Multi-mode AIGC cold chain cargo state tracking and tracing method and system

Abstract

The invention relates to the technical field of cargo state analysis, in particular to a multi-mode AIGC cold chain cargo state tracking and tracing method and system. The method comprises the following steps: acquiring real-time state parameters of multi-dimensional cold chain goods; abnormal data cleaning is carried out on the real-time state parameters of the multi-dimensional cold chain goods, heterogeneous state fusion is carried out, and therefore goods heterogeneous state features are obtained; carrying out cargo vibration time-frequency mining on the multi-dimensional cold chain cargo real-time state parameters, and carrying out transportation scene identification so as to obtain a cargo transportation scene; carrying out multi-time-point position calculation according to a cargo transportation scene, carrying out real-time scene modeling, and constructing a dynamic scene transportation track map; and carrying out one-by-one state position matching on the dynamic scene transportation track map based on the cargo heterogeneous state features, carrying out dynamic state rendering, and constructing a dynamic cargo transportation twin model. According to the method, the safety of cold chain cargo transportation is improved, and the real-time state of the cargo is efficiently analyzed.

Description

Technical Field

[0001] The present invention relates to the technical field of cargo status analysis, and in particular to a multimodal AIGC cold chain cargo status tracking and tracing method and system. Background Art

[0002] With the growing demand for cold chain logistics around the world, cold chain management has become increasingly important, especially in the transportation of high-value goods such as medicines, food, and electronic products that have strict temperature and humidity requirements. The core of cold chain logistics is to ensure that the goods are always under specific environmental conditions during transportation to prevent temperature fluctuations, humidity changes, or other external factors from affecting the quality of the goods. Traditional cold chain logistics management methods mainly rely on manual monitoring, manual recording, and regular inspections, but these traditional methods have many deficiencies in real-time, accuracy, and efficiency, and cannot fully respond to complex transportation scenarios and emergencies.

[0003] In modern cold chain logistics, with the rapid development of intelligent technology, especially the widespread application of the Internet of Things (IoT), artificial intelligence (AI) and data analysis technology, the management mode of cold chain logistics is gradually shifting towards intelligence and automation. The widespread use of sensor equipment enables the temperature, humidity, location and other conditions of goods to be monitored and recorded in real time. However, the status data provided by a single sensor often cannot fully reflect the true status of the goods when faced with complex transportation environments and various interference factors. For example, changes in temperature and humidity cause errors due to different transportation environments, vibrations of transportation tools, or differences in the characteristics of the goods themselves. These single-dimensional monitoring data cannot effectively reveal the potential risks in cold chain transportation.

[0004] Existing cold chain tracking and tracing technologies still face many challenges. For example, how to effectively process and fuse large amounts of data from different types of sensors, how to build accurate anomaly detection models, and how to achieve intelligent prediction in complex cold chain scenarios are still bottlenecks in the current technological development. Therefore, a new cold chain cargo status tracking and tracing method that integrates advanced artificial intelligence technology and multimodal data analysis is urgently needed to improve the accuracy, real-time and intelligent level of cold chain management and ensure that the cargo is always in the best condition during transportation. Summary of the invention

[0005] In order to solve the above technical problems, the present invention proposes a multimodal AIGC cold chain cargo status tracking and tracing method and system to solve at least one of the above technical problems.

[0006] To achieve the above object, the present invention provides a multi-modal AIGC cold chain cargo status tracking and tracing method, comprising the following steps: Step S1: obtaining multi-dimensional real-time status parameters of cold chain goods; performing abnormal data cleaning on the multi-dimensional real-time status parameters of cold chain goods, and performing heterogeneous state fusion to obtain heterogeneous state characteristics of goods; Step S2: mining the cargo vibration time-frequency of multi-dimensional cold chain cargo real-time status parameters and identifying the transportation scene to obtain the cargo transportation scene; Step S3: Perform multi-time point location calculation according to the cargo transportation scenario, perform real-time scenario modeling, and construct a dynamic scenario transportation trajectory map; Step S4: Based on the heterogeneous state characteristics of the goods, the dynamic scene transportation trajectory map is matched with the state position one by one, and dynamic state rendering is performed to build a dynamic cargo transportation twin model; Step S5: Detect abnormal status of cargo in the dynamic cargo transportation twin model and trace the abnormal location, thereby generating the starting traceability location of the abnormal status; Step S6: Perform abnormal fault risk diagnosis based on the abnormal state starting traceability position, make intelligent fault decisions, and build a cargo intelligent state decision model.

[0007] The present invention obtains multi-dimensional real-time status parameters of cold chain goods, cleans abnormal data and integrates heterogeneous states to obtain heterogeneous state characteristics of goods, effectively monitors the state of goods, identifies and cleans abnormal data, integrates multi-dimensional data, and improves the reliability and stability of the cold chain transportation process. Through vibration time-frequency mining and transportation scene recognition, the cargo transportation scene is obtained, different transportation scenes are identified, and the understanding and monitoring of the cargo state are improved, providing a basis for subsequent position calculation and modeling. According to the cargo transportation scene, multi-time point position calculation and real-time scene modeling are performed, and a dynamic scene transportation trajectory map is constructed to realize real-time monitoring and dynamic modeling of the cargo transportation process, and improve the accuracy and Timeliness: Dynamic state position matching and rendering are performed based on the heterogeneous state characteristics of the goods, and a dynamic cargo transportation twin model is constructed to achieve accurate matching and presentation of the cargo state, improve the understanding and monitoring level of the cargo transportation process, perform abnormal state detection and position traceability on the dynamic cargo transportation twin model, generate the abnormal state starting traceability position, and achieve timely identification and positioning of abnormal cargo conditions, providing data support for subsequent fault risk diagnosis, and perform abnormal fault risk diagnosis and intelligent decision-making based on the abnormal state starting traceability position, and build a cargo intelligent state decision-making model, which can quickly diagnose and make intelligent decisions on cargo failure risks, improve cargo transportation safety and efficiency, and reduce potential risks.

[0008] Preferably, step S1 specifically comprises the following steps: Step S11: Monitor the entire cold chain cargo transportation process based on sensors to obtain multi-dimensional real-time status parameters of the cold chain cargo; Step S12: performing abnormal data cleaning on multi-dimensional cold chain cargo real-time status parameters, thereby generating cleaning and optimization cargo real-time status parameters; Step S13: performing temperature distribution analysis on the real-time state parameters of the cleaning optimized cargo to generate cargo temperature distribution characteristic data; Step S14: Calculate the humidity fluctuation of the real-time state parameters of the cleaning optimization cargo to obtain the cargo humidity fluctuation parameters; Step S15: heterogeneous state fusion is performed on the cargo temperature distribution characteristic data and cargo humidity fluctuation parameters to obtain cargo heterogeneous state characteristics.

[0009] The present invention monitors the status of the entire cold chain cargo transportation process through sensors, obtains multi-dimensional real-time status parameters, realizes all-round monitoring of the cargo status, improves the visualization and real-time performance of the cargo transportation process, cleans the multi-dimensional real-time status parameters for abnormal data, generates cleaning-optimized cargo real-time status parameters, improves data quality, reduces interference factors, ensures the accuracy and reliability of subsequent analysis, performs temperature distribution analysis on the cleaning-optimized real-time status parameters, generates cargo temperature distribution characteristic data, provides cargo temperature change trends based on data analysis, and provides an important basis for subsequent status characteristic analysis, calculates humidity fluctuations on the cleaning-optimized real-time status parameters, obtains cargo humidity fluctuation parameters, analyzes humidity fluctuations, and provides data support of an additional dimension for cargo status evaluation, performs heterogeneous state fusion on cargo temperature distribution characteristic data and cargo humidity fluctuation parameters, obtains cargo heterogeneous state characteristics, integrates multi-dimensional data, improves the understanding and analysis depth of cargo status, and provides a more comprehensive basis for subsequent status tracking and abnormality detection.

[0010] Preferably, the specific steps of step S2 are: Step S21: mining the cargo vibration time-frequency of multi-dimensional cold chain cargo real-time status parameters to extract cargo vibration time-frequency features; Step S22: Calculating the vibration frequency of the cargo based on the time-frequency characteristics of cargo vibration to generate cargo vibration frequency parameters; Step S23: analyzing the vibration amplitude characteristics of the cargo vibration time-frequency characteristics to obtain the vibration amplitude characteristics; Step S24: performing time-series vibration fitting on the vibration frequency parameters and vibration amplitude characteristics of the goods to construct a time-series vibration curve for the goods transportation; Step S25: Performing transportation scene recognition on the cargo transportation time series vibration curve to obtain the cargo transportation scene.

[0011] The present invention mines the vibration time-frequency of cold chain goods in real-time status parameters of multi-dimensional goods, extracts the vibration time-frequency characteristics of goods, helps to deeply understand the vibration of goods during transportation, extracts the vibration time-frequency characteristics of goods to provide basic data for subsequent analysis and processing, helps to track the status of goods and identify abnormal situations, calculates the vibration frequency of the vibration time-frequency characteristics of goods, generates vibration frequency parameters of goods, helps to accurately describe the frequency characteristics of vibration of goods, and generates vibration frequency parameters of goods to help understand the vibration frequency range to which goods are subjected, and provides a basis for subsequent vibration analysis and control, analyzes the vibration amplitude characteristics of the vibration time-frequency characteristics of goods, obtains the vibration amplitude characteristics, helps to evaluate the intensity and amplitude changes of vibration of goods, and extracts vibration The vibration amplitude characteristics help identify the vibration amplitude range of the goods, provide a reference for the evaluation and protection of the goods status, perform time-series vibration fitting on the vibration frequency parameters and vibration amplitude characteristics of the goods, and construct a time-series vibration curve for cargo transportation, which is helpful for comprehensive analysis of the vibration characteristics of the goods. The construction of the time-series vibration curve for cargo transportation can more comprehensively describe the vibration conditions of the goods at different stages of transportation, and provide data support for subsequent scene identification and traceability. The transportation scene identification is performed on the time-series vibration curve for cargo transportation to obtain the cargo transportation scene, which helps to understand the specific conditions and environment of the goods during transportation. The transportation scene identification helps monitor the cargo transportation path and environmental changes, realizes the tracking and traceability of the cargo status, and improves the controllability and safety of cold chain cargo transportation.

[0012] Preferably, step S3 specifically comprises the following steps: Step S31: extracting real-time positioning data of goods based on multi-dimensional real-time status parameters of cold chain goods; Step S32: performing multi-time point position calculation on the real-time positioning data of the cargo, thereby obtaining the cargo positioning coordinates at multiple time points; Step S33: performing key stop node analysis on the cargo positioning coordinates at multiple time points, thereby generating key nodes for cargo transportation; Step S34: Dynamically track the cargo location coordinates at multiple time points to construct a time-series dynamic cargo trajectory; Step S35: Visualizing the real-time trajectory of the time-series dynamic cargo trajectory according to the key nodes of cargo transportation, thereby constructing a dynamic trajectory visualization map; Step S36: Perform real-time scene modeling on the dynamic trajectory visualization map according to the cargo transportation scene points to construct a dynamic scene transportation trajectory map.

[0013] The present invention extracts real-time positioning data of goods based on real-time status parameters of multi-dimensional cold chain goods, which is helpful for real-time monitoring of the location information of goods. The extracted real-time positioning data of goods provides accurate basic data for subsequent position calculation and trajectory tracking, performs multi-time point position calculation on the real-time positioning data of goods, and obtains the positioning coordinates of goods at multiple time points, which is helpful for accurately recording the position of goods at different time points. The generation of cargo positioning coordinates at multiple time points helps to understand the position changes of goods during transportation, and provides data support for subsequent trajectory analysis. The key stop node analysis is performed on the cargo positioning coordinates at multiple time points, and the key nodes of cargo transportation are generated, which is helpful for identifying the key stop points of goods in the transportation process. The generation of key nodes of cargo transportation helps to track the transportation path and important nodes of goods, and provides a basis for subsequent trajectory analysis and scenario modeling. It can track the dynamic transport trajectory of the cargo and build a time-series dynamic cargo trajectory, which is helpful to track the transport trajectory of the cargo in real time. It can build a time-series dynamic cargo trajectory to show the changes in the transport trajectory of the cargo at different time points, help monitor the real-time status of the cargo transportation process, and visualize the real-time trajectory of the time-series dynamic cargo trajectory according to the key nodes of cargo transportation. It can build a dynamic trajectory visualization map, which is helpful to intuitively display the transport trajectory of the cargo. The real-time trajectory visualization map helps users to view the transportation status of the cargo in real time, improve the efficiency of cold chain cargo status tracking and monitoring, and perform instant scene modeling on the dynamic trajectory visualization map according to the cargo transportation scene points, and build a dynamic scene transportation trajectory map, which is helpful to comprehensively display the transportation scene of the cargo. The dynamic scene transportation trajectory map helps users to have a more comprehensive understanding of the transportation environment and scene changes of the cargo, and improve the visualization effect of cold chain cargo status tracking and traceability.

[0014] Preferably, the specific steps of step S4 are: Step S41: Perform multi-time point feature recognition on the heterogeneous state characteristics of the goods, and extract the state characteristics of each time node; Step S42: Performing time-series state evolution on the state characteristics of each time node, thereby generating cargo state evolution characteristics; Step S43: matching the cargo positioning coordinates at multiple time points one by one according to the state characteristics of each time node, thereby obtaining the state positioning of each time node; Step S44: According to the state positioning of each time node, the cargo state evolution characteristics are dynamically rendered on the dynamic scene transportation trajectory map to build a dynamic cargo transportation twin model.

[0015] The present invention performs multi-time point feature recognition on heterogeneous state features of goods and extracts the state features of each time node, which helps to fully understand the state performance of goods at different time points. The state features of each time node are extracted to provide detailed feature descriptions for subsequent state analysis and matching, which helps to identify changes and evolutions in the state of goods. The state features of each time node are subjected to time-series state evolution to generate goods state evolution features, which helps to describe the change process of the state of goods over time. The generated goods state evolution features help to understand the evolution law and trend of the state of goods, and provide basic data for state matching and rendering. According to the state features of each time node, the positioning coordinates of goods at multiple time points are matched one by one. State-position matching, obtaining the state positioning of each time node, helps to combine state characteristics with location information. Realizing the state positioning of each time node helps to accurately describe the state and position relationship of the goods at different time points, and provide accurate data support for subsequent state rendering. According to the state positioning of each time node, the state evolution characteristics of the goods are dynamically rendered on the dynamic scene transportation trajectory map to build a dynamic cargo transportation twin model, which helps to show the dynamic changes of the cargo status. Building a dynamic cargo transportation twin model combines the cargo status information with the transportation trajectory map to achieve dynamic visualization of cargo status changes and improve the information presentation effect of cold chain cargo status tracking and traceability.

[0016] Preferably, the specific steps of step S5 are: Step S51: Perform full-cycle transportation monitoring analysis on the dynamic cargo transportation twin model and extract full-cycle transportation monitoring response data; Step S52: Perform cargo abnormal status detection on the full-cycle transportation monitoring response data to generate cargo transportation abnormal status data; Step S53: Tracing back the abnormal state time length of the cargo transportation abnormal state data to identify the initial time point of the abnormal state; Step S54: Based on the initial time point of the abnormal state, the dynamic cargo transportation twin model is traced to the abnormal location, thereby generating the starting traceability location of the abnormal state.

[0017] The present invention performs full-cycle transportation monitoring and analysis on the dynamic cargo transportation twin model, extracts full-cycle transportation monitoring response data, helps to comprehensively monitor the transportation process of the cargo, extracts full-cycle transportation monitoring response data to provide comprehensive data support for subsequent abnormal state detection and traceability analysis, helps to identify abnormal situations in the transportation process, performs cargo abnormal state detection on the full-cycle transportation monitoring response data, generates cargo transportation abnormal state data, helps to timely discover and record abnormal situations in cargo transportation, generates cargo transportation abnormal state data to help identify abnormal state changes of cargo, and provides a data basis for subsequent abnormal state time length backtracing, performs abnormal state time length backtracing on the cargo transportation abnormal state data, identifies the initial time point of the abnormal state, helps to determine the starting time of the abnormal state, identifies the initial time point of the abnormal state to help accurately record the occurrence time of the abnormal state, and provides a time reference for subsequent abnormal position tracing, traces the abnormal position of the dynamic cargo transportation twin model based on the initial time point of the abnormal state, generates the abnormal state starting traceability position, helps to determine the starting position of the abnormal state, generates the abnormal state starting traceability position to help locate the starting position information of the abnormal state, and provides a positioning basis for subsequent abnormal state processing and problem solving.

[0018] Preferably, the specific steps of step S6 are: Step S61: performing abnormal fault risk diagnosis on the cargo transportation abnormal state data to generate abnormal fault diagnosis results; Step S62: When the abnormal fault diagnosis result is that the cargo can be transported normally, a cargo abnormal fault signal is generated; Step S63: When the abnormal fault diagnosis result is that normal transportation is not possible, the cargo fault emergency transportation route is planned based on the abnormal state starting traceability position, thereby generating an emergency transportation planning route; Step S64: Make intelligent fault decisions on abnormal fault signals of cargo and emergency transportation planning paths, and build a cargo intelligent status decision model.

[0019] The present invention performs abnormal fault risk diagnosis on abnormal state data of cargo transportation and generates abnormal fault diagnosis results, which is helpful to evaluate the risks and problems existing in cargo transportation. The abnormal fault diagnosis results are generated to provide a diagnostic basis for subsequent fault processing and decision-making, and help to take corresponding countermeasures in time. When the abnormal fault diagnosis result is that the cargo can be transported normally, a cargo abnormal fault signal is generated, which is helpful to identify the cargo that can continue to be transported, reduce false alarms and reduce the processing burden. The cargo abnormal fault signal is generated to help distinguish normal cargo from abnormal cargo, and provide a reference basis for subsequent transportation decisions. When the abnormal fault diagnosis result is that it cannot be transported normally, the cargo fault emergency transportation path planning is performed based on the abnormal state starting traceability position, and the emergency transportation planning path is generated, which is helpful to respond to emergency situations. The emergency transportation planning path is generated to help plan the optimal emergency transportation path, reduce the impact of the fault on cargo transportation, and ensure the safety and integrity of the cargo. Intelligent fault decision-making is performed on the cargo abnormal fault signal and the emergency transportation planning path, and a cargo intelligent state decision model is constructed, which is helpful to intelligently handle abnormal situations in cargo transportation. The cargo intelligent state decision model is constructed to help realize intelligent decision-making and processing of abnormal situations, improve the safety and reliability of cold chain cargo transportation, and reduce losses and risks.

[0020] Preferably, the specific steps of step S63 are: When the abnormal fault diagnosis result is that normal transportation is impossible, the abnormal state data of cargo transportation is subjected to multi-period fault evolution to generate abnormal fault evolution data of multiple time periods; Perform abnormal fault prediction on abnormal state evolution data in multiple time periods to generate time series cargo fault prediction data; The current transportation route time length is calculated based on the abnormal state starting traceability position, thereby obtaining the current cargo transportation route time length; According to the current cargo transportation route duration, the time series cargo failure prediction data is used to plan the cargo failure emergency transportation route, thereby generating an emergency transportation planning route.

[0021] The present invention helps to track the development process of abnormal status and understand the evolution law of faults through abnormal fault evolution data of multiple time periods, and tracks abnormal fault evolution data to provide a time series data basis for subsequent fault prediction and planning, help predict the fault conditions that will occur, perform abnormal fault prediction to identify the fault conditions that will occur in advance, help take preventive measures, reduce risks in cargo transportation, generate time series cargo fault prediction data to provide prediction results for subsequent emergency transportation planning, help formulate targeted fault handling plans, calculate the current cargo transportation path duration to help evaluate the current transportation status, understand the time required for cargo transportation, and provide a time reference for subsequent planning, and obtain the current cargo transportation path duration to provide basic data for emergency transportation planning, help determine the urgency and urgency of the emergency transportation path, perform emergency transportation planning on the fault prediction data according to the current transportation path duration, formulate flexible transportation plans according to actual conditions to cope with emergencies, and generate emergency transportation planning paths to help respond quickly when faults occur, ensure the safe arrival of goods at the destination, and minimize losses.

[0022] In this specification, a multimodal AIGC cold chain cargo status tracking and tracing system is provided, which is used to execute the multimodal AIGC cold chain cargo status tracking and tracing method as described above, including: The cargo heterogeneous state module is used to obtain the real-time state parameters of multi-dimensional cold chain cargo; perform abnormal data cleaning on the real-time state parameters of multi-dimensional cold chain cargo, and perform heterogeneous state fusion to obtain the cargo heterogeneous state characteristics; The transportation scenario module is used to mine the vibration time-frequency of multi-dimensional cold chain cargo real-time status parameters and identify the transportation scenario to obtain the cargo transportation scenario; The track map module is used to calculate the multi-time point positions according to the cargo transportation scenario, perform real-time scenario modeling, and build a dynamic scenario transportation track map; The twin model module is used to match the state positions of the dynamic scene transportation trajectory map one by one based on the heterogeneous state characteristics of the goods, and perform dynamic state rendering to build a dynamic cargo transportation twin model; The abnormal tracing module is used to detect abnormal cargo status in the dynamic cargo transportation twin model and trace the abnormal location, thereby generating the starting traceability location of the abnormal status; The intelligent status decision module is used to diagnose abnormal fault risks based on the abnormal status starting traceability location, make intelligent fault decisions, and build a cargo intelligent status decision model.

[0023] The present invention helps to fully understand the status of cold chain goods by obtaining multi-dimensional real-time status parameters of cold chain goods, provides rich data support, performs abnormal data cleaning and heterogeneous state fusion to improve data quality, obtains more accurate and reliable heterogeneous state characteristics of goods, and provides a reliable basis for subsequent analysis and decision-making. Cargo vibration time-frequency mining and transportation scene recognition help to understand the behavioral characteristics of goods in different transportation environments, and obtains cargo transportation scenes to help understand the situations faced during cargo transportation, providing important information for subsequent trajectory map modeling. Multi-point position calculation and instant scene modeling are performed according to the cargo transportation scene to achieve real-time monitoring and modeling of the cargo transportation process. Construction of dynamic scene transportation trajectory map helps to visualize the transportation path and state changes of goods, and provides data support for subsequent twin models. The status characteristics perform status position matching and dynamic status rendering on the dynamic scene transportation trajectory map to achieve real-time tracking and display of the cargo status. Building a dynamic cargo transportation twin model helps to compare the actual transportation situation with the model, improve the understanding and prediction of the cargo status, and perform abnormal status detection and abnormal position tracing on the dynamic cargo transportation twin model to quickly discover abnormal situations and determine the location where the abnormality occurred. Generating the starting traceability position of the abnormal state helps to track the root cause of the abnormal situation and provide a decision-making basis for subsequent intelligent state decisions. Abnormal fault risk diagnosis and intelligent fault decision-making based on the starting traceability position of the abnormal state improve the efficiency and accuracy of handling abnormal situations. Building a cargo intelligent state decision-making model helps to intelligently handle abnormal situations in cargo transportation and ensure the safety and stability of cargo transportation. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic diagram of the steps of a multi-modal AIGC cold chain cargo status tracking and tracing method of the present invention; Figure 2 Detailed implementation flow chart of step S1; Figure 3 Detailed implementation flow chart of step S2; Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0025] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0026] The present application example provides a multimodal AIGC cold chain cargo status tracking and tracing method and system. The execution subject of the multimodal AIGC cold chain cargo status tracking and tracing method and system includes but is not limited to the following: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio and image management system, an information management system, and a cloud data management system.

[0027] See also Figures 1 to 4 The present invention provides a multi-modal AIGC cold chain cargo status tracking and tracing method, and the multi-modal AIGC cold chain cargo status tracking and tracing method comprises the following steps: Step S1: obtaining multi-dimensional real-time status parameters of cold chain goods; performing abnormal data cleaning on the multi-dimensional real-time status parameters of cold chain goods, and performing heterogeneous state fusion to obtain heterogeneous state characteristics of goods; Step S2: mining the cargo vibration time-frequency of multi-dimensional cold chain cargo real-time status parameters and identifying the transportation scene to obtain the cargo transportation scene; Step S3: Perform multi-time point location calculation according to the cargo transportation scenario, perform real-time scenario modeling, and construct a dynamic scenario transportation trajectory map; Step S4: Based on the heterogeneous state characteristics of the goods, the dynamic scene transportation trajectory map is matched with the state position one by one, and dynamic state rendering is performed to build a dynamic cargo transportation twin model; Step S5: Detect abnormal status of cargo in the dynamic cargo transportation twin model and trace the abnormal location, thereby generating the starting traceability location of the abnormal status; Step S6: Perform abnormal fault risk diagnosis based on the abnormal state starting traceability position, make intelligent fault decisions, and build a cargo intelligent state decision model.

[0028] The present invention obtains multi-dimensional real-time status parameters of cold chain goods, cleans abnormal data and integrates heterogeneous states to obtain heterogeneous state characteristics of goods, effectively monitors the state of goods, identifies and cleans abnormal data, integrates multi-dimensional data, and improves the reliability and stability of the cold chain transportation process. Through vibration time-frequency mining and transportation scene recognition, the cargo transportation scene is obtained, different transportation scenes are identified, and the understanding and monitoring of the cargo state are improved, providing a basis for subsequent position calculation and modeling. According to the cargo transportation scene, multi-time point position calculation and real-time scene modeling are performed, and a dynamic scene transportation trajectory map is constructed to realize real-time monitoring and dynamic modeling of the cargo transportation process, and improve the accuracy and Timeliness: Dynamic state position matching and rendering are performed based on the heterogeneous state characteristics of the goods, and a dynamic cargo transportation twin model is constructed to achieve accurate matching and presentation of the cargo state, improve the understanding and monitoring level of the cargo transportation process, perform abnormal state detection and position traceability on the dynamic cargo transportation twin model, generate the abnormal state starting traceability position, and achieve timely identification and positioning of abnormal cargo conditions, providing data support for subsequent fault risk diagnosis, and perform abnormal fault risk diagnosis and intelligent decision-making based on the abnormal state starting traceability position, and build a cargo intelligent state decision-making model, which can quickly diagnose and make intelligent decisions on cargo failure risks, improve cargo transportation safety and efficiency, and reduce potential risks.

[0029] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of a multi-modal AIGC cold chain cargo status tracking and tracing method of the present invention. In this example, the steps of the multi-modal AIGC cold chain cargo status tracking and tracing method include: Step S1: obtaining multi-dimensional real-time status parameters of cold chain goods; performing abnormal data cleaning on the multi-dimensional real-time status parameters of cold chain goods, and performing heterogeneous state fusion to obtain heterogeneous state characteristics of goods; In this embodiment, a variety of sensors are installed on cold chain transport vehicles and goods, including temperature sensors, humidity sensors, GPS locators, vibration sensors, etc. These sensors should be able to collect the state parameters of the goods in real time. Ensure that each sensor can send data to the central data storage system through a wireless network (such as GSM, LoRa or Wi-Fi). The data format of each sensor should be a unified JSON format for subsequent processing. Set the data collection frequency, such as recording data every minute or every ten minutes, to ensure that the state changes of the goods during transportation can be captured. Real-time data transmission is achieved through a data integration platform (such as Apache Kafka or MQTT protocol), and the state parameters from different sensors are aggregated into a central database (such as MySQL, MongoDB or InfluxDB). In the data storage system, the data from different sensors are integrated to form a comprehensive data set containing multi-dimensional state parameters. This data set should include information such as temperature, humidity, location, vibration, etc. at each time point. Set the data table structure to ensure the readability and queryability of the data, which is convenient for subsequent data cleaning and analysis. Choose appropriate anomaly detection techniques, such as the Z-score method, IQR (interquartile range) method, or machine learning models (such as isolation forests, support vector machines) to analyze the data and identify abnormal data points. Determine the threshold for anomalies, which is usually set based on the statistical characteristics of historical data. For example, for temperature data, set a normal range based on historical statistical analysis. Check the parameters of the integrated data set one by one to filter out abnormal data points that exceed the set threshold. For detected outliers, choose to delete them directly, replace them with the mean or median, or use interpolation to repair them. Record the type and number of abnormal data processed during the cleaning process for subsequent analysis and reporting.Store the cleaned data back into the database to ensure data integrity and consistency, generate a cleaning report, record the processing process and results of abnormal data in detail, and provide a basis for subsequent analysis. Define heterogeneous state features, including state parameters (such as temperature, humidity, position, vibration, etc.) obtained from different sensors, and map them into a unified feature space to ensure that different features have the same metric. Set a fusion strategy, such as using weighted average method, principal component analysis (PCA) or feature selection technology to fuse state features of different dimensions. Fuse the cleaned data and combine each state parameter into a comprehensive state feature through set weights or algorithms. Use Python's NumPy and Pandas libraries for data processing and calculation to generate a fused state feature dataset, ensuring that each time point corresponds to a complete heterogeneous state feature. Store the fused cargo heterogeneous state features in the database to form a unified data output for subsequent analysis and visualization. Generate a fusion feature report to record the fusion process and results for subsequent use.

[0030] Step S2: mining the cargo vibration time-frequency of multi-dimensional cold chain cargo real-time status parameters and identifying the transportation scene to obtain the cargo transportation scene; In this embodiment, data collected by the vibration sensor are identified from the multi-dimensional real-time status parameters of cold chain goods extracted in the previous step. These data are usually recorded in the form of time series, including vibration amplitude and frequency at different time points, to ensure the integrity and continuity of the data, handle missing values ​​and outliers, and facilitate subsequent analysis. The vibration data is preprocessed, including denoising, smoothing and normalization. A filter (such as a low-pass filter) is used to remove high-frequency noise to ensure the accuracy of the data. Interpolation is used to handle missing values ​​to ensure the integrity of the data set for subsequent time-frequency analysis. An appropriate time-frequency analysis method is selected, such as short-time Fourier transform (STFT), wavelet transform or Hilbert-Huang transform (HHT), to extract the time-frequency characteristics of the vibration signal. The window size and overlap ratio of the analysis are determined to balance the time resolution and frequency resolution. The vibration signal is transformed using the selected time-frequency analysis method to extract the frequency components and their time-varying characteristics. This process uses Python's scipy and pywt The library is used to facilitate the implementation of short-time Fourier transform or wavelet transform, and a time-frequency feature matrix is ​​generated, in which each column represents a frequency component in a time period and each row represents a different frequency. The most representative time-frequency features for subsequent scene recognition are selected by calculating the importance of the features (such as using variance analysis or principal component analysis). The dimensionality reduction process uses principal component analysis (PCA) to reduce the computational complexity and improve the efficiency of subsequent models. An appropriate scene recognition model is selected, such as support vector machine (SVM), random forest or deep learning model (such as convolutional neural network (CNN)), to classify transportation scenes according to the extracted time-frequency features. Training data is prepared, including labeled samples of different transportation scenes (such as smooth transportation, bumpy transportation, emergency braking, etc.) for model learning, and the extracted time-frequency features are used as input data. According to the data, the selected model is trained, and the hyperparameters of the model are tuned through cross-validation, grid search and other methods to improve the accuracy and generalization ability of the model. The model performance is evaluated, and indicators such as accuracy, recall and F1-score are used to ensure that the model can effectively distinguish different transportation scenarios. The new vibration time-frequency features are input into the trained model for transportation scene recognition. The model will output the category of the current transportation scene, record the recognition results, and associate them with other real-time status parameters to form complete transportation scene data. The identified transportation scene information is stored in the database, including timestamps, scene categories and related status parameters, to ensure that the data structure is clear and convenient for subsequent query and analysis. A transportation scene recognition report is generated, which summarizes the recognition process and results, and provides a basis for subsequent cold chain transportation monitoring and optimization.

[0031] Step S3: Perform multi-time point location calculation according to the cargo transportation scenario, perform real-time scenario modeling, and construct a dynamic scenario transportation trajectory map; In this embodiment, real-time location information is collected from the GPS positioning system of the transport vehicle to ensure the accuracy and timeliness of the data. The location data should include longitude, latitude and timestamp for subsequent analysis. The transport scene information extracted in the previous step is integrated to ensure that each location point is associated with the corresponding transport scene. An interpolation algorithm (such as linear interpolation or spline interpolation) is used to calculate the real-time location data to supplement the missing values ​​that appear in the data collection process. The calculation time interval is determined, for example, a new location point is generated every minute or every five minutes to ensure the smoothness and continuity of the trajectory. The collected GPS data is processed to calculate the location information of each time point and combine it with the transport scene data. The time series data is processed using Python's Pandas library to ensure that each time point has corresponding scene information. A multi-time point location data set is generated, including timestamps, longitude and latitude, scene categories and other related state parameters. An appropriate modeling method is selected, such as GIS-based dynamic scene modeling, or three-dimensional modeling software (such as Blender, Unity) is used to visualize the transport trajectory. The key elements of modeling are determined, including the geographic information of the scene, the transport path, the cargo status, etc. GIS is used to calculate the location information of each time point. Software (such as ArcGIS or QGIS) can be used to import multi-time point location data and generate dynamic transportation trajectory maps. Each location point can be represented by a marker and distinguished by different colors or symbols according to changes in the transportation scenario. If a 3D modeling tool is used, the trajectory data can be imported to create a 3D scene, including the height changes of the transportation path, the surrounding environment, etc., to enhance the visualization effect, and build a real-time update mechanism to automatically update the dynamic scene model when new location information or scene data is available. Use stream processing technology (such as Apache Kafka or Redis) to ensure the processing and update of real-time data streams, generate time series data graphs, show the dynamic changes of goods during transportation, and help analyze the evolution of transportation scenarios. Design the interface of the dynamic scene transportation trajectory map, including the map zoom function, timeline selection, scene information display, etc., to ensure that the user interface is friendly and can intuitively display the status changes during transportation. Determine the base map source of the map (such as OpenStreetMap, Google Maps) and integrate related APIs to achieve real-time loading and updating of the map. Use GIS Or visualization tools generate dynamic scenario transportation trajectory maps, combine multi-time point location data with scenario data to form interactive visualization maps, and provide export functions so that users can save the generated maps in different formats (such as PDF, pictures) for subsequent analysis or reporting. The dynamic scenario transportation trajectory maps and their related data are stored in the database for subsequent query and analysis, and the time of map generation, the data source used, and the version information of the map are recorded. A map update report is generated to record the content and reasons of each update, providing a basis for subsequent decision-making.

[0032] Step S4: Based on the heterogeneous state characteristics of the goods, the dynamic scene transportation trajectory map is matched with the state position one by one, and dynamic state rendering is performed to build a dynamic cargo transportation twin model; In this embodiment, from the heterogeneous state features of the goods extracted in the previous step, state parameters related to the transportation process (such as temperature, humidity, vibration, position, etc.) are collected and the integrity of the data is ensured, the location information in the dynamic scene transportation trajectory map is integrated, and each location point is ensured to have a corresponding timestamp and scene data. A timestamp-based matching algorithm is used to ensure that the state features of the goods correspond to the location data in the dynamic scene map one by one. A linear interpolation or nearest neighbor matching method is used to handle inconsistent time intervals. A matching threshold, such as the maximum allowable range of the time difference, is determined to ensure the accuracy of the matching. The extracted state features of the goods are matched one by one with the data of the dynamic scene transportation trajectory map. During the matching process, the state features and the corresponding geographic location at each time point are recorded to generate a new data set containing the state features of each location point to form a complete dynamic state data set for subsequent analysis and rendering. A suitable visualization tool or library (such as D3.js, Three.js) is selected. or Unity) for dynamic state rendering. These tools can process real-time data and generate interactive visualization effects. Determine the key elements of rendering, such as state parameters, location markers, dynamic effects, etc. Use the selected rendering tool to combine the matched state features with the dynamic scene transportation trajectory map to generate dynamic visualization effects. For example, use markers of different colors or shapes to indicate the state changes of the goods (such as normal, abnormal), and achieve dynamic effects, such as state parameters that change over time, real-time updated cargo locations, etc., to ensure that the visualization results are intuitive and interactive for users. Establish a real-time update mechanism to automatically update the dynamic state rendering results when new cargo state features or transportation trajectory data are available. Use stream processing technology (such as Apache Kafka, Redis) to ensure real-time data processing, generate dynamic rendering reports, record the content and changes of each update, and provide a basis for subsequent decision-making. Design the overall framework of the dynamic cargo transportation twin model, including the combination of physical models and virtual models, to ensure that the model can reflect the state changes in the actual transportation process, determine the input and output parameters of the model, such as real-time state characteristics, location data, environmental factors, etc., to ensure the integrity and effectiveness of the model, and use the previously matched state characteristics and dynamic rendering results to build a dynamic cargo transportation twin model. Use simulation software (such as AnyLogic, MATLAB Simulink) for modeling and simulation to ensure the real-time and accuracy of the model, and be able to dynamically adjust and optimize according to real-time data. Verify the constructed dynamic cargo transportation twin model, use historical data to compare with real-time data, evaluate the accuracy and reliability of the model, generate a model evaluation report, and record the performance indicators of the model to provide a basis for subsequent optimization and improvement.

[0033] Step S5: Detect abnormal status of cargo in the dynamic cargo transportation twin model and trace the abnormal location, thereby generating the starting traceability location of the abnormal status; In this embodiment, real-time cargo status parameters, including temperature, humidity, vibration, location, etc., are extracted from the dynamic cargo transportation twin model to ensure the integrity and accuracy of the data for subsequent analysis. Historical data and real-time data are integrated to form a comprehensive data set for abnormal state detection. Appropriate anomaly detection algorithms are selected, such as statistical methods (Z-score, IQR) or machine learning-based methods (isolation forest, support vector machine, etc.). These algorithms can effectively identify abnormal data points that deviate significantly from the normal state and determine the abnormal judgment criteria. Usually, a threshold range is set, such as temperature exceeds the set range, humidity changes too much, etc. The extracted state parameters are input into the selected anomaly detection model for real-time analysis to identify abnormal states. The model will output whether the state at each time point is abnormal, mark the type of anomaly and its severity, record the detection results, including the time, location and abnormal characteristics of the anomaly, form an abnormal state report, and adopt a traceability method based on a graph algorithm (such as Dijkstra algorithm or A* The algorithm can effectively find the source of the abnormal state in dynamic scenarios and determine the key parameters of traceability, including the timestamp, location data and related state characteristics of the abnormal state. According to the detected abnormal state, the graph algorithm is used to track the path of the abnormal state in the dynamic cargo transportation twin model. The algorithm will analyze the historical location data, find the initial location of the abnormal state, and record the traceability results, including the latitude and longitude of the starting location, the time of the abnormality, the state characteristics and other information. The generated abnormal state starting traceability location and related information are stored in the database to ensure the integrity and traceability of the data, generate an abnormality traceability report, record the starting location, time and characteristics of the abnormal state in detail, so as to facilitate subsequent analysis and decision-making, set up a regular review mechanism, and regularly review the process of anomaly detection and traceability, so as to continuously optimize the model and detection algorithm, improve the accuracy of abnormal state recognition, conduct in-depth analysis of the detected abnormal state, identify common abnormal patterns and potential causes, and use data visualization tools (such as Tableau Or Matplotlib) to display the distribution trend and frequency of abnormal conditions, and analyze the causes of abnormalities in combination with environmental factors and operating behaviors during transportation to provide a basis for subsequent improvements. According to the results of abnormal state detection and traceability analysis, specific optimization suggestions are put forward, such as adjusting transportation conditions, improving cargo packaging, or strengthening monitoring measures. The optimization suggestions are fed back to relevant decision makers to form an improvement plan to continuously improve the safety and reliability of cold chain transportation.

[0034] Step S6: Perform abnormal fault risk diagnosis based on the abnormal state starting traceability position, make intelligent fault decisions, and build a cargo intelligent state decision model.

[0035] In this embodiment, the data of the origin of the abnormal state is extracted from the previous step, including the time, location, state characteristics and related environmental factors (such as temperature, humidity, vibration, etc.) of the abnormality, to ensure the integrity and consistency of the data for subsequent analysis, to integrate the historical fault data and the corresponding environmental conditions, to form a comprehensive fault data set, to select a suitable risk assessment model, such as fault tree analysis (FTA), failure mode and effect analysis (FMEA) or a risk prediction model based on machine learning, which can systematically analyze potential faults and their impacts, to determine the input parameters of the model, including abnormal state characteristics, environmental factors and historical fault records, and to input the integrated data into the selected risk assessment model. In the model, risk diagnosis of abnormal failures is carried out. The model will identify potential failure types, probability of occurrence and degree of impact, record risk assessment results, including risk levels of different failure states, probability of occurrence and their impact on the transportation process, form a risk diagnosis report, and design the framework of the intelligent failure decision model, including data input, processing algorithm and decision output. The model should be able to process data in real time and generate decision suggestions for abnormal states, determine key decision indicators, such as failure probability, cargo importance, transportation time requirements, etc., to facilitate subsequent decision-making, and select appropriate decision algorithms, such as decision trees, random forests or deep learning models (such as neural networks), so as to make intelligent decisions based on risk assessment results. The algorithm should have processing The ability to process multi-dimensional data. For specific application scenarios, consider combining expert systems, integrating professional knowledge and data-driven methods to improve the accuracy of decision-making. Input risk assessment results and related status data into the intelligent decision-making model to generate decision-making suggestions for each abnormal state. Decisions include taking emergency measures, adjusting transportation routes, increasing monitoring frequency, etc. Record the decision-making process and results to ensure that each decision has sufficient basis and transparent records, form an intelligent decision-making report, integrate the results of risk diagnosis and intelligent decision-making into a unified intelligent state decision-making model, ensure that the model has real-time processing capabilities, adapt to the ever-changing transportation environment, set the model's feedback mechanism, and enable the model to continuously self-learn based on new data Learning and optimization improve the accuracy of decision-making, verify the constructed intelligent state decision-making model, use historical data for testing, evaluate the accuracy and effectiveness of the model, use cross-validation method to ensure that the model has good generalization ability, generate a model evaluation report, and record the performance indicators of the model, such as accuracy, recall rate, F1-score, etc., to provide a basis for subsequent optimization, apply the constructed intelligent state decision-making model to the actual transportation process, monitor the status of the goods in real time, provide intelligent decision-making support, ensure that the decision can quickly respond to abnormal conditions during transportation, reduce potential risks, collect new data and feedback generated in the application process, continuously optimize the model, form a closed-loop management, and improve overall transportation safety and efficiency.

[0036] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Step S11: Monitor the entire cold chain cargo transportation process based on sensors to obtain multi-dimensional real-time status parameters of the cold chain cargo; Step S12: performing abnormal data cleaning on multi-dimensional cold chain cargo real-time status parameters, thereby generating cleaning and optimization cargo real-time status parameters; Step S13: performing temperature distribution analysis on the real-time state parameters of the cleaning optimized cargo to generate cargo temperature distribution characteristic data; Step S14: Calculate the humidity fluctuation of the real-time state parameters of the cleaning optimization cargo to obtain the cargo humidity fluctuation parameters; Step S15: heterogeneous state fusion is performed on the cargo temperature distribution characteristic data and cargo humidity fluctuation parameters to obtain cargo heterogeneous state characteristics.

[0037] In this embodiment, in order to ensure the quality and safety of the goods during cold chain transportation, multiple sensors need to be installed on the transport vehicles and goods. These sensors include temperature sensors, humidity sensors, vibration sensors and position sensors. First, select sensors with high precision and reliability to ensure the accuracy and stability of the data. For example, the temperature sensor should have a fast response and a wide measurement range, and the humidity sensor should have a high humidity measurement accuracy. Next, the sensors need to be arranged reasonably according to the characteristics of the goods. The temperature sensor should be placed at the center and edge of the goods to monitor the temperature distribution of the entire goods; the humidity sensor should be placed in a well-ventilated location to obtain the real ambient humidity. The vibration sensor can be installed on the chassis of the vehicle to monitor the vibration that affects the goods during transportation. At the same time, the position sensor (such as GPS) is used to track the path and position of the transport vehicle in real time. The configuration of all sensors needs to ensure that their acquisition frequency is high enough, usually set to collect once a minute to obtain real-time status data. In addition, the communication protocol between the sensor and the data acquisition system needs to select a suitable type (such as LoRa, Zigbee, GPRS, etc.) to ensure that the data can be stably and timely transmitted to the cloud database for storage and analysis. The implementation of real-time data collection is the core of the cold chain monitoring system. After starting the data collection system, the system will automatically start monitoring and recording various status parameters of cold chain goods. Specifically, the system will collect data such as temperature, humidity, vibration and location in real time. The temperature and humidity data will be recorded together in the form of timestamps to ensure that each data point can be traced back to a specific time and state. This process ensures the reliability and integrity of the data. During the data recording process, the system needs to continuously monitor the operating status of the sensor to ensure its normal operation. If a sensor failure or data loss is found, the system should be able to issue an alarm in time and take corresponding fault handling measures. In addition, the real-time monitoring data should be transmitted to the cloud storage through a secure communication channel to ensure that the data will not be tampered with or lost during transmission. The data storage uses an efficient database system (such as MySQL, MongoDB, etc.) to support high-concurrency data writing and reading. All real-time status parameters will be stored in the database for subsequent analysis and processing. The format of the data should be unified to ensure that the required information can be quickly extracted and processed during subsequent analysis. After obtaining the real-time status parameters of the cold chain goods, the data needs to be cleaned for abnormal data. First of all, defining the abnormality standard is a key step in this process. Abnormal data includes temperature outside the set range (e.g. the temperature of cold chain goods should be kept between -18°C and 4°C), humidity fluctuation exceeding the predetermined threshold (e.g. relative humidity should be kept between 30% and 70%), or sensor failure (e.g. data does not change, the value is zero or an extreme value). In order to identify these abnormal data, statistical methods and machine learning algorithms are used. Common statistical methods include Z-score and interquartile range (IQR).The Z-score method calculates the Z-score of each data point to determine whether it exceeds the set threshold (such as |Z|>3); the IQR method calculates the quartiles of the data to identify values ​​that exceed 1.5 times the IQR. Machine learning methods, such as isolation forests or support vector machines (SVM), are also used to detect outliers, and these methods perform well when processing complex data sets. The implementation process of data cleaning requires the removal of identified abnormal data. First, traverse the real-time status parameter data and apply the selected anomaly detection algorithm to identify and mark abnormal data points. For each data point, the system will make a judgment based on the defined criteria and record the index and cause of the abnormal data. This process not only helps to improve the accuracy of the data, but also provides a reliable basis for subsequent analysis. For data identified as abnormal, different processing methods are adopted: one is to completely remove it, the second is to fill it with the average value of the surrounding normal data, and the third is to mark it and keep it in the data set for subsequent analysis. Finally, the optimized real-time status parameters of the cleaned goods should be stored in a new database table to ensure its accuracy and reliability. This process lays the foundation for the subsequent temperature distribution analysis and humidity fluctuation calculation.

[0038] Extract temperature data from the cleaned data set to ensure data accuracy and completeness. Select appropriate time windows (such as every hour) to group temperature data for subsequent analysis. Use statistical analysis methods (such as descriptive statistics, histogram analysis) to analyze temperature data and calculate the mean, variance, minimum and maximum values ​​of temperature. Plot temperature data into a histogram to visualize the distribution of temperature and facilitate identification of the range and frequency of temperature fluctuations. Generate temperature distribution feature data, including the distribution range of temperature, common temperature range, and the frequency of abnormal high and low temperatures. Record these feature data in the database for subsequent use and analysis. Extract humidity data from the cleaned data set to ensure data accuracy. Similar to temperature analysis, select appropriate time windows for grouping. Calculate humidity fluctuation parameters, including standard deviation, maximum fluctuation value and minimum fluctuation value. Use moving average method to smooth humidity data to eliminate short-term fluctuations and highlight long-term trends. Plot humidity fluctuation data into a line chart to show the trend of humidity over time and help identify abnormal fluctuations. Record humidity fluctuation parameters, including average humidity fluctuation, peak and valley values, and store them in the database. Determine the method of heterogeneous state fusion and select a suitable fusion algorithm, such as weighted average method, principal component analysis (PCA) or ensemble learning algorithm in machine learning (such as random forest). Treat temperature distribution characteristics and humidity fluctuation parameters as feature vectors for fusion analysis. Combine the characteristic data of temperature and humidity to generate comprehensive state characteristics through fusion algorithm. This step aims to find the relationship between temperature and humidity and identify their impact on the state of the goods. The generated feature data includes comprehensive state score, state category (such as normal, abnormal) and corresponding risk assessment value. Store the fused heterogeneous state characteristics in the database to ensure that they can be used for real-time monitoring and report generation. Generate an analysis report that outlines the comprehensive impact of temperature and humidity on the state of the goods to facilitate subsequent decision-making and implementation of optimization measures.

[0039] In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Step S21: mining the cargo vibration time-frequency of multi-dimensional cold chain cargo real-time status parameters to extract cargo vibration time-frequency features; Step S22: Calculating the vibration frequency of the cargo based on the time-frequency characteristics of cargo vibration to generate cargo vibration frequency parameters; Step S23: analyzing the vibration amplitude characteristics of the cargo vibration time-frequency characteristics to obtain the vibration amplitude characteristics; Step S24: performing time-series vibration fitting on the vibration frequency parameters and vibration amplitude characteristics of the goods to construct a time-series vibration curve for the goods transportation; Step S25: Performing transportation scene recognition on the cargo transportation time series vibration curve to obtain the cargo transportation scene.

[0040] In this embodiment, multi-dimensional cargo status parameters, including temperature, humidity and vibration data, are obtained from the real-time monitoring system. The vibration data should be collected by an acceleration sensor and usually presented in the form of three-axis acceleration. The sampling frequency of the data should be high enough to capture the details of the vibration signal. It is recommended to sample at least 50 times per second. The vibration data is analyzed in time and frequency using methods such as short-time Fourier transform (STFT) or wavelet transform. These methods convert time signals into time-frequency domains to facilitate the analysis of frequency components and time changes of vibrations. Appropriate window length and overlap rate are selected to balance time resolution and frequency resolution. Usually, the window length is set to 256. points, with an overlap rate of 50%. Apply the selected time-frequency analysis method to convert the vibration signal into a time-frequency diagram and extract the time-frequency features. The time-frequency diagram shows the distribution of the vibration signal in time and frequency. Identify and extract the main frequency components and their corresponding time periods. Record these features for subsequent analysis. Calculate the main vibration frequency from the frequency components extracted from the time-frequency features. Use the peak detection algorithm to identify the peak position in the spectrum to determine the main frequency of the vibration. When calculating the frequency, use the result of Fourier transform to extract the frequency value corresponding to the maximum amplitude. Generate the vibration frequency parameters of the goods, including the main frequency and its variation range. Record these frequency values ​​for subsequent analysis and comparison. Store the vibration frequency parameters in a database and associate them with other state parameters for comprehensive analysis. Calculate the amplitude of each frequency component through the time-frequency diagram, usually by obtaining the root mean square (RMS) value of the vibration signal. Or peak value to represent amplitude, record vibration amplitude characteristics, including the average, maximum and minimum amplitude, etc. These characteristics help identify the vibration of goods during transportation, store vibration amplitude characteristics in the database, and associate them with vibration frequency parameters to facilitate subsequent analysis and comparison, integrate vibration frequency parameters and vibration amplitude characteristics into time series data, ensure that the timestamps of the data are consistent, so as to perform time series analysis, create a comprehensive data set containing timestamps, frequencies and amplitudes, use linear regression, spline regression or other suitable time series analysis methods to fit the integrated data, and select a suitable model to effectively capture the changing trend of the data, for example, use a polynomial fitting model to fit the frequency and amplitude change curves to obtain a smooth time series vibration curve, generate a time series vibration curve and visualize it, and show the vibration changes of goods during transportation, using Matplotlib Or other visualization tools to draw curve graphs, store the vibration curve data and fitting results in the database for subsequent analysis and report generation, and use machine learning algorithms (such as K-means clustering, support vector machine or decision tree) to identify transportation scenarios based on the generated time-series vibration curves. It is necessary to first label the vibration data, build a training data set, and select feature vectors, including frequency, amplitude, vibration energy, etc., as input features.The labeled data set is divided into a training set and a test set. The training set is used to train the model and the parameters are adjusted to obtain the best results. After the model training is completed, the accuracy of the model is evaluated using the test set to ensure that it can effectively identify different transportation scenarios. The identified transportation scenario classification results are stored in the database, including the feature description of each scenario and its corresponding vibration curve. A transportation scenario recognition report is generated, which summarizes the recognition results and their analysis to facilitate subsequent decision support.

[0041] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Step S31: extracting real-time positioning data of goods based on multi-dimensional real-time status parameters of cold chain goods; Step S32: performing multi-time point position calculation on the real-time positioning data of the cargo, thereby obtaining the cargo positioning coordinates at multiple time points; Step S33: performing key stop node analysis on the cargo positioning coordinates at multiple time points, thereby generating key nodes for cargo transportation; Step S34: Dynamically track the cargo location coordinates at multiple time points to construct a time-series dynamic cargo trajectory; Step S35: Visualizing the real-time trajectory of the time-series dynamic cargo trajectory according to the key nodes of cargo transportation, thereby constructing a dynamic trajectory visualization map; Step S36: Perform real-time scene modeling on the dynamic trajectory visualization map according to the cargo transportation scene points to construct a dynamic scene transportation trajectory map.

[0042] In this embodiment, a GPS positioning module (such as NEO-6M or SIM808) is installed in the cold chain transport vehicle to obtain the location information of the goods in real time, ensuring that each cargo is equipped with a corresponding positioning device, and collecting other dimensional status parameters (such as temperature and humidity). These data will be integrated with the positioning data to form complete cargo status information. Wireless communication technology (such as GSM, LoRa or Wi-Fi) is used to transmit the positioning data to the central server in real time. The data format should be JSON for subsequent processing. The data collection frequency is set to ensure that the positioning data is updated every few seconds to provide continuous location information. The collected real-time positioning data is stored in a cloud database (such as Amazon DynamoDB or MongoDB). The data table should include fields such as timestamp, longitude, latitude, speed, etc. to ensure the integrity and consistency of the data for subsequent analysis and processing. The time window is set according to business needs, and the appropriate time period (for example, every 10 minutes or 30 minutes) is selected to extract the positioning data to form a multi-time point data set. For each time point, the corresponding GPS Coordinates and other status parameters. For the latitude and longitude data at each time point, ensure that it uses a unified coordinate system (such as WGS84) to facilitate subsequent analysis and visualization. Calculate the movement trajectory of the goods at different time points to analyze the transportation dynamics of the goods. Store the calculated cargo positioning coordinates at multiple time points in the database, record the location information, timestamp and related status parameters at each time point, ensure data structuring, and facilitate subsequent retrieval and analysis. Set the judgment criteria for the stay node. For example, if the goods stay at the same location for more than 5 minutes, it is considered a key stay node. By analyzing the positioning coordinates of multiple time points, identify the stop location and stay time of the goods, and perform cluster analysis on the cargo positioning data. Use DBSCAN Algorithms such as CNN and CNN are used to identify dense positioning points as potential stop nodes. The latitude and longitude, stay time and related state parameters (such as temperature, humidity, etc.) of the stop nodes are recorded. The identified key stop nodes are stored in the database to form a list of key nodes for subsequent analysis and visualization. The characteristic information of each key node, such as location, stay time, status, etc., is recorded. According to the cargo positioning coordinates at multiple time points, a smooth dynamic transportation trajectory is generated using linear interpolation or Bezier curve fitting algorithm. The trajectory should show the path of the cargo during transportation. The positioning coordinates of multiple time points are connected to form a continuous linear trajectory. The dynamic transportation trajectory is drawn on the map using visualization tools (such as Matplotlib, Plotly or Leaflet) to intuitively display the cargo transportation path. The key stop nodes are marked in the visualization to analyze the behavior of the cargo during transportation. The generated dynamic transportation trajectory and its visualization results are stored in the database. The latitude and longitude, timestamp and related status of the trajectory are recorded.Generate transportation trajectory reports and provide them to relevant personnel for analysis and decision-making. Use real-time data visualization tools (such as Grafana, D3.js or Google Maps) The dynamic trajectory visualization system is built based on the data processing and data flow processing (API) to ensure that the system can be updated in real time to reflect the latest cargo transportation status and trajectory changes. A data stream processing mechanism is established to ensure that the latest cargo positioning data and key node information are obtained from the database in real time. When new positioning data arrives, the dynamic trajectory on the visualization map is automatically updated. The dynamic trajectory of the cargo is displayed on the visualization map. The key stop nodes are represented by different marks (such as dots or icons). A rich information hierarchy is provided, and interactive functions are provided. Users click on the nodes to view detailed information (such as stop time, status parameters, etc.). Machine learning or deep learning algorithms (such as convolutional neural networks) are used to model the dynamic trajectory and identify different transportation scenarios (such as normal transportation, delays, cold chain failure, etc.). Training data sets are prepared, including trajectory features and corresponding labels in different scenarios. The scene recognition model is trained using the labeled data sets, and the parameters are adjusted to improve the accuracy of the model. After the training is completed, the performance of the model is verified using the test set to ensure that it can effectively identify different transportation scenarios. The results of the scene model are superimposed on the dynamic trajectory visualization map. Different colors or icons are used to represent different transportation scenarios. A scene modeling report is generated, which summarizes the recognition results and analysis of different scenarios and provides them to decision makers to optimize the transportation plan.

[0043] In this embodiment, step S4 includes the following steps: Step S41: Perform multi-time point feature recognition on the heterogeneous state characteristics of the goods, and extract the state characteristics of each time node; Step S42: Performing time-series state evolution on the state characteristics of each time node, thereby generating cargo state evolution characteristics; Step S43: matching the cargo positioning coordinates at multiple time points one by one according to the state characteristics of each time node, thereby obtaining the state positioning of each time node; Step S44: According to the state positioning of each time node, the cargo state evolution characteristics are dynamically rendered on the dynamic scene transportation trajectory map to build a dynamic cargo transportation twin model.

[0044] In this embodiment, the heterogeneous state feature data of the goods are collected from the previous steps, including temperature, humidity, vibration frequency and amplitude, positioning data, etc., and the data are ensured to have a timestamp for time series analysis. All relevant features at each time point are integrated to form a structured data set for subsequent processing. Feature selection methods in machine learning (such as principal component analysis (PCA) and recursive feature elimination (RFE)) are applied to identify the most representative state features at each time node. These features help understand the state of the goods at different time points. The importance score of each feature at different time points is calculated by an algorithm, and features with high importance are selected for subsequent analysis. The identified state features are recorded in a database, ensuring that the state features of each time node have a detailed description and timestamp, and generating a feature identification report to summarize the important state features and their changes at each time node, providing a basis for subsequent evolution analysis. Appropriate time series analysis methods (such as time series modeling, hidden Markov model (HMM) or long short-term memory network (LSTM)) are selected to analyze the evolution of state features. These methods can capture the law of feature changes over time, determine the time window and step size, and ensure that the model can effectively capture the evolution trend of state features. The state feature data extracted above is used to train the selected model, and the hyperparameters are adjusted to improve the prediction accuracy. Cross-validation is performed to ensure that the model performs well in different time periods and can accurately predict the changes in state features. The evolution features of the cargo state are generated, including the state category (such as normal, abnormal), state change rate, etc. of each time node. The evolution features are stored in the database for subsequent analysis and visualization. A state evolution report is formed. A suitable matching algorithm (such as the nearest neighbor algorithm, dynamic time warping (DTW) or weighted distance measurement) is selected to match the state features and positioning coordinates of each time node one by one. The matching criteria are determined, such as selecting the minimum distance between the state features and the positioning coordinates within a specific time window. For each time node, its state features are extracted and the corresponding coordinates are searched in the positioning data. The state positioning of successful matching is recorded. Many-to-one or one-to-many situations are handled to ensure that each state feature can find a suitable positioning coordinate. The state positioning information of each time node is stored in the database, including timestamp, longitude and latitude, state features, etc. A state location matching report is generated, which summarizes the results of the matching algorithm and its accuracy, and provides basic data for subsequent dynamic rendering. A suitable visualization tool (such as Three.js, D3.js or Mapbox) to build a dynamic rendering system to display the state evolution characteristics of the cargo transportation process in real time, ensure that the rendering system supports dynamic updates, can reflect the cargo status and trajectory changes in real time, and map the cargo status evolution characteristics to the dynamic scene transportation trajectory map according to the status positioning information of each time node. Different status characteristics are represented by different colors, icons or animation effects. The time series of state changes are displayed in the visualization. Users select a specific time point to view the status of the cargo during transportation, generate a dynamic cargo transportation twin model, and display the real-time state evolution of the cargo during transportation. This visualization can help relevant personnel quickly identify abnormal situations and make timely decisions. It provides interactive functions and allows users to select different time nodes through the time slider to dynamically view the changes in cargo status. .

[0045] In this embodiment, step S5 includes the following steps: Step S51: Perform full-cycle transportation monitoring analysis on the dynamic cargo transportation twin model and extract full-cycle transportation monitoring response data; Step S52: Perform cargo abnormal status detection on the full-cycle transportation monitoring response data to generate cargo transportation abnormal status data; Step S53: Tracing back the abnormal state time length of the cargo transportation abnormal state data to identify the initial time point of the abnormal state; Step S54: Based on the initial time point of the abnormal state, the dynamic cargo transportation twin model is traced to the abnormal location, thereby generating the starting traceability location of the abnormal state.

[0046] In this embodiment, the real-time and historical data already available in the dynamic cargo transportation twin model are used to integrate the status information of each time node, including temperature, humidity, vibration data and positioning information, to ensure that the data covers the entire transportation cycle, all time nodes from shipment to delivery, to form a complete data set, and select appropriate analysis tools (such as Python's Pandas, NumPy and SciPy libraries) to perform statistical analysis and visualization on the integrated data, set monitoring indicators, such as temperature variation range, humidity fluctuation, residence time, etc., as the basis for analysis, perform time series analysis on key status parameters in the transportation process, extract full-cycle monitoring response data, and the response data should include the change trend of status parameters, abnormal point identification, etc., generate a full-cycle transportation monitoring analysis report, describe the status changes in the transportation process in detail, pay attention to any parameters that deviate from the normal range, and select appropriate anomaly detection algorithms, such as statistical-based methods (such as Z-score or IQR) or machine learning methods (such as isolation forest, support vector machine) to identify abnormal conditions and determine the threshold of abnormalities, which is usually set according to the statistical characteristics of historical data. Input the monitoring response data into the selected anomaly detection model, identify the abnormal state points in the data, record the occurrence time of the abnormal state, the state parameters involved and their specific values, ensure the traceability of the data, generate a cargo transportation abnormal state data set, including all detected abnormal conditions and their detailed information, store the abnormal state data in the database, and generate an abnormal state detection report for subsequent analysis and decision-making. Select appropriate time series analysis tools (such as Pandas) The time series function of the abnormal state is used to analyze the abnormal state data, trace back the time length of the abnormal state, determine the time window for the traceback, which can usually be set as the time period before and after the abnormal state occurs. For each abnormal state, analyze its duration and record the time points when the abnormal state starts and ends. This process can be achieved by comparing the time periods of the abnormal state and the normal state. Generate an abnormal state time length traceback report, summarize the duration of each abnormality and its frequency of occurrence, record the initial time point of the abnormal state and its duration in the database, form an abnormal state time length traceback data set, generate a traceback analysis report, and provide basic data support for subsequent abnormal tracing. Select Appropriate traceability analysis methods usually use graph databases or time series databases (such as Neo4j, InfluxDB) to connect time nodes and location data and determine the logic of traceability, that is, trace back from the initial time point of the abnormal state, find the corresponding positioning data, start from the initial time point of the abnormal state, gradually trace back to the previous time point, extract the positioning coordinates and status data at that time, record the abnormal position and its corresponding status characteristics, and analyze the factors that affect the status of the goods before and after the abnormal state, such as transportation conditions, changes in the external environment, etc., and store the traced abnormal state starting position and its corresponding status information in the database.Form an abnormal state traceability data set, generate a traceability analysis report, provide detailed information on the starting position of the abnormal state, and related state parameters, and provide a basis for subsequent decision-making and improvement.

[0047] In this embodiment, the specific steps of step S6 are: Step S61: performing abnormal fault risk diagnosis on the cargo transportation abnormal state data to generate abnormal fault diagnosis results; Step S62: When the abnormal fault diagnosis result is that the cargo can be transported normally, a cargo abnormal fault signal is generated; Step S63: When the abnormal fault diagnosis result is that normal transportation is not possible, the cargo fault emergency transportation route is planned based on the abnormal state starting traceability position, thereby generating an emergency transportation planning route; Step S64: Make intelligent fault decisions on abnormal fault signals of cargo and emergency transportation planning paths, and build a cargo intelligent status decision model.

[0048] In this embodiment, the abnormal state data generated from the previous step is collected to ensure that the data contains specific parameters, timestamps, locations and other information of the abnormal state. The relevant historical transportation data, including cargo type, transportation conditions, environmental factors, etc., are integrated to provide a comprehensive background for fault risk diagnosis. A rule-based diagnosis system or machine learning model (such as random forest, decision tree, etc.) is used to diagnose abnormal fault risks. The selected model should be able to process multi-dimensional features and identify potential fault risks. The standards and thresholds for fault diagnosis are set. The characteristics of normal and abnormal states are analyzed based on historical data. The abnormal state data is input into the risk diagnosis model for real-time analysis to identify potential fault risks and generate abnormal fault diagnosis results. The existence of faults, fault types and their impact are recorded. The results are stored in a database, and a report is generated and provided to relevant decision makers to determine what the standard for "normal transportation" is. This usually depends on the output of the diagnosis model. If the model indicates that the cargo state is within a tolerable range, the cargo can be considered to be normal. Normal transportation of goods. When the diagnosis result is that the goods can be transported normally, a corresponding abnormal fault signal is generated, and the detailed information of the signal is recorded, including the signal generation time, state parameters, etc. The signal is stored in the monitoring system for subsequent tracking and analysis. The monitoring system is updated to show that the current state of the goods is a normal transportation state, and the existing abnormal fault signals are marked to ensure that all relevant personnel can obtain status updates in real time. Determine what the standard of "normal transportation" is, which usually depends on the output results of the diagnostic model. If the model indicates that the state of the goods is within the tolerable range, the goods can be considered to be transported normally. When the diagnosis result is that the goods can be transported normally, a corresponding abnormal fault signal is generated, and the detailed information of the signal is recorded, including the signal generation time, state parameters, etc. The signal is stored in the monitoring system for subsequent tracking and analysis. The monitoring system is updated to show that the current state of the goods is a normal transportation state, and the existing abnormal fault signals are marked to ensure that all relevant personnel can obtain status updates in real time. The shortest path algorithm (such as Dijkstra Algorithms or A* search algorithms) are used for emergency transportation route planning. These algorithms are suitable for handling path optimization problems in dynamic environments and can be adjusted according to real-time traffic conditions and road conditions. According to the starting traceability location and destination of the abnormal state, the best emergency transportation route is calculated, and factors such as traffic conditions, road conditions, and transportation time are considered to ensure the safety and efficiency of the route. The emergency transportation planning route is generated, and detailed information of the route (such as waypoints, estimated time, etc.) is recorded. The generated emergency transportation route information is stored in the database, and the monitoring system is updated to ensure that all relevant personnel can obtain the latest emergency transportation plan. An emergency transportation plan report is generated, outlining the basis and expected results of the route planning, and selecting a suitable intelligent decision-making model, such as a rule-based system, fuzzy logic control, or reinforcement learning model, is intended to generate reasonable decisions based on the current state and emergency path, and set the input parameters of the decision-making model.Including abnormal fault signals, emergency path information, cargo status, etc., according to the input data, run the intelligent decision-making model to generate suggestions and guidance on cargo transportation, such as whether to start emergency transportation immediately or take other measures. The model should consider the risks and benefits under various scenarios to ensure the best decision, record the decision results in the system, including recommended action steps, expected effects and related risk assessments, and generate intelligent status decision reports for decision makers to refer to, ensuring information transparency and timely feedback.

[0049] In this embodiment, the specific steps of step S63 are: When the abnormal fault diagnosis result is that normal transportation is impossible, the abnormal state data of cargo transportation is subjected to multi-period fault evolution to generate abnormal fault evolution data of multiple time periods; Perform abnormal fault prediction on abnormal state evolution data in multiple time periods to generate time series cargo fault prediction data; The current transportation route time length is calculated based on the abnormal state starting traceability position, thereby obtaining the current cargo transportation route time length; According to the current cargo transportation route duration, the time series cargo failure prediction data is used to plan the cargo failure emergency transportation route, thereby generating an emergency transportation planning route.

[0050] In this embodiment, the abnormal state data generated previously are collected to ensure that the data covers all time periods of the entire transportation cycle, including the state characteristics and their changes at each time point, and the relevant environmental factors and transportation parameters are integrated to form a comprehensive fault evolution data set. A time series analysis model (such as a hidden Markov model (HMM) or a long short-term memory network (LSTM)) is used to perform fault evolution analysis on the abnormal state data. These models can effectively capture patterns and trends in time series data and determine the time period for analysis. For example, the entire transportation cycle is divided into multiple time periods (such as every hour, each transportation stage, etc.) to provide a basis for subsequent analysis. The abnormal state data is input into the selected time series model to generate abnormal fault evolution data for each time period. These data should include changes in fault status, fault type, fault severity, etc. The evolution results of each time period are recorded to ensure that the data has a clear timestamp and state description, and the evolution data is stored. In the database, for subsequent use, select a suitable prediction model, such as time series prediction (ARIMA), machine learning model (such as random forest, XGBoost) or deep learning model (such as LSTM) for fault prediction, determine the input characteristics of the model, including abnormal state evolution data and related parameters for multiple time periods, use historical data to train the selected prediction model, adjust the parameters to improve the accuracy and reliability of the prediction, use cross-validation and other methods to evaluate the performance of the model to ensure its effectiveness in different time periods, generate time series cargo failure prediction data, including the type of failure, failure probability and impact of failure in the future time period, store the generated time series cargo failure prediction data in the database, record the prediction results and related information for each time period, generate a fault prediction report, summarize the prediction process and results, provide a basis for decision-making, use GIS (geographic information system) tools or path analysis algorithms (such as Dijkstra Algorithm) is used to calculate the time length of the current transportation path, and the tool is ensured to support real-time traffic data to improve the calculation accuracy. The starting traceability location and destination are determined as the basis for path calculation. Starting from the starting traceability location of the abnormal state, the selected tool is used to calculate the time length of the transportation path to the destination, taking into account the mode of transportation, traffic conditions and real-time road conditions, and the calculation results are recorded to ensure that the path time length has a detailed description, including the estimated arrival time and traffic conditions. The calculated time length of the current transportation path is recorded in the database, and a path length calculation report is generated to ensure that the information is transparent and traceable. The shortest path algorithm (such as A* algorithm or genetic algorithm) is used in combination with real-time traffic information to plan emergency transportation paths. These algorithms can quickly find the optimal path under dynamic conditions and determine the key parameters of emergency path planning, such as transportation time, road conditions, environmental factors, etc. According to the time length of the current cargo transportation path and fault prediction data, the best emergency transportation path is calculated. Path planning should consider factors such as the shortest time, lowest risk and lowest cost.Generate emergency transport planning routes and record detailed information on the routes, including transit points, estimated transport time, and risk areas. Store the generated emergency transport route information in the database, update the monitoring system to ensure that all relevant personnel can obtain emergency transport plans in real time, and generate emergency transport planning reports that outline the basis for route planning, expected results, and emergency response strategies to provide support for subsequent decision-making.

[0051] In this embodiment, a multimodal AIGC cold chain cargo status tracking and tracing system is provided, which is used to execute the multimodal AIGC cold chain cargo status tracking and tracing method as described above, including: The cargo heterogeneous state module is used to obtain the real-time state parameters of multi-dimensional cold chain cargo; perform abnormal data cleaning on the real-time state parameters of multi-dimensional cold chain cargo, and perform heterogeneous state fusion to obtain the cargo heterogeneous state characteristics; The transportation scenario module is used to mine the vibration time-frequency of multi-dimensional cold chain cargo real-time status parameters and identify the transportation scenario to obtain the cargo transportation scenario; The track map module is used to calculate the multi-time point positions according to the cargo transportation scenario, perform real-time scenario modeling, and build a dynamic scenario transportation track map; The twin model module is used to match the state positions of the dynamic scene transportation trajectory map one by one based on the heterogeneous state characteristics of the goods, and perform dynamic state rendering to build a dynamic cargo transportation twin model; The abnormal tracing module is used to detect abnormal cargo status in the dynamic cargo transportation twin model and trace the abnormal location, thereby generating the starting traceability location of the abnormal status; The intelligent status decision module is used to diagnose abnormal fault risks based on the abnormal status starting traceability location, make intelligent fault decisions, and build a cargo intelligent status decision model.

[0052] The present invention helps to fully understand the status of cold chain goods by obtaining multi-dimensional real-time status parameters of cold chain goods, provides rich data support, performs abnormal data cleaning and heterogeneous state fusion to improve data quality, obtains more accurate and reliable heterogeneous state characteristics of goods, and provides a reliable basis for subsequent analysis and decision-making. Cargo vibration time-frequency mining and transportation scene recognition help to understand the behavioral characteristics of goods in different transportation environments, and obtains cargo transportation scenes to help understand the situations faced during cargo transportation, providing important information for subsequent trajectory map modeling. Multi-point position calculation and instant scene modeling are performed according to the cargo transportation scene to achieve real-time monitoring and modeling of the cargo transportation process. Construction of dynamic scene transportation trajectory map helps to visualize the transportation path and state changes of goods, and provides data support for subsequent twin models. The status characteristics perform status position matching and dynamic status rendering on the dynamic scene transportation trajectory map to achieve real-time tracking and display of the cargo status. Building a dynamic cargo transportation twin model helps to compare the actual transportation situation with the model, improve the understanding and prediction of the cargo status, and perform abnormal status detection and abnormal position tracing on the dynamic cargo transportation twin model to quickly discover abnormal situations and determine the location where the abnormality occurred. Generating the starting traceability position of the abnormal state helps to track the root cause of the abnormal situation and provide a decision-making basis for subsequent intelligent state decisions. Abnormal fault risk diagnosis and intelligent fault decision-making based on the starting traceability position of the abnormal state improve the efficiency and accuracy of handling abnormal situations. Building a cargo intelligent state decision-making model helps to intelligently handle abnormal situations in cargo transportation and ensure the safety and stability of cargo transportation.

[0053] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0054] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A multi-modal AIGC cold chain cargo status tracking and tracing method, characterized in that: The following steps are involved: Step S1: obtaining multi-dimensional real-time status parameters of cold chain goods; performing abnormal data cleaning on the multi-dimensional real-time status parameters of cold chain goods, and performing heterogeneous state fusion to obtain heterogeneous state characteristics of goods; Step S2: mining the cargo vibration time-frequency of multi-dimensional cold chain cargo real-time status parameters and identifying the transportation scene to obtain the cargo transportation scene; Step S3: Perform multi-time point location calculation according to the cargo transportation scenario, perform real-time scenario modeling, and construct a dynamic scenario transportation trajectory map; Step S4: Based on the heterogeneous state characteristics of the goods, the dynamic scene transportation trajectory map is matched with the state position one by one, and dynamic state rendering is performed to build a dynamic cargo transportation twin model; Step S5: Detect abnormal status of cargo in the dynamic cargo transportation twin model and trace the abnormal location, thereby generating the starting traceability location of the abnormal status; Step S6: Perform abnormal fault risk diagnosis based on the abnormal state starting traceability position, make intelligent fault decisions, and build a cargo intelligent state decision model.

2. The multimodal AIGC cold chain cargo status tracking and tracing method according to claim 1 is characterized in that: The specific steps of step S1 are: Step S11: Monitor the entire cold chain cargo transportation process based on sensors to obtain multi-dimensional real-time status parameters of the cold chain cargo; Step S12: performing abnormal data cleaning on multi-dimensional cold chain cargo real-time status parameters, thereby generating cleaning and optimization cargo real-time status parameters; Step S13: performing temperature distribution analysis on the real-time state parameters of the cleaning optimized cargo to generate cargo temperature distribution characteristic data; Step S14: Calculate the humidity fluctuation of the real-time state parameters of the cleaning optimization cargo to obtain the cargo humidity fluctuation parameters; Step S15: heterogeneous state fusion is performed on the cargo temperature distribution characteristic data and cargo humidity fluctuation parameters to obtain cargo heterogeneous state characteristics.

3. The multimodal AIGC cold chain cargo status tracking and tracing method according to claim 1 is characterized in that: The specific steps of step S2 are: Step S21: mining the cargo vibration time-frequency of multi-dimensional cold chain cargo real-time status parameters to extract cargo vibration time-frequency features; Step S22: Calculating the vibration frequency of the cargo based on the time-frequency characteristics of cargo vibration to generate cargo vibration frequency parameters; Step S23: analyzing the vibration amplitude characteristics of the cargo vibration time-frequency characteristics to obtain the vibration amplitude characteristics; Step S24: performing time-series vibration fitting on the vibration frequency parameters and vibration amplitude characteristics of the goods to construct a time-series vibration curve for the goods transportation; Step S25: Performing transportation scene recognition on the cargo transportation time series vibration curve to obtain the cargo transportation scene.

4. The multimodal AIGC cold chain cargo status tracking and tracing method according to claim 1 is characterized in that: The specific steps of step S3 are: Step S31: extracting real-time positioning data of goods based on multi-dimensional real-time status parameters of cold chain goods; Step S32: performing multi-time point position calculation on the real-time positioning data of the cargo, thereby obtaining the cargo positioning coordinates at multiple time points; Step S33: performing key stop node analysis on the cargo positioning coordinates at multiple time points, thereby generating key nodes for cargo transportation; Step S34: Dynamically track the cargo location coordinates at multiple time points to construct a time-series dynamic cargo trajectory; Step S35: Visualizing the real-time trajectory of the time-series dynamic cargo trajectory according to the key nodes of cargo transportation, thereby constructing a dynamic trajectory visualization map; Step S36: Perform real-time scene modeling on the dynamic trajectory visualization map according to the cargo transportation scene points to construct a dynamic scene transportation trajectory map.

5. The multimodal AIGC cold chain cargo status tracking and tracing method according to claim 1 is characterized in that: The specific steps of step S4 are: Step S41: Perform multi-time point feature recognition on the heterogeneous state characteristics of the goods, and extract the state characteristics of each time node; Step S42: Performing time-series state evolution on the state characteristics of each time node, thereby generating cargo state evolution characteristics; Step S43: matching the cargo positioning coordinates at multiple time points one by one according to the state characteristics of each time node, thereby obtaining the state positioning of each time node; Step S44: According to the state positioning of each time node, the cargo state evolution characteristics are dynamically rendered on the dynamic scene transportation trajectory map to build a dynamic cargo transportation twin model.

6. The multimodal AIGC cold chain cargo status tracking and tracing method according to claim 1 is characterized in that: The specific steps of step S5 are: Step S51: Perform full-cycle transportation monitoring analysis on the dynamic cargo transportation twin model and extract full-cycle transportation monitoring response data; Step S52: Perform cargo abnormal status detection on the full-cycle transportation monitoring response data to generate cargo transportation abnormal status data; Step S53: Tracing back the abnormal state time length of the cargo transportation abnormal state data to identify the initial time point of the abnormal state; Step S54: Based on the initial time point of the abnormal state, the dynamic cargo transportation twin model is traced to the abnormal location, thereby generating the starting traceability location of the abnormal state.

7. The multimodal AIGC cold chain cargo status tracking and tracing method according to claim 1 is characterized in that: The specific steps of step S6 are: Step S61: performing abnormal fault risk diagnosis on the cargo transportation abnormal state data to generate abnormal fault diagnosis results; Step S62: When the abnormal fault diagnosis result is that the cargo can be transported normally, a cargo abnormal fault signal is generated; Step S63: When the abnormal fault diagnosis result is that normal transportation is not possible, the cargo fault emergency transportation route is planned based on the abnormal state starting traceability position, thereby generating an emergency transportation planning route; Step S64: Make intelligent fault decisions on abnormal fault signals of cargo and emergency transportation planning paths, and build a cargo intelligent status decision model.

8. The multimodal AIGC cold chain cargo status tracking and tracing method according to claim 7 is characterized in that: The specific steps of step S63 are: When the abnormal fault diagnosis result is that normal transportation is impossible, the abnormal state data of cargo transportation is subjected to multi-period fault evolution to generate abnormal fault evolution data of multiple time periods; Perform abnormal fault prediction on abnormal state evolution data in multiple time periods to generate time series cargo fault prediction data; The current transportation route time length is calculated based on the abnormal state starting traceability position, thereby obtaining the current cargo transportation route time length; According to the current cargo transportation route duration, the time series cargo failure prediction data is used to plan the cargo failure emergency transportation route, thereby generating an emergency transportation planning route.

9. A multi-modal AIGC cold chain cargo status tracking and tracing system, characterized in that: Used to execute the multimodal AIGC cold chain cargo status tracking and tracing method as claimed in claim 1, comprising: The cargo heterogeneous state module is used to obtain the real-time state parameters of multi-dimensional cold chain cargo; perform abnormal data cleaning on the real-time state parameters of multi-dimensional cold chain cargo, and perform heterogeneous state fusion to obtain the cargo heterogeneous state characteristics; The transportation scenario module is used to mine the vibration time-frequency of multi-dimensional cold chain cargo real-time status parameters and identify the transportation scenario to obtain the cargo transportation scenario; The track map module is used to calculate the multi-time point positions according to the cargo transportation scenario, perform real-time scenario modeling, and build a dynamic scenario transportation track map; The twin model module is used to match the state positions of the dynamic scene transportation trajectory map one by one based on the heterogeneous state characteristics of the goods, and perform dynamic state rendering to build a dynamic cargo transportation twin model; The abnormal tracing module is used to detect abnormal cargo status in the dynamic cargo transportation twin model and trace the abnormal location, thereby generating the starting traceability location of the abnormal status; The intelligent status decision module is used to diagnose abnormal fault risks based on the abnormal status starting traceability location, make intelligent fault decisions, and build a cargo intelligent status decision model.

Images