A shipping logistics tracking method

Abstract

The application relates to the technical field of logistics tracking, in particular to a shipping logistics tracking method.The method comprises the following steps: detecting a signal interruption time point and a signal interruption duration to obtain a signal interruption interval; according to the signal interruption interval, extracting a last recorded sailing speed, direction and position of a ship before interruption from a historical sailing track database to generate an initial state vector; predicting a future transportation path of the ship through a corrected position sequence to generate a predicted path set; according to the predicted path set, constructing a ship motion track in a three-dimensional virtual scene to generate a visual track model; combining a confidence degree of the calculated position sequence to obtain a decision support view; and through the decision support view, updating a port scheduling plan in real time to generate an optimized loading and unloading time sequence.The application solves the problem of low calculation accuracy of the position of the ship in a complex scene with signal interruption, and improves the position accuracy of the ship and the prediction accuracy of the future transportation path.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of logistics tracking, in particular to a shipping logistics tracking method. BACKGROUND

[0002] In modern shipping logistics, the real-time position and navigation trajectory of a ship are the core technologies to ensure shipping safety and improve transportation efficiency. However, due to various factors such as adverse weather, equipment failure, signal interference, etc., ships often face signal interruption during navigation. This signal interruption not only leads to the loss of ship position data, but also affects the execution of port scheduling plans, further increasing the complexity and uncertainty of shipping logistics. Traditional ship tracking technology relies on real-time sensor data to provide ship position and state information, but once the signal is interrupted, the system cannot accurately obtain the dynamic position of the ship, posing great challenges to port scheduling and transportation path prediction.

[0003] Currently, for the tracking problem after ship signal interruption, traditional methods usually use static compensation or rely on a single data source for calculation. These methods often cannot effectively solve the problem of accurate calculation of ship position in complex dynamic sea conditions and variable weather conditions. In addition, existing technologies mainly rely on historical data for path prediction, without fully considering real-time dynamics of environmental changes, ship state, and state estimation and correction after signal interruption. With the continuous expansion of global shipping scale, the improvement of port scheduling and logistics efficiency puts forward higher requirements for real-time and accurate trajectory calculation and path prediction. Therefore, how to accurately calculate the position of the ship and predict the future transportation path in the complex scenario of signal interruption has become a key problem to be solved in the field of shipping logistics.

[0004] The present application provides a new shipping logistics tracking method, aiming to solve the problem of ship position calculation and path prediction in the case of signal interruption. Through multi-source data fusion and advanced state estimation technology, the present application can accurately calculate the motion trajectory of the ship during signal loss and provide a visual ship trajectory model in a three-dimensional virtual scene, providing scientific decision support for port scheduling and transportation path optimization. This method not only makes up for the shortcomings of existing technologies, but also significantly improves the efficiency and safety of shipping logistics. SUMMARY

[0005] The present application provides a shipping logistics tracking method, mainly including:

[0006] By collecting ship sensor data and external environmental data, detecting the signal interruption time point and duration, and obtaining the signal interruption interval;

[0007] According to the signal interruption interval, the last recorded sailing speed, direction and position of the ship before the interruption are extracted from the historical sailing trajectory database to generate an initial state vector;

[0008] Through the initial state vector, combined with the real-time acquired sea state data and weather data, the possible position of the ship during the signal interruption is calculated through the Kalman filtering algorithm to obtain a calculation position sequence;

[0009] If the calculation position sequence deviates from the historical trajectory by more than a preset threshold, the port scheduling data and AIS data of nearby ships are integrated to correct the calculation position sequence and obtain a corrected position sequence;

[0010] Through the corrected position sequence, the future transportation path of the ship is predicted through the particle filtering algorithm to generate a prediction path set;

[0011] According to the prediction path set, the ship motion trajectory is constructed in a three-dimensional virtual scene, the real position data and the calculation position data are distinguished by color coding, and a visual trajectory model is generated;

[0012] Through the visual trajectory model, combined with the confidence of the calculation position sequence, the reliability of the speculative data is presented through the transparency mapping to obtain a decision support view;

[0013] Through the decision support view, the port scheduling plan is updated in real time to generate an optimized loading and unloading time sequence.

[0014] The technical scheme provided by the embodiment of the application can include the following beneficial effects:

[0015] The application discloses a technical scheme of trajectory calculation and scheduling optimization in a shipping logistics ship signal interruption scenario, and provides a complete solution for the business problems of missing trajectory and low port scheduling efficiency caused by ship signal interruption. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1A shipping logistics tracking method flow chart in an embodiment of the present application;

[0017] Figure 2 A shipping logistics tracking system structure diagram in an embodiment of the present application. DETAILED DESCRIPTION

[0018] The technical solutions of the present application will be described clearly and completely in combination with embodiments. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the present application.

[0019] As Figure 1 , the shipping logistics tracking method in the embodiment can specifically include:

[0020] Step S1: acquiring ship sensor data and external environment data, determining a signal interruption interval through time series analysis; according to the signal interruption interval, extracting the last sailing state of the ship before interruption from a historical sailing track database, and generating an initial state vector;

[0021] The determination of the signal interruption interval through time series analysis includes:

[0022] The sensor data and environment data are acquired through ship sensors and external environment acquisition equipment, the time point of signal interruption in the data stream is determined through a time series analysis method, if the sensor data or the environment data is below a preset threshold, the starting time point of signal interruption is obtained through a sliding window algorithm to detect continuous abnormal points, the duration of interruption is acquired through a time difference calculation method according to the starting time point of signal interruption, the ending time point of signal interruption is determined, the sensor data and the environment data are weighted and integrated through a data fusion technology to generate a fusion data set, if the environment data fluctuation in the fusion data set exceeds a preset threshold, the correlation between the environment data and the signal interruption is analyzed through a decision tree algorithm to determine the interruption interval classification, the interruption interval sequence is separated from the normal interval through a time series segmentation method according to the interruption interval classification result, and the time distribution characteristics of the interruption interval are generated through feature extraction of the interruption interval sequence through time series analysis.

[0023] In particular, the above technical solution is based on the continuous acquisition of ship sensor data such as ship position, speed, heading, and external environmental data such as wind speed, wave height, and air pressure by ship sensors and environmental collection devices. After time stamp alignment, an ordered time series is formed. Through time series analysis, the data stream is continuously monitored, and through the collection time difference between adjacent two data, it is judged whether the data has been interrupted in the collection process. When the above collection time difference is greater than the set time, the interruption occurs between the above two data, otherwise, it is not interrupted. If it is detected that the sensor data or environmental data is lower than the set threshold, the sliding window algorithm is automatically used to perform window statistical analysis on the continuous sampling points, identify continuous abnormal points, and take the first abnormal point time as the signal interruption start time. According to the time difference between the signal interruption start time and the subsequent data recovery time, the interruption duration and signal interruption end time are directly calculated. In order to further improve the detection accuracy, the sensor data and environmental data are fused by weighting to form a fused data set to enhance the sensitivity to signal interruption under complex working conditions. If the environmental data fluctuation of the fused data within the interruption interval exceeds the threshold, the decision tree algorithm is called to analyze the correlation between environmental factors and signal interruption, and the interruption interval is automatically classified to realize accurate attribution and classification of signal interruption. Based on the classification result of the interruption interval, the time series is sliced by the time series segmentation method to accurately separate the interruption interval and the normal interval, and the complete interruption interval sequence is extracted. The time series analysis method is applied to the interruption interval sequence again to extract time characteristic information including start and end time, duration, interval distribution, etc., to generate the time distribution characteristics of the interruption interval, and realize the regular modeling of the signal interruption event. In actual shipping scenarios, for example, a ship loses GPS signal for several seconds due to bad weather. Based on the above algorithm, the interruption is automatically determined and classified, and the interval distribution data is generated to provide a data basis for subsequent trajectory compensation and scheduling optimization. This technical means closely combines sliding window, weighted data fusion, decision tree classification, time series segmentation and feature extraction algorithm with actual collection system to form an automatic detection, attribution, segmentation and modeling closed loop process, which can realize efficient and accurate identification and classification of signal interruption, overcome the inaccuracy and timeliness of existing technology relying on single signal or manual determination, improve the intelligent level of data processing under abnormal state in shipping logistics scene, and ensure the quality of basic data for subsequent trajectory calculation and scheduling decision.

[0024] Further, the initial state vector is generated, including: obtaining the last recorded navigation data of the ship before the signal interruption from the historical navigation trajectory database to obtain a record set containing the navigation speed, heading and position; if the record set is not empty, sorting according to the last record time, extracting the record with the latest timestamp to obtain the last navigation state; converting the navigation speed, heading and position in the last navigation state into a standardized initial state vector through a vector conversion method; smoothing the initial state vector to obtain an optimized state vector. If the deviation of the optimized state vector from the mode of the historical navigation trajectory exceeds a preset threshold, the initial state vector is corrected according to the statistical characteristics of the historical navigation trajectory to obtain a corrected initial state vector.

[0025] Specifically, the above technical solution is based on the historical navigation trajectory database. When the signal interruption occurs, the last recorded navigation speed, heading and position data before the interruption are obtained to form a navigation record set, and the latest record is extracted by sorting according to the timestamp. The navigation speed, heading and position data contained in the record are standardized into an initial state vector by a vector conversion method. Then, the vector is smoothed by a Kalman filter algorithm to reduce the noise effect, and an optimized state vector is obtained. When there is a large deviation between the optimized state vector and the mode of the historical navigation trajectory, the state vector is automatically corrected according to the statistical characteristics of the historical trajectory, so as to form a corrected initial state vector that is more consistent with the actual navigation characteristics. The above technical solution can obtain accurate initial state data, which lays a foundation for calculating the ship position sequence.

[0026] Step S2: According to the initial state vector, real-time sea state data and weather data, the ship position during the signal interruption is calculated through state estimation to obtain a calculated position sequence; if the deviation between the calculated position sequence and the historical trajectory exceeds a preset threshold, the port scheduling data and AIS data of nearby ships are fused to correct the calculated position sequence to obtain a corrected position sequence;

[0027] The ship position during the signal interruption is calculated by state estimation to obtain a calculated position sequence, including: obtaining an initial state vector, real-time sea state data and real-time weather data, and obtaining a synchronized state data set through data time synchronization processing; obtaining a preliminary state estimation value through state estimation by combining the synchronized state data set with a ship motion model through a Kalman filtering algorithm; if the deviation of the preliminary state estimation value from the sensor data fusion exceeds a preset threshold, adjusting the state estimation through the updating step of the Kalman filtering algorithm to obtain a corrected state estimation value; obtaining a single-step position prediction value by predicting the position at the next time through the prediction step of the Kalman filtering algorithm according to the corrected state estimation value and fusing the environmental dynamic influence; obtaining a continuous position sequence by iteratively executing the prediction and updating steps of the Kalman filtering algorithm through the single-step position prediction value and combining the ship motion model; if the time span of the continuous position sequence exceeds the signal interruption time, verifying the sequence consistency through data time synchronization processing to obtain a final calculated position sequence; and obtaining a smooth calculated position sequence by optimizing the sequence smoothness according to the final calculated position sequence and fusing the real-time sea state data and weather data.

[0028] Specifically, the initial state vector of the ship, i.e. the current position, speed and heading of the ship, is obtained, and real-time sea state data and weather data are obtained at the same time. After time synchronization processing of these data, a unified synchronized state data set is formed, and the above data set is input into the ship motion model to obtain the state estimation value of the ship at the next time under the influence of the above data set through the above Kalman filtering algorithm, i.e. the state transition equation. The ship motion model is trained based on historical synchronized state data sets and corresponding ship positions, speeds and headings. The state estimation value includes the predicted position, speed and heading of the ship. If the deviation of the preliminary estimation value from the fusion result of the real-time sensor data exceeds the preset threshold, the state estimation value is corrected based on the sensor data, and the sensor data includes the position, speed and heading of the ship. Through this step, the calculation of the current position of the ship during the signal interruption can be more accurate. In the next step, the corrected state estimation value is used to calculate the specific position of the ship at the next time through the prediction step of the Kalman filtering algorithm, and a single-step position prediction value is generated. After obtaining the single-step position prediction value, the algorithm iteratively executes the prediction and updating steps of the Kalman filtering in combination with the ship motion model, thereby obtaining a series of continuous position sequences. These position sequences can reflect the motion trajectory of the ship during the signal interruption. If the time span of the continuous position sequence exceeds the signal interruption time, further data synchronization processing is performed to verify the consistency of the calculation result, and the final calculated position sequence is ensured to be consistent with the actual situation in time.

[0029] The technical scheme realizes dynamic tracking of the ship during signal interruption by multi-source data fusion, combination of the ship motion model and accurate estimation of the Kalman filtering algorithm, can effectively solve the problem that the ship position cannot be updated in real time, and improves the stability and accuracy of the ship navigation system. In addition, the calculation result of the system provides reliable data support for port scheduling and shipping route optimization, further optimizes the efficiency of shipping logistics, especially in complex environments, significantly improves the accuracy and availability of ship position prediction.

[0030] For example, in actual application scenarios, when the ship encounters signal loss or interference, traditional methods cannot accurately determine the position of the ship, which affects the port scheduling plan. By adopting the technical scheme, the system can calculate the accurate position of the ship in real time, update the port scheduling plan in time, avoid the delay of loading and unloading caused by inaccurate information in the port, and improve the efficiency of the entire shipping logistics system. The combination of multi-source data fusion and Kalman filtering algorithm overcomes the uncertainty caused by signal interruption and provides more reliable technical support for the shipping industry.

[0031] Further, the corrected position sequence is obtained by correcting the calculated position sequence, comprising: if the deviation of the calculated position sequence and the historical trajectory exceeds the preset threshold, the deviation value is calculated by comparing the calculated position sequence and the historical trajectory to obtain the deviation distribution characteristics; the calculated position sequence is optimized by Kalman filtering algorithm by fusing the port scheduling data and the AIS data of the nearby ship to obtain the preliminary corrected position sequence; the position sequence deviation is adjusted by particle filtering algorithm according to the preliminary corrected position sequence combined with the ship motion state to obtain the optimized corrected position sequence; if the consistency of the optimized corrected position sequence and the historical trajectory is lower than the preset threshold, the time stamp in the port scheduling data and the speed information in the AIS data are extracted by multi-source data integration to update the position sequence to obtain the high-precision position sequence; the real-time requirement of the position sequence is judged according to the high-precision position sequence to obtain the real-time corrected position sequence; the final corrected position sequence is obtained by smoothing the position sequence by weighted average according to the real-time corrected position sequence combined with the trajectory consistency; and whether the position sequence accuracy meets the preset accuracy threshold is judged according to the final corrected position sequence.

[0032] Specifically, the above technical solution corrects the position sequence dynamically to obtain a high-precision and continuous ship movement trajectory. For the case that the deviation between the position sequence and the historical trajectory exceeds the preset threshold, the system automatically compares the two and calculates the deviation, extracts the deviation distribution characteristics, introduces the Kalman filtering algorithm to optimize the calculated trajectory in combination with the port scheduling data and nearby ship AIS data, fuses multiple source dynamic data in real time, generates a preliminary corrected position sequence, and on this basis, introduces the particle filtering algorithm according to the actual movement state of the ship to further adjust the position deviation, so that the sequence fits the complex and nonlinear movement change. When the consistency between the optimized and corrected position sequence and the historical trajectory is still lower than the standard, the system calls the time stamp and AIS speed information of the port scheduling data, integrates multiple sources, dynamically completes and corrects the trajectory points, and obtains a high-precision position sequence. Further, the high-precision position sequence is analyzed for its timeliness to determine whether it meets the real-time updating requirements of the system, realizing the closed-loop correction of the whole-process trajectory data. Through comparison of the real-time corrected position sequence and its consistency with the historical trajectory, the system smoothes the trajectory by weighted average method, and outputs the smoothed and continuous corrected trajectory. Then, the trajectory accuracy is verified to ensure that it meets the set precision threshold. Taking the case of a container ship entering a port area and signal interference causing data drift as an example, the system can correct the trajectory obtained by integrating AIS, scheduling, historical and other data to a trajectory close to the actual movement state through the above multi-dimensional dynamic correction and smoothing mechanism, and significantly improve the continuity of the trajectory and the data reliability.

[0033] Step S3: predicting a future transport path of the ship through the particle filtering algorithm based on the corrected position sequence, and generating a prediction path set;

[0034] The generation of the prediction path set includes: obtaining the corrected position sequence, obtaining the continuous ship movement state through data cleaning and interpolation processing; extracting dynamic characteristics from the ship movement state through the particle filtering algorithm to determine an initial prediction set of the future transport path; adjusting the weight of the particle filtering algorithm through environmental influencing factors to obtain an optimized prediction path set; if the path prediction accuracy of the prediction path set is lower than a preset threshold, updating the particle state through iteration to obtain a new prediction path set; calculating the evaluation index of the path optimization target based on the optimized prediction path set to determine the optimal transport path; updating the ship movement state through real-time data input to obtain a dynamically adjusted prediction path set; and outputting the final prediction path set through the evaluation result of the path prediction accuracy.

[0035] Specifically, by data cleaning and interpolation on the corrected ship position sequence, missing values and abnormal points are eliminated to form a time-continuous motion state input sequence. The particle filter algorithm is based on this sequence to establish a large number of particle individuals, each of which represents a possible future state change path of the ship. The system extracts dynamic characteristics of each particle such as speed, heading, acceleration, etc. and introduces different environmental factors such as real-time wind, wave height, ocean current, etc. into the particle weight adjustment link to realize dynamic weight distribution and probability modeling of path possibility in the particle filtering process. If the path prediction accuracy of the generated prediction path set is lower than the system set threshold, the system automatically iteratively adjusts the particle state, resamples and evolves the particle distribution, and continuously optimizes the prediction path set structure until the overall accuracy meets the requirements. According to the optimized path set, the system comprehensively evaluates multiple target indicators such as transportation timeliness, path safety, fuel consumption, etc. to select the optimal transportation path. On this basis, the system dynamically updates the motion state by real-time collection of the latest ship and environmental data, further corrects and supplements the particle distribution, and realizes adaptive adjustment of the prediction path set. The path prediction accuracy evaluation result is directly used as the criterion for path set output threshold, and the path set output with high confidence is called by subsequent three-dimensional visualization, scheduling optimization and other modules. Taking the transoceanic transportation of a certain cargo ship in complex sea conditions as an example, the system can push multiple optional paths according to the corrected trajectory and environmental changes, and automatically select the optimal path combined with the port operation plan, overcoming the problems of limited prediction accuracy of single model and untimely response to environmental changes in the prior art, forming a closed-loop system of data flow, model prediction and decision optimization, realizing high-reliability prediction of ship transportation path, and providing effective technical support for intelligent port scheduling and risk control. The introduction of the above particle filtering algorithm and the dynamic adjustment mechanism of the environment, the technical features and the actual data and needs in the application scenario are closely combined to form an integrated solution that supports each other in function, realizes the automation and intelligentization of path prediction, and improves the overall transportation safety and operational efficiency.

[0036] Step S4: Constructing the ship motion trajectory in the three-dimensional virtual scene, distinguishing the real position data and the calculated position data by color coding, and generating a visual trajectory model;

[0037] The generating of the visualized trajectory model comprises: obtaining real position data and calculated position data from the set of predicted paths, obtaining a set of position data in a unified format through data fusion processing, obtaining a ship movement trajectory through trajectory generation algorithm by analyzing the set of position data in a unified format, determining an initial trajectory coordinate set, constructing a three-dimensional trajectory model in a three-dimensional virtual scene according to the initial trajectory coordinate set through scene rendering, obtaining a rendered trajectory framework, distinguishing real position data and calculated position data through color coding for the rendered trajectory framework, generating a color-coded trajectory, and obtaining the visualized trajectory model; if a trajectory point in the visualized trajectory model deviates from the set of predicted paths by more than a preset threshold, recalculating the trajectory point coordinate through position data analysis, updating the visualized trajectory model; adjusting the visualized trajectory model through trajectory dynamic updating technology according to real-time ship position data, obtaining a dynamically updated three-dimensional trajectory model; and performing light and shadow processing and perspective optimization on the dynamically updated three-dimensional trajectory model through scene rendering technology, and obtaining a final visualized trajectory model.

[0038] Specifically, the above technical solution processes the real position data and the calculated position data from the set of predicted paths into a unified data format, forms a consistent-structure position data set after data fusion, analyzes the data set through trajectory generation algorithm, automatically generates a continuous ship movement trajectory and determines a trajectory coordinate set, calls scene rendering technology in a three-dimensional virtual scene, constructs a three-dimensional trajectory model according to the trajectory coordinate set, significantly distinguishes real position and calculated position through color coding, forms an intuitive and identifiable color-coded trajectory, realizes the construction of the visualized trajectory model, and ensures that the model accurately reflects the actual and predicted dynamics of the ship. In the case of deviation of the trajectory point from the set of predicted paths, the position data of the deviated point is analyzed, the trajectory point coordinate is automatically recalculated, and the visualized trajectory model is updated in real time. After the real-time ship position data is input, the trajectory dynamic updating technology automatically adjusts the three-dimensional trajectory model, continuously completes and corrects the trajectory according to the new data stream, meets the requirement of dynamic change of ship position data, and further improves the readability and spatial level of the trajectory model through light and shadow processing and perspective optimization and other scene rendering, so that the decision maker can intuitively master the relationship between the ship running trajectory and the predicted path from multiple angles. In actual application, if a ship calculates multiple future paths through a prediction algorithm during signal interruption, the system uniformly fuses and visualizes the real and calculated trajectory data, the management personnel can accurately distinguish the deviation between the current trajectory and the prediction based on the three-dimensional visualized model, and assist in dynamically adjusting the port scheduling plan. This technical means cooperates the algorithm features of data fusion, trajectory generation, dynamic updating, three-dimensional rendering and color coding with the technical features of three-dimensional modeling and real-time scene application, forms a full-process closed loop of shipping logistics trajectory visualization, solves the technical problems of non-intuitive presentation of trajectory and insufficient decision support data in the signal interruption scenario, and improves the timeliness and reliability of port scheduling decision.

[0039] Step S5: generate a decision support view through transparency mapping by visualizing the trajectory model combined with the confidence of the estimated position sequence;

[0040] Wherein, the decision support view generated by transparency mapping comprises: obtaining position sequence data by visualizing the trajectory model; smoothing the position sequence by Kalman filtering to obtain smoothed trajectory data; calculating the confidence of each position point by Bayesian inference according to the smoothed trajectory data to obtain a position confidence sequence; mapping the position confidence sequence to visualized transparency values by a transparency mapping function to obtain transparency mapping data; generating a trajectory view according to the transparency mapping data by visualized rendering to obtain a preliminary decision support view; if the transparency of the trajectory point in the preliminary decision support view is lower than a preset threshold, supplementing the missing position data by an interpolation algorithm to obtain an optimized trajectory view; extracting the trajectory trend by sequence analysis according to the optimized trajectory view to obtain trajectory trend data; fusing the trajectory trend data with the optimized trajectory view by view superposition method to obtain the final decision support view.

[0041] Specifically, after obtaining the ship trajectory position sequence data, the above technical solution performs dynamic smoothing processing on all collected trajectory points through the Kalman filtering algorithm, so that the noise and mutations in the original trajectory are effectively weakened, thereby obtaining smooth trajectory data with strong continuity and conforming to the actual trajectory change rule. On this basis, the system further applies the Bayesian inference method to estimate the confidence probability of each smoothed trajectory point, and according to the historical observation, sensor state and environmental data of the current point, the Kalman filtering algorithm is used to predict the position and state of the current point by using the historical observation and the dynamics model of the control input ship, and according to the measurement accuracy of the sensor and the influence of the environmental data, a predicted state estimation value is calculated, and the prediction uncertainty, that is, the prediction error, is also obtained. The confidence value of each trajectory point is obtained, and a position confidence sequence is formed. The system takes the sequence as input, performs normalization processing through the transparency mapping function, directly converts the confidence level into the transparency form of the visualized trajectory, and the trajectory points with high confidence level present high opacity, and the trajectory points with low confidence level present high transparency. Through the visual rendering technology, the system projects the trajectory into the three-dimensional virtual scene or two-dimensional interface in the color and transparency combination mode, generates an intuitive and usable preliminary decision support view, identifies which trajectory sections have high credibility and which sections need to be focused on. If the transparency of part of the trajectory points is lower than the system preset threshold, it means that the data credibility of this section is insufficient, and the interpolation algorithm is automatically called to supplement the missing points or optimize the abnormal points according to the context spatiotemporal sequence, so as to enhance the continuity of the trajectory. The completed trajectory is subjected to overall trend extraction through sequence analysis, so as to realize intelligent identification of the trajectory change trend. The trajectory change trend and the optimized trajectory view are combined by using the view superposition method, and a decision support view which can directly reflect the trajectory confidence distribution and motion trend is output. Taking the case that a ship sails to the nearshore area and encounters signal interference, resulting in a decrease in the confidence of the positioning data in part of the section, as an example, the system automatically displays the trajectory of this section as high transparency, and fills in the missing information through the interpolation algorithm. The optimized trajectory is superimposed with the overall trend, which helps the port logistics scheduling center to master the ship dynamics in real time. The technical solution cooperates the trajectory data processing algorithm with the visual presentation means, constructs a decision support system integrating trajectory reliability intuitive expression and trend insight, significantly improves the intelligent level of port scheduling and ship safety management, overcomes the problems in the prior art that the trajectory credibility is difficult to be quantitatively and intuitively expressed, and the abnormal sections are difficult to be positioned and completed, and has outstanding substantial features and practical engineering application value.

[0042] Step S6: According to the decision support view, update the port scheduling plan, and generate an optimized unloading time sequence.

[0043] Specifically, step S6 above uses multi-dimensional data such as trajectory confidence, location prediction, and cargo loading / unloading demand generated by the decision support view to link in real time with business data such as the port scheduling system's operational status, logistics vessel entry and exit dynamics, and berth utilization. It automatically compares and analyzes parameters such as vessel arrival, berthing, departure times, and operation durations calculated by the trajectory model with the current scheduling plan. By evaluating constraints such as resource allocation, equipment availability, and berth layout for loading / unloading tasks, and using optimization methods such as genetic algorithms, it iteratively solves the vessel queue, operation sequence, and resource allocation scheme through multiple rounds to form a new loading / unloading time sequence. Efficient synchronization and feedback are achieved between data points. Vessel predicted paths and confidence data are intuitively displayed to the scheduling system through the decision support view. Each loading / unloading operation node is dynamically adjusted under the algorithm's guidance, and scheduling instructions are distributed to the port in real time. Various operational terminals rapidly reconstruct scheduling plans based on real-time changes in actual ship dynamics and predicted trajectories, improving port operational efficiency and resource utilization. Taking container ship berthing as an example, the system automatically collects the predicted arrival time and berth arrangement, determines the optimal operational sequence, allocates scheduling resources to the berths and equipment combinations with the highest operational efficiency, and provides immediate feedback and adjustments for emergencies such as equipment failures and abnormal weather, thereby reducing ship waiting time and improving loading and unloading efficiency. Among the technical features, trajectory calculation and scheduling optimization algorithms are closely integrated to jointly support real-time optimization of port scheduling plans. The synergistic effect of technical and algorithmic features solves the practical problems of dynamic port scheduling and cargo loading and unloading sequence optimization under ship signal interruption conditions, achieving beneficial technical effects such as improving port operational efficiency, reducing resource idleness, and optimizing overall logistics timeliness.

[0044] The present invention also provides a shipping logistics tracking system for implementing the above-described method, such as... Figure 2 As shown, the system includes:

[0045] The generation unit is used to acquire ship sensor data and external environment data, determine the signal interruption interval through time series analysis, and extract the ship's last navigation state before the interruption from the historical navigation trajectory database based on the signal interruption interval to generate an initial state vector.

[0046] The correction unit is used to estimate the ship's position during the signal interruption period by combining the initial state vector with real-time sea state data and weather data, and obtain the estimated position sequence. If the deviation between the estimated position sequence and the historical trajectory exceeds a preset threshold, the port scheduling data and the AIS data of nearby ships are fused to correct the estimated position sequence and obtain the corrected position sequence.

[0047] The prediction unit is configured to predict a future transportation path of the ship by a particle filter algorithm through a corrected position sequence to generate a prediction path set; and construct a ship movement trajectory in a three-dimensional virtual scene, distinguish real position data from calculated position data by color coding, and generate a visual trajectory model.

[0048] The mapping unit is configured to generate a decision support view by transparency mapping through the visual trajectory model in combination with the confidence of the calculated position sequence.

[0049] The application also provides a computer readable storage medium, which stores instructions, and the instructions are executed by a processor to implement the above method.

[0050] In summary, the application obtains ship sensor and external environment data, identifies signal interruption intervals by using a time series analysis method, extracts the last sailing state of the ship from a historical sailing trajectory database to form an initial state vector, combines real-time sea conditions and weather data, performs state estimation by Kalman filtering, calculates the position of the ship during signal interruption, and corrects the position when the position calculation deviation is too large by integrating port scheduling and AIS data of nearby ships; the corrected ship position data is used to predict the future path of the ship by a particle filter algorithm to form multiple prediction path sets; in a three-dimensional virtual scene, the prediction path and the actual position are combined to generate a visual trajectory model, and the real position and the calculated position are distinguished by color coding to visually display the movement trajectory of the ship. Meanwhile, the confidence of the calculated position sequence is displayed by transparency mapping to further provide a visual view for decision support; the decision support view is combined to optimize the port scheduling plan to generate a new loading and unloading time sequence, thereby improving the operation efficiency of the port; the whole process gradually advances from accurate detection of signal interruption, state estimation, path prediction, three-dimensional visualization, and decision support to form a closed-loop feedback mechanism, ensuring efficient scheduling and accurate decision-making of shipping logistics in complex environments; the above scheme ensures the efficiency and reliability of ship position calculation and future path prediction in complex sea conditions and weather environments during signal interruption, improves the accuracy of logistics ship position calculation and the flexibility of port scheduling, and significantly optimizes the efficiency and safety of the overall shipping logistics.

[0051] The above specific embodiments further illustrate the purpose, technical solutions and advantages of the application, and it should be understood that the above description is only a specific embodiment of the application and is not used to limit the protection scope of the application, and any modification, equivalent replacement, improvement, etc. within the spirit and principles of the application should be included in the protection scope of the application.

Claims

1. A method of shipping logistics tracking, characterized by, The method comprises: Step S1: acquiring ship sensor data and external environment data, determining a signal interruption interval through time series analysis, extracting the last sailing state of the ship before the interruption from a historical sailing trajectory database according to the signal interruption interval, and generating an initial state vector; Step S2: calculating the ship position during the signal interruption interval through state estimation by combining the initial state vector, real-time sea state data and weather data, obtaining a calculated position sequence; if the deviation of the calculated position sequence from the historical trajectory exceeds a preset threshold, fusing port scheduling data and AIS data of nearby ships to correct the calculated position sequence, and obtaining a corrected position sequence; Step S3: predicting the future transportation path of the ship through a particle filtering algorithm by using the corrected position sequence, and generating a predicted path set; Step S4: obtaining real position data and calculated position data from the predicted path set, constructing a ship motion trajectory in a three-dimensional virtual scene, distinguishing the real position data from the calculated position data through color coding, and generating a visual trajectory model; Step S5: generating a decision support view through transparency mapping by using the visual trajectory model and the confidence of the calculated position sequence; Step S6: updating the port scheduling plan according to the decision support view, and generating an optimized loading and unloading time sequence; Wherein, the corrected position sequence is obtained by: if the deviation of the calculated position sequence from the historical trajectory exceeds a preset threshold, calculating the deviation value by comparing the calculated position sequence with the historical trajectory, obtaining the deviation distribution characteristics; optimizing the calculated position sequence through Kalman filtering algorithm by fusing the port scheduling data and the AIS data of nearby ships, obtaining the preliminary corrected position sequence; adjusting the position sequence deviation through particle filtering algorithm according to the preliminary corrected position sequence and the ship motion state, obtaining the optimized corrected position sequence; if the consistency of the optimized corrected position sequence with the historical trajectory is lower than the preset threshold, updating the position sequence by extracting the timestamp in the port scheduling data and the speed information in the AIS data through multi-source data integration, obtaining the high-precision position sequence; analyzing the data real-time according to the high-precision position sequence, judging whether the position sequence meets the real-time updating requirement, obtaining the real-time corrected position sequence; smoothing the position sequence through weighted average method according to the real-time corrected position sequence and the trajectory consistency, obtaining the final corrected position sequence; The decision support view is generated through transparency mapping, which comprises: obtaining position sequence data through the visual trajectory model; smoothing the position sequence to obtain smoothed trajectory data; calculating the confidence of each position point according to the smoothed trajectory data to obtain a position confidence sequence; mapping the position confidence sequence to visual transparency value through a transparency mapping function to obtain transparency mapping data; generating a trajectory view according to the transparency mapping data through visual rendering technology to obtain a preliminary decision support view.

2. The method of claim 1, wherein, The signal interruption interval is determined through time series analysis, which comprises: The sensor data and the environment data are acquired through a ship sensor and an external environment acquisition device, a time point of signal interruption in a data stream is determined through a time series analysis method, if the sensor data or the environment data is lower than a preset threshold, a sliding window algorithm is used to detect continuous abnormal points to obtain a starting time point of signal interruption, a time difference calculation method is used to acquire a duration of interruption according to the starting time point of signal interruption to determine an ending time point of signal interruption, a data fusion technology is used to integrate the sensor data and the environment data to generate a fusion data set, if the environment data in the fusion data set fluctuates beyond a preset threshold, a decision tree algorithm is used to analyze the correlation between the environment data and the signal interruption to determine an interruption interval classification, a time series segmentation method is used to separate the interruption interval and a normal interval according to the interruption interval classification result to obtain an interruption interval sequence, and a time series analysis is used to extract features of the interruption interval sequence to generate time distribution features of the interruption interval.

3. The method of claim 1, wherein, An initial state vector is generated, including: acquiring, from a historical voyage trajectory database, voyage data of a ship recorded last before signal interruption to obtain a record set containing a voyage speed, a heading and a position, if the record set is not empty, sorting the record set according to a last record time to extract a record with a latest time stamp to obtain a last voyage state, and converting the voyage speed, the heading and the position in the last voyage state into a standardized initial state vector through a vector conversion method, and smoothing the initial state vector through a Kalman filtering algorithm to obtain an optimized state vector.

4. The method of claim 1, wherein, The ship position during signal interruption is calculated through state estimation to obtain a calculated position sequence, including: acquiring the initial state vector, real-time sea state data and real-time weather data to obtain a synchronized state data set through data time synchronization processing, performing state estimation through a Kalman filtering algorithm in combination with a ship motion model based on the synchronized state data set to obtain a preliminary state estimation value, adjusting the state estimation through the updating step of the Kalman filtering algorithm if the deviation of the preliminary state estimation value from the sensor data fusion exceeds a preset threshold to obtain a corrected state estimation value, calculating a position at a next time through the prediction step of the Kalman filtering algorithm based on the corrected state estimation value and the fusion of environmental dynamic influences to obtain a single-step position prediction value, iteratively performing the prediction and updating steps of the Kalman filtering algorithm based on the single-step position prediction value and the ship motion model to obtain a continuous position sequence, and verifying the sequence consistency through data time synchronization processing if the time span of the continuous position sequence exceeds the signal interruption duration to obtain a final calculated position sequence, and optimizing the sequence smoothness based on the final calculated position sequence, the real-time sea state data and the weather data to obtain a smoothed calculated position sequence.

5. The method of claim 1, wherein, The ship future transportation path is predicted by the particle filtering algorithm, and a prediction path set is generated, including: obtaining a corrected position sequence, obtaining continuous ship motion states through data cleaning and interpolation processing, extracting dynamic characteristics from the ship motion states through the particle filtering algorithm, determining an initial prediction set of the future transportation path, adjusting the weight of the particle filtering algorithm through environmental influencing factors, and obtaining an optimized prediction path set; if the path prediction accuracy of the prediction path set is lower than a preset threshold, iteratively updating the particle state to obtain a new prediction path set; according to the optimized prediction path set, an evaluation index of a path optimization target is calculated to determine an optimal transportation path; through real-time data input, the ship motion state is updated to obtain a dynamically adjusted prediction path set; and through the evaluation result of the path prediction accuracy, a final prediction path set is output.

6. The method of claim 1, wherein, The visual trajectory model is generated, including: obtaining real position data and calculated position data from the prediction path set, obtaining a unified format position data set through data fusion processing, analyzing the unified format position data set, generating a ship motion trajectory through a trajectory generation algorithm, and determining an initial trajectory coordinate set; in a three-dimensional virtual scene, a three-dimensional trajectory model is constructed according to the initial trajectory coordinate set through scene rendering technology, and a rendered trajectory framework is obtained; for the rendered trajectory framework, the real position data and the calculated position data are distinguished through color coding, a color-coded trajectory is generated, and a visual trajectory model is obtained; if the trajectory points in the visual trajectory model deviate from the prediction path set by more than a preset threshold, the trajectory point coordinates are recalculated through position data analysis, and the visual trajectory model is updated; according to the real-time position data of the ship, the visual trajectory model is adjusted through trajectory dynamic updating technology, and a dynamically updated three-dimensional trajectory model is obtained; through scene rendering technology, light and shadow processing and perspective optimization are performed on the dynamically updated three-dimensional trajectory model, and a final visual trajectory model is obtained.

7. The method of claim 1, wherein, The decision support view is generated through transparency mapping, including: if the transparency of the trajectory points in the preliminary decision support view is lower than a preset threshold, the missing position data is supplemented through an interpolation algorithm to obtain an optimized trajectory view; according to the optimized trajectory view, the trajectory change trend is extracted through sequence analysis technology to obtain trajectory trend data; the trajectory trend data and the optimized trajectory view are fused through a view superposition method to obtain a final decision support view.

8. A shipping logistics tracking system for implementing the method of any one of claims 1-7, characterized by, The system comprises: A generation unit is configured to obtain ship sensor data and external environment data, determine a signal interruption interval through time series analysis, extract the last sailing state of the ship before the interruption from a historical sailing trajectory database according to the signal interruption interval, and generate an initial state vector; A correction unit is configured to calculate the position of the ship during the signal interruption interval through state estimation by combining the initial state vector, real-time sea state data and weather data, obtain a calculated position sequence, and correct the calculated position sequence by fusing port scheduling data and AIS data of nearby ships if the deviation between the calculated position sequence and the historical trajectory exceeds a preset threshold to obtain a corrected position sequence. The prediction unit is configured to predict a future transport path of the ship by a particle filter algorithm through a correction of the position sequence, and generate a prediction path set; construct a ship motion trajectory in a three-dimensional virtual scene, distinguish real position data from calculated position data by color coding, and generate a visual trajectory model; The mapping unit is configured to generate a decision support view by transparency mapping through the visual trajectory model in combination with the confidence of the calculated position sequence.

9. A computer-readable storage medium having stored thereon instructions, the computer-readable storage medium comprising: The instructions, when executed by the processor, implement the method of any one of claims 1-7.

Images