconsistent and need to be synchronized. In this embodiment, linear interpolation is used for time alignment to estimate the target position when data is missing.

[0058] ; ; in: , is the interpolated target longitude and latitude, in degrees (°); , is the target longitude and latitude at the previous time point, in degrees (°); , is the target longitude and latitude at the next time point, in degrees (°); is the current time point to be calculated, in seconds (s); , The timestamps of adjacent data points are in seconds (s).

[0059] In order to remove measurement errors and mutation data, this embodiment uses a sliding window mean filtering method to smooth the data: ; ; in: , is the target latitude and longitude after filtering, in degrees (°); , For History The target latitude and longitude at each time point, in degrees (°); is the sliding window size, usually 3-5.

[0060] For data with abnormal values ​​such as speed and heading, this embodiment uses median filtering to remove abnormal values ​​to ensure the continuity and stability of the data.

[0061] By covering format conversion, data deduplication, time alignment, interpolation completion and abnormal data processing, the stability and consistency of data input are ensured, providing high-quality data support for subsequent data fusion.

[0062] S3, data fusion, matching and fusion of processed data based on time interleaving algorithm and spatial clustering algorithm; The core goal of data fusion is to integrate the observation information of multiple data sources to make the target state estimation more accurate and improve the stability of trajectory prediction. Due to the differences in observation characteristics, error models and data update frequencies of different data sources, a single data source may have missing or inaccurate information. Therefore, it is necessary to use data fusion methods to reasonably weight the observation results of different data sources and combine them with the state estimation algorithm to obtain a more accurate and stable target motion state. In this embodiment, weighted fusion, extended Kalman filter (EKF), joint probabilistic data association (JPDA) and other methods are used to improve the reliability and stability of data fusion.

[0063] In this embodiment, weighted fusion is first performed to calculate the contribution of different data sources to the target state estimation. The observations in the data sources are , , …, , then the calculation formula of weighted fusion is: ; in: is the estimated value of the target state after fusion, in meters (m); For the The measurement value of the target by the data source, in meters (m); For the The weight coefficients of the data sources satisfy the normalization constraint: ; In this embodiment, the weight coefficient is calculated based on the measurement error variance of the data source. , which is calculated as follows: ; in: For the The measurement error variance of the data source, in m 2 , indicating the stability of the observation value of the data source.

[0064] In one possible implementation, if the error characteristics of some data sources change dynamically over time, a recursive least squares (RLS) or adaptive Kalman filtering method is used to dynamically adjust the weights to ensure the accuracy of data fusion.

[0065] In this embodiment, in order to improve the accuracy of target state estimation, an extended Kalman filter (EKF) is used for state prediction and update. The state vector of the target is defined as: ; in: For the goal at the moment The two-dimensional position coordinates of , in meters (m); For the purpose and The speed in a direction is in meters per second (m / s).

[0066] State prediction equation: ; in: For the goal at the moment The state vector of is the state transfer matrix, describing the target motion model: ; in: is the time step, in seconds (s); is process noise, with mean 0 and covariance matrix Gaussian distribution of , indicating the influence of system noise.

[0067] Measurement equation: ; in: is the measurement vector, containing the target position observed by the sensor; is the measurement noise, with a mean of 0 and a covariance matrix of Gaussian distribution of , indicating the influence of measurement error; is the measurement matrix, defined as: ; Kalman gain calculation: ; in: is the Kalman gain matrix; is the prediction covariance matrix; is the measurement noise covariance matrix; is the weighting matrix used to calculate the gain, taking into account the measurement error and the prior estimation error; is the observation matrix, which represents the relationship between the system state and the measurement space.

[0068] Status update formula: ; ; in: is the unit matrix to ensure the correctness of matrix operations; is the actual measured value, indicating the Observed quantity; is the updated error covariance matrix, indicating that at time Uncertainty about the error in estimating the system state; is the prior error covariance matrix, indicating that at time The covariance of the errors in previous estimates of the system state.

[0069] In a multi-target environment, there may be multiple observations corresponding to different targets, so this embodiment uses a joint probabilistic data association (JPDA) method to calculate the association probability between each observation and the target.

[0070] In this embodiment, Mahalanobis distance is used for target association: ; in: is the Mahalanobis distance, which measures the degree of match between the observed value and the predicted target; For the Observation data, in meters (m); is the target state estimation value, in meters (m); is the residual covariance matrix, defined as: ; If the Mahalanobis distance satisfies: ; in: is the set correlation threshold, the observation is considered to match the target.

[0071] This embodiment provides a complete data fusion method, which covers weighted fusion, extended Kalman filter (EKF) and joint probabilistic data association (JPDA). By reasonably allocating data source weights, estimating target states, and combining association methods to improve target matching accuracy, the data fusion result is made more reliable, laying a solid foundation for subsequent target trajectory prediction and situation analysis.

[0072] S4, target tracking, in the absence of target data, the Kalman filter method is used to predict the target trajectory; S4 combines motion models, filtering algorithms, smoothing methods, and deep learning techniques in the trajectory prediction process. Through these methods, the target's motion characteristics and environmental factors can be considered to better estimate the target's future position and state.

[0073] In the first step of trajectory prediction, the target’s motion model is used to describe the change of the target’s state over time, using the “state vector definition” disclosed by S3.

[0074] In one possible implementation, the state transition of the target can be represented by a linear motion model using the “state prediction equation” disclosed in S3.

[0075] In certain cases, if the target's motion has higher dynamic characteristics, the model can be extended to an acceleration model, for example: ; in: For the goal at the moment The x-axis position of the , in meters (m); For the goal at the moment The y-axis position of the , in meters (m); For the goal at the moment The x-axis speed is in meters per second (m / s); For the goal at the moment The y-axis speed is in meters per second (m / s); is the acceleration component of the target in meters per second squared.

[0076] For the acceleration model, the state transfer matrix F\mathbf{F}F needs to be adjusted accordingly to include the acceleration information.

[0077] In order to improve the accuracy of prediction, filtering and smoothing techniques are used to reduce errors. Generally speaking, short-term observations may be interfered by noise, resulting in trajectory fluctuations. Therefore, it is necessary to smooth the historical trajectory of the target.

[0078] In this embodiment, a sliding average filter is used to smooth the trajectory, and the formula is: ; in: is the smoothed target state vector in meters (m) and meters per second (m / s); is the sliding window size; is the target state at the past N time points.

[0079] The sliding average filter can effectively reduce the impact of mutation data on target state estimation by calculating the mean of historical data, thereby improving the stability of the prediction.

[0080] In order to further improve the reliability of the prediction, this embodiment uses Bayesian estimation to perform joint estimation of the state. The core idea of ​​Bayesian estimation is to gradually optimize the estimation of the target state by updating the posterior probability distribution. In the state prediction process, the formula of Bayesian estimation is: ; in: Represents given past observation data , the goal is at time The posterior probability of the state; is the state transition probability of the target, which is usually given by the motion model; For the goal at the moment The posterior probability distribution of the state.

[0081] This method can improve the accuracy of target state estimation by integrating historical information and current observation data by recursively updating the posterior probability.

[0082] For complex nonlinear target state estimation, this embodiment uses a particle filter (PF) to handle nonlinear and non-Gaussian processes. The particle filter represents the target state as a group of particles and assigns a weight to each particle to represent its contribution to the target state.

[0083] The state update process of particle filtering can be expressed as: ; ; in: For the Particles at time Status; For the The weight of each particle; is the observation likelihood function, which indicates the degree of match between the particle state and the observed value.

[0084] Particle filtering eliminates low-weight particles through a resampling process and concentrates particle distribution in high-weight areas, thereby improving the accuracy of prediction.

[0085] In some embodiments, this embodiment also uses a long short-term memory network (LSTM) to further optimize trajectory prediction. LSTM can capture long-term dependencies in time series through its special gating structure, effectively improving the accuracy of prediction.

[0086] The state update formula of LSTM is as follows: ; in: are activation values ​​of the forget gate, input gate, and output gate, respectively, ranging from [0,1], which are used to control the selection and update of information flow; is the weight matrix of LSTM; is the bias term of LSTM; is the memory state of the LSTM unit; is the hidden state output of LSTM; is a candidate memory unit, which indicates the influence of the current input information on the memory state; OutputGate determines how much the current memory state will affect the output; It is a hyperbolic tangent activation function, which is used to generate outputs ranging from -1 to 1 and is usually used to smooth information. The hidden state of the previous moment is part of the input of the current moment, providing contextual information from the previous moment to the current moment; is the input at the current moment, usually the observation data or input features at the current moment.

[0087] In this way, LSTM is able to learn complex patterns in the target trajectory, thereby providing more accurate state predictions.

[0088] Step S4 comprehensively improves the accuracy of target trajectory prediction by combining motion models, filtering methods, particle filtering, Bayesian estimation, and deep learning techniques. The parameter definitions involved in each formula have been disclosed in detail to ensure that people in the technical field can accurately understand the meaning and function of each formula, thereby providing reliable technical support for more complex trajectory prediction and dynamic tracking.

[0089] S5, monitor output, generate fused data set, and output dynamic information of the target; S5 can identify the type of target by analyzing the target state information and trajectory prediction results obtained in the previous steps, and provide necessary decision support for the system. Target recognition and classification is not only a simple distinction of targets, but also an accurate judgment through in-depth analysis of the target's motion pattern, appearance characteristics and other relevant information. Therefore, the implementation of this step requires a detailed analysis based on the multi-dimensional characteristics of the target (including dynamic characteristics and static characteristics), and the use of appropriate classification algorithms to complete the target recognition and classification work.

[0090] In the process of target recognition and classification, it is first necessary to extract representative features from the target state information. The target state information usually includes dynamic features such as the target's position, speed, acceleration, etc. However, relying solely on these dynamic features may not be sufficient to fully identify the target. Therefore, this embodiment further introduces information such as the target's motion mode and appearance features, so as to classify the target through more comprehensive features.

[0091] This embodiment uses classification algorithms such as multi-layer perceptron (MLP) and support vector machine (SVM) to achieve target classification. By using the target state information predicted in the previous steps, these classification models are trained based on historical data to accurately classify the target type.

[0092] In this method, the target state vector As input features, after training, the classification model outputs the category label of each target The input of the target state vector and the output of the classification label are optimized through the loss function. The loss function generally adopts the cross entropy loss function, which is in the form of: ; in: For the goal at the moment The state vector of , in meters (m) and meters per second (m / s); is the classification label of the target, indicating the target type; is the total number of target types, indicating the number of target categories; For the target belongs to the category The predicted probability is usually the probability value output by the classification model.

[0093] The loss function updates the parameters (such as weights and biases) of the classification model during the training process by optimizing the gap between the target output and the actual label. This process can continuously improve the classification accuracy and enable the classification model to make correct classification judgments when facing unknown targets.

[0094] Once the target is classified, the classification results will provide key information for subsequent decision-making and action planning. In some embodiments, the present invention also combines the target's motion trajectory characteristics to further improve the accuracy of target recognition. For example, if a target exhibits a similar motion pattern to a known target, the system will confirm it in combination with this trajectory information, thereby improving the accuracy of classification.

[0095] After the target is successfully classified, the system will take different decision-making measures according to different target categories. For example, the system may select different trajectory tracking strategies according to the target type, or perform corresponding countermeasures for hostile targets, while selecting appropriate track keeping strategies for friendly targets.

[0096] The classification result can further affect the estimation of the target state. If the system finds that the type of the target has changed through target recognition, the classification result will immediately drive the re-estimation of the target state. For example, corrections can be made to dynamic features such as target speed and acceleration to ensure the accuracy and consistency of the prediction.

[0097] In order to improve the performance of target recognition, some embodiments also use a deep learning-based convolutional neural network (CNN) to process the target's image data or radar echo data, thereby further improving the classification of the target. In this implementation, CNN can extract important feature information from more complex target patterns by automatically learning the spatial and temporal features in the input data, thereby improving recognition accuracy.

[0098] S5 extracts multi-dimensional features by combining the results of the aforementioned trajectory prediction and data fusion, and the classification model can effectively distinguish different types of targets. The formulas and parameter definitions involved in each step have been disclosed in detail to ensure that people in this technical field can fully understand their functions and significance. These details not only provide reliable technical support for the target recognition and classification process, but also lay a solid foundation for the subsequent decision-making process.

[0099] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A marine data monitoring method based on a wireless communication system, characterized in that: The following steps are involved: S1. Data collection, obtaining the target’s maritime monitoring data; S2, data preprocessing, performing format conversion, deduplication, time alignment and data interpolation processing on the data obtained by the data collection; S3, data fusion, matching and fusion of processed data based on time interleaving algorithm and spatial clustering algorithm; S4, target tracking, in the absence of target data, the Kalman filter method is used to predict the target trajectory; S5. Monitor output, generate fused data set, and output dynamic information of the target.

2. The marine data monitoring method based on a wireless communication system according to claim 1, characterized in that: The data collection includes: Receive ship automatic identification system data through the AIS data receiving module and analyze the static and dynamic information of the ship; The radar detection data receiving module is used to obtain the radar detection data of the marine target and extract the position information and motion parameters of the target; Acquire the data of the mobile terminal device through the mobile communication terminal collection and positioning module, and analyze the terminal's location information and unique identification; The data storage module is used to cache the collected data to ensure data integrity.

3. The marine data monitoring method based on a wireless communication system according to claim 1, characterized in that: The data preprocessing includes: The time window method is used to deduplicate AIS data, radar data and mobile communication terminal data; Adopting the geodetic coordinate conversion model, the location information of different data sources is standardized by unified geographic coordinates; Linear interpolation method is used to fill in missing data and ensure the continuity of data time series; The filtering algorithm is used to denoise the original data to reduce the impact of data outliers.

4. The marine data monitoring method based on a wireless communication system according to claim 1, characterized in that: The data fusion includes: A time interleaving algorithm is used to align the timestamps of different data sources and remove data whose time error exceeds the threshold. A spatial clustering algorithm is used to match the target location information from different data sources, and the matching threshold is set according to the spatial distance; The target association method is used to match AIS data, radar data and mobile communication terminal data, and a multi-source fusion target data set is established; The data consistency detection method is used to optimize the fused data to improve the reliability of the data.

5. The marine data monitoring method based on a wireless communication system according to claim 1, characterized in that: The target tracking includes: The Kalman filter method is used to estimate the target state and predict the target position at the next moment based on the state transition equation; In the absence of AIS data, radar data and mobile communication terminal data are used to predict target trajectory and dynamically adjust state estimation parameters; In the absence of radar data, AIS data and mobile communication terminal data are used to predict target trajectory and compensate for trajectory deviation; In the absence of mobile communication terminal data, AIS data and radar data are used to predict target trajectory and correct target speed changes; The data confidence assessment method is used to judge the data accuracy of the tracking target and adjust the tracking parameters.

6. The marine data monitoring method based on a wireless communication system according to claim 1, characterized in that: The monitoring outputs include: Generate a fused data set based on the target’s time synchronization information and store the target’s historical data; Output the fused target dynamic information, including the target's position, speed, heading and timestamp; According to the source information of the target data, identify the data source category and adjust the data source weight; Adopt data storage and management module to classify and store fusion data and provide data calling interface; The data visualization module is used to convert the fused data into a graphical display and support real-time monitoring query.

7. The marine data monitoring method based on a wireless communication system according to claim 4, characterized in that: The time interleaving algorithm comprises: Calculate the time difference between different data sources and set the time matching window to ensure data time synchronization; Use interpolation calculation method to correct the timestamps of different data sources and perform time alignment processing; Eliminate data whose time error exceeds the set threshold, and retain valid data for subsequent fusion; A time synchronization mechanism is used to time align data from different data sources and ensure the time consistency of the data.

8. The marine data monitoring method based on a wireless communication system according to claim 4, characterized in that: The spatial clustering algorithm includes: Calculate the spatial distance between AIS data, radar data and mobile communication terminal data, and set the distance matching threshold; The target clustering method is used to spatially cluster targets from different data sources and merge data belonging to the same target; Use data association analysis methods to compare target trajectories and ensure the stability of target matching; The trajectory matching method is used to perform correlation analysis on the historical trajectory and real-time trajectory of the target, and optimize the matching degree of the target position.

9. The marine data monitoring method based on a wireless communication system according to claim 5, characterized in that: The Kalman filtering method comprises: The state transfer equation is used to predict the state of the target at the next moment and calculate the target motion parameters; The target state is corrected according to the observation equation and the Kalman gain matrix is ​​updated to adjust the target prediction result; Adopt the target state estimation algorithm to optimize the target trajectory and improve the accuracy of target positioning; The error correction method is used to correct the target trajectory prediction error and optimize the target's motion trajectory.

10. The marine data monitoring method based on a wireless communication system according to claim 1, characterized in that: The source information of the target data includes: When the target relies only on AIS data, the data source is identified as AIS, and the track information of the AIS target is stored; When the target relies only on radar data, the data source is identified as radar, and the detection data of the radar target is stored; When the target relies only on mobile communication terminal data, the data source is identified as the mobile communication terminal, and the location information of the mobile terminal is stored; When the target fuses multiple data sources, the data source is identified as fused data, and the weight of the fused data is calculated based on the confidence of the data source; A data classification management method is adopted to classify and store target data from different data sources, and a multi-level data management interface is provided.

Images