Multi-airport luggage full-process tracking method and system

Abstract

The present invention relates to a method and system for full-process tracking of luggage at multiple airports. The method steps are as follows: data collection, where tags collect temperature, humidity, and acceleration data, high-definition cameras obtain luggage tag images, synchronize airline flight dynamic information, collect real-time airport operation data, passenger feedback, and personalized requirements; data conversion, encoding according to specific rules into sequence tokens suitable for large model processing; multi-head attention mechanism, focusing on key parameters such as luggage timeliness and equipment load; intelligent tag and blockchain collaboration, dynamically optimizing sensor parameters, recording key luggage status information and providing personalized feedback; automatic identification and path planning, combining computer vision and multi-head attention mechanism to identify tags in real time and optimize the transfer path; virtual luggage companion service, providing intelligent interaction and immersive experience; cross-airport collaborative management, real-time monitoring of global airport dynamics, establishing a risk warning model, and coordinating resources to execute transfer adjustment plans.

Description

Technical Field

[0001] The present invention belongs to the technical field of air transportation and logistics management, and specifically relates to a method and system for full-process tracking of luggage at multiple airports. Background Art

[0002] In the process of modern air transportation, the efficiency and accuracy of luggage management are of great significance for improving passenger satisfaction and ensuring transportation safety. Existing technologies mainly focus on luggage identification and tracking, but there are still certain limitations in multi-airport collaborative management, data security, and intelligent decision-making.

[0003] Chinese invention patent CN116309692B discloses a method, device, and medium for binding people and luggage at airport security checks based on deep learning. This method obtains real-time surveillance videos when users place their luggage, extracts effective images, and uses a deep learning model to extract user identity features and luggage item features to achieve the binding and real-time tracking of luggage and users. Although this method improves the security during the security check process, its main limitation is that it is limited to luggage binding and tracking at a single airport, lacks the ability of cross-airport collaborative management, and the efficiency and accuracy in processing a large amount of multi-source data still need to be improved.

[0004] Chinese invention patent CN115228753B discloses a method and device for visual tracking and display of luggage. This method performs luggage position detection and display through a combination of barcode reading and optoelectronic induction, and uses optoelectronic sensors to monitor the luggage status in real time and perform real-time tracking on the display screen. Although this method realizes the visual tracking of luggage positions, its main limitation is that it depends on the distribution of physical sensors, has low scalability and data security, and it is difficult to achieve unified management and intelligent optimization of luggage among multiple airports. Summary of the Invention

[0005] In view of the above deficiencies of the prior art, the present invention proposes a method and system for full-process tracking of luggage at multiple airports, aiming to achieve efficient, intelligent, and secure full-process tracking and management of luggage among multiple airports through the integration of advanced intelligent tags, blockchain technology, and multi-head attention mechanisms, and improve the overall efficiency of luggage management and the user experience.

[0006] To solve the above problems, the present invention provides a method and system for full-process tracking of luggage at multiple airports, including the following steps:

[0007] S1: Data collection, including:

[0008] Collecting sensor data from intelligent tags, collecting luggage tag images; collecting flight dynamic information, real-time operation data; collecting luggage query, feedback, and personalized demand information submitted by terminals;

[0009] S2: Data Transformation. Transform the multi-source data collected into sequence tokens that can be processed by large models, including: converting the baggage label text recognized by computer vision into text tokens of a fixed dimension through natural language processing techniques; converting the flight departure and arrival times into timestamp sequence tokens, encoding the equipment status in the real-time airport operation data in binary, and encoding the baggage flow according to the number of baggages passing through key monitoring points; classifying and encoding the passenger interaction information according to the interaction type, and encoding the query information into sequence tokens containing keywords and time.

[0010] S3: Use the multi-head attention mechanism to focus on key baggage parameters, including:

[0011] Use the timeliness attention head to focus on the timeliness requirements of the baggage, use the equipment load attention head to pay attention to the airport equipment load situation, use the resource attention head to pay attention to the resource reserve and allocation capabilities of different airports, and use the weather attention head to monitor the impact of weather changes on airports in different regions.

[0012] S4: Intelligent label and blockchain collaborative processing, including:

[0013] According to the analysis of the transformed sequence tokens by the large model, dynamically optimize the operating parameters of the intelligent label sensors; use blockchain technology to record the key status information of the baggage to ensure the immutability and transparency of the data; feedback the baggage status information to the passengers in a personalized and easy-to-understand manner through the blockchain.

[0014] S5: Automatic identification and dynamic path planning optimization, including:

[0015] Real-time identify baggage labels through the computer vision system to improve the identification accuracy; use the multi-head attention mechanism combined with real-time data to generate and dynamically optimize the transfer path of the baggage to ensure efficient transportation.

[0016] S6: Virtual baggage companion service, including:

[0017] Provide intelligent interaction services based on passenger needs to respond to passenger inquiries and feedback; provide immersive experiences.

[0018] S7: Cross-airport collaborative management execution, including:

[0019] Real-time monitor the weather, equipment and flight dynamics of airports around the world, and establish a risk warning model; when the warning is triggered, coordinate the resources of multiple airports, formulate and execute the baggage transfer adjustment plan to ensure the stability and efficiency of the baggage transportation network.

[0020] In the S2 data conversion step, the data encoding of the intelligent tag includes encoding temperature data to one decimal place in degrees Celsius, encoding humidity data in percentage form, and encoding acceleration data according to the three-dimensional coordinate axis direction components respectively.

[0021] The S4 intelligent tag and blockchain collaborative processing step includes: dynamically adjusting the acquisition frequency and accuracy of the intelligent tag sensor according to the transportation stage of the luggage; recording the key status information of the luggage through blockchain technology, and customizing the information feedback strategy according to the passenger type.

[0022] The S5 automatic recognition and dynamic path planning optimization step includes: using the multi-head attention mechanism to improve the recognition accuracy of luggage tags in different lighting, angle, and stacking scenarios; dynamically generating and optimizing the transfer path of the luggage based on real-time flight dynamics and airport operation data.

[0023] The multi-airport luggage full-process tracking system includes: a data acquisition module for executing the S1 step;

[0024] A data conversion module for executing the S2 step in claim 1;

[0025] A large model processing module with a built-in multi-head attention mechanism for executing the S3 step;

[0026] An intelligent tag and blockchain collaborative module for executing the S4 step;

[0027] An automatic recognition and dynamic path planning module for executing the S5 step;

[0028] A virtual luggage companion service module for executing the S6 step;

[0029] A cross-airport collaborative management module for executing the S7 step.

[0030] The data conversion module includes: an encoding rule unit for encoding multi-source data; an embedding combination unit for semantically fusing the encoded data to generate a sequence token suitable for large model processing.

[0031] The large model processing module includes: a multi-head attention mechanism for respectively focusing on the key parameters of luggage timeliness, equipment load, resource reserve, and weather impact; a model training unit that uses federated learning and incremental learning methods to continuously optimize the model performance to adapt to the dynamically changing airport operation environment.

[0032] The intelligent tag and blockchain collaborative module includes: an environment prediction unit for predicting the possible environmental changes that the luggage may encounter based on machine learning and deep learning algorithms; an information feedback unit for pushing the luggage status information through multiple channels according to the passenger portrait.

[0033] The automatic identification and dynamic path planning module includes: a label recognition model that combines multi-head attention and deep neural network technologies to improve the accuracy of luggage label recognition; a path planning unit that generates an optimal transfer path based on real-time data and dynamically adjusts the luggage transfer order to ensure the priority handling of high-priority luggage.

[0034] The cross-airport collaborative management module includes: a risk warning unit that monitors the weather, equipment, and flight dynamics of airports around the world in real time, and establishes and updates a risk warning model; an emergency linkage unit that coordinates the resources of multiple airports and formulates and executes a luggage transfer adjustment plan when a warning is triggered; an execution supervision unit that continuously monitors the execution of each airport during the emergency handling process, and allocates resources in a timely manner to ensure the efficiency and smoothness of the emergency handling process.

[0035] In summary, the present application has the following beneficial effects:

[0036] 1. By installing intelligent tags on luggage and deploying high-definition cameras at key transfer nodes, accurate identification and real-time tracking of luggage in each transfer link are achieved. The temperature, humidity, and acceleration sensors built into the intelligent tags can monitor the environmental status of the luggage in real time, while the high-definition cameras ensure clear capture of the luggage label images, thereby significantly reducing the risk of luggage loss or mistransportation. The multi-head attention mechanism is used to focus on and analyze the key parameters of the luggage's time limit requirements, equipment load, resource reserve and allocation capabilities, and weather changes, improving the accuracy of luggage tracking and the comprehensiveness of decision-making, and ensuring the efficient management of luggage throughout the transportation process.

[0037] 2. Through the automatic identification and dynamic path planning module, combined with real-time data and the multi-head attention mechanism, it is possible to dynamically generate and optimize the luggage transfer path. This not only ensures the priority handling of high-priority luggage but also effectively avoids equipment congestion and resource waste, thus significantly improving the overall efficiency of luggage transportation. The cross-airport collaborative management module monitors the operation dynamics of airports around the world in real time, establishes a risk warning model, and coordinates the resources of multiple airports and formulates and executes a luggage transfer adjustment plan when a warning is triggered. This collaborative management mechanism can respond quickly in case of anomalies, ensuring the stable and efficient operation of the luggage transportation network and further improving the transportation efficiency and system resilience.

[0038] 3. The large model processing module is built with a multi-head attention mechanism, which can deeply analyze multi-source data, focus on the key parameters of luggage, and achieve intelligent decision-making and automated control. This greatly reduces manual intervention, improves the automation level and operation efficiency of the system. The virtual luggage companion service module provides personalized services based on passenger needs through intelligent interaction and immersive experiences. This not only enhances the user experience but also strengthens the emotional connection between passengers and luggage through intelligent means, further promoting the intelligent development of the system.

[0039] 4. Smart label and blockchain collaboration module. By using blockchain technology, it records the key status information of the luggage, ensuring the immutability and transparency of the data. This not only enhances the data security of the system but also improves the user's trust in the luggage management system. The environmental prediction unit, based on machine learning and deep learning algorithms, can predict in advance the environmental changes that the luggage may encounter and dynamically optimize the parameters of the smart label sensors. This prediction and optimization mechanism ensure the safety and stability of the luggage in different environments, further enhancing the overall security and reliability of the system.

[0040] 5. Cross-airport collaborative management module. Through the risk warning unit, it monitors the weather, equipment, and flight dynamics of airports around the world in real time, establishes and updates the risk warning model. When a warning is triggered, the system can quickly coordinate the resources of multiple airports, formulate and execute the luggage transfer adjustment plan to ensure the stable and efficient operation of the luggage transportation network under abnormal circumstances. The execution supervision unit, by continuously monitoring the implementation of each airport during the emergency handling process, allocates resources in a timely manner to ensure the efficiency and smoothness of the emergency handling process. This supervision and allocation mechanism effectively improves the response speed and handling ability of the system in the face of emergencies, enhancing the overall stability and resilience of the system.

[0041] 6. Information feedback unit. It customizes the information feedback strategy according to different types of passengers (such as business passengers, passengers carrying special items), and pushes the luggage status information in a personalized and easy-to-understand way through the mobile application or airport display screen, improving the service experience and satisfaction of passengers. The virtual luggage companion service module, through augmented reality (AR) or virtual reality (VR) technology, provides an immersive luggage transportation scenario for passengers, enhancing the emotional connection between passengers and luggage and further improving the overall experience of users.

[0042] 7. Data conversion module. Through the encoding rule unit and the embedding combination unit, it can efficiently encode and semantically fuse multi-source data to generate sequence tokens suitable for large model processing. This efficient data processing mechanism ensures that the system can quickly and accurately process massive amounts of data, improving the overall data processing efficiency. The system architecture adopts a modular design, and each functional module works independently and collaboratively, facilitating the expansion, maintenance, and upgrade of the system. Such a design not only improves the overall reliability and flexibility of the system but also provides convenience for future function expansion and technology upgrade, ensuring the long-term sustainable development of the system.

[0043] 8. The large model processing module and the cross-airport collaborative management module use big data analysis technology to establish a risk warning model, which can insight potential operation crises in advance, provide strong decision-making support for operation management, improve the scientificity and predictability of overall operation management. The multi-head attention mechanism can intelligently analyze and focus on the key parameters of luggage, generate optimization decisions, and achieve intelligent control during the luggage transfer process. This intelligent decision-making and optimization control mechanism not only improves operation efficiency but also ensures the high efficiency and safety of the luggage transportation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 is the workflow diagram of this application;

[0045] Figure 2 is the system architecture diagram of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] In order to make the objectives, technical solutions, and beneficial effects of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the drawings for the convenience of those skilled in the art to understand.

[0047] Embodiment 1:

[0048] As Figures 1 to 2 shown, this embodiment details a method for full-process tracking of luggage at multiple airports, covering key steps such as data collection, data conversion, application of the multi-head attention mechanism, collaborative processing of intelligent tags and blockchain, automatic identification, and dynamic path planning optimization. The specific implementation methods of each step will be described step by step below.

[0049] Step S1: Data collection

[0050] Intelligent tag data collection:

[0051] Install intelligent tags on each piece of luggage. The intelligent tags are built-in with temperature sensors, humidity sensors, and acceleration sensors. The sensors continuously collect the temperature and humidity of the environment where the luggage is located and the movement state of the luggage (such as acceleration changes). The sensor data is regularly uploaded to the central data server through low-power wireless communication technology (such as Bluetooth Low Energy BLE).

[0052] Luggage tag image collection:

[0053] Install high-definition cameras at key nodes of luggage transfer at each airport (such as security checkpoints, sorting centers, boarding gates). The cameras continuously capture the luggage tag images passing by and transmit the image data to the computer vision system for real-time recognition processing.

[0054] Flight dynamic information collection:

[0055] The system interfaces with the airline flight information system through a secure and efficient interface protocol (such as API). It obtains dynamic data such as flight departure and arrival times, delays, and aircraft type information in real time and stores them in the central database.

[0056] Airport real-time operation data collection:

[0057] By interfacing with the airport operation management system, real-time operation data including equipment status (such as the operation status of conveyor belts and sorting equipment), personnel configuration (number of on-duty staff), and baggage flow (number of bags passing through key monitoring points per minute) is obtained. The data is updated in real time to the central data server through a secure encryption transmission method to ensure the real-time and accuracy of the data.

[0058] Passenger interaction information collection:

[0059] Passengers can submit baggage query, feedback, and personalized demand information through mobile applications, self-service terminals, or manned counters. The system classifies and encodes the collected interaction information and stores it in the central database for subsequent processing.

[0060] Step S2: Data conversion

[0061] Baggage label text conversion:

[0062] The baggage label text recognized by the computer vision system first undergoes text cleaning by natural language processing (NLP) technology to remove noise characters and correct spelling mistakes. The cleaned text is converted into text tokens of a fixed dimension through a word vector model for easy processing by the large model.

[0063] Flight departure and arrival time conversion:

[0064] The flight departure and arrival times are converted into a timestamp sequence token in the international standard time format.

[0065] The delay situation is quantitatively encoded in minutes, and the aircraft type information is encoded according to the International Air Transport Association (IATA) standard code.

[0066] Airport real-time operation data conversion:

[0067] The equipment status is represented by binary encoding for running (1) or faulty (0) status. The personnel configuration data is encoded by the actual number of on-duty staff. The baggage flow data is encoded according to the number of bags passing through the key monitoring points per minute.

[0068] Passenger interaction information conversion:

[0069] Classify and code passenger interaction information according to the interaction type. The query information includes keywords and time, which are converted into sequence tokens. The feedback information includes the feedback content and the feedback emotion tendency, which are converted into corresponding sequence tokens. The personalized demand information is coded according to the demand category and priority.

[0070] Step S3: Use the multi-head attention mechanism to focus on the key parameters of the luggage

[0071] Timeliness attention:

[0072] Use the timeliness attention head to focus on the timeliness requirements of the luggage, and combine the flight connection time and the transit stay duration to accurately calculate the optimal transfer speed and route selection of the luggage. For example, at the initial stage of the luggage entering the airport transfer process, the timeliness attention head obtains the flight connection time and the transit stay duration corresponding to the luggage, and combines the internal airport map data and the preset transfer speed models in different regions to calculate the set of routes that can theoretically reach the next link fastest.

[0073] Equipment load attention:

[0074] Use the equipment load attention head to monitor the real-time load conditions of each conveyor belt and sorting equipment in the airport, and mark the congested sections and idle sections. Every predetermined time period (such as 5 minutes), the equipment load attention head dynamically switches the luggage transfer route according to the change of equipment load, and preferentially guides the luggage to flow to the idle sections.

[0075] Resource attention:

[0076] Use the resource attention head to pay attention to the resource reserve and allocation capabilities of different airports, and reasonably arrange the transfer destinations of the luggage. According to the idle conditions of manpower and equipment at each airport, select the most suitable transfer equipment for the luggage.

[0077] Weather attention:

[0078] Use the weather attention head to monitor the weather changes at airports in different regions, anticipate in advance the flight cancellations or delays that may be caused by bad weather, and adjust the luggage transportation strategy. For example, during the daily monitoring stage, when it is detected that the bad weather indicators at an airport in a certain region exceed the set threshold, trigger a potential operation crisis warning and formulate a luggage transfer adjustment plan in advance.

[0079] Step S4: Collaborative processing of intelligent tags and blockchain

[0080] Dynamically optimize the parameters of intelligent tag sensors:

[0081] Analyze the transformed sequence tokens based on the large model and dynamically adjust the operating parameters of the intelligent tag sensors. For example, according to the transportation stage of the luggage, combined with the analysis of historical luggage transportation data of the same type of flights and the current weather and airport busyness, predict the possible environmental changes that the luggage may encounter. When it is predicted that the luggage will pass through a high-temperature area, instruct the temperature sensor of the intelligent tag to increase the acquisition frequency to 10 times per minute, and adjust the accuracy to two decimal places to accurately capture potential risks.

[0082] The blockchain technology records the key status information of the luggage:

[0083] Use blockchain technology to record the key status information of the luggage to ensure the immutability and transparency of the data. Every time the luggage status changes (such as temperature and humidity changes, arrival at transfer nodes, etc.), it will be recorded on the blockchain to form an immutable luggage status log.

[0084] Personalized and easy-to-understand feedback on luggage status information:

[0085] S5: Automatic identification and dynamic path planning optimization, including:

[0086] Real-time identify luggage tags through a computer vision system to improve the identification accuracy; use the multi-head attention mechanism combined with real-time data to generate and dynamically optimize the transfer path of the luggage to ensure efficient transportation;

[0087] In a multi-airport scenario, in order to achieve efficient transportation of luggage, it is necessary to dynamically select the transfer path based on real-time data. The key factors affecting the selection of the transfer path usually include:

[0088] Time Sensitivity: The departure and arrival times of the flights corresponding to the luggage, the transfer time of the passengers, etc.;

[0089] Equipment Load: The real-time busyness of each conveyor belt and sorting equipment;

[0090] Resource Availability: The available personnel and facility reserves at each airport;

[0091] Weather Impact: Delays or reduced equipment availability that may be caused by bad weather.

[0092] To fully consider the above factors, a corresponding dynamic cost function can be defined on a graph model or a network model, and combined with the multi-head attention mechanism to assign variable weights to each path or edge, so as to optimize the luggage transfer decision at different times.

[0093] Combination of multi-head attention mechanism and path planning

[0094] Weights of the multi-head attention mechanism

[0095] The multi-head attention mechanism consists of K "attention heads", each corresponding to a different focus, such as:

[0096] The 1st attention head: Head time Timeliness attention

[0097] The 2nd attention head: Head load Device load attention

[0098] The 3rd attention head: Head resource Resource reserve and allocation attention

[0099] The 4th attention head: Head weather Weather attention

[0100] By analyzing the current state of the luggage and the real-time operation data of the airport, the large model will output an attention weight vector that changes over time as follows:

[0101]

[0102] Transshipment network model

[0103] Abstract the airport and its internal transfer nodes as G=(V,E), where:

[0104] V is the set of nodes (such as security checkpoints, sorting centers, boarding gates, etc.), |V| = N;

[0105] E is the set of directed edges (representing conveyor belts or passable transfer routes), |E| = M;

[0106] (u,v)∈E means there is a feasible transfer channel from node u to node v.

[0107] Dynamic cost function

[0108] For each edge (u,v)∈E, define a time-varying "comprehensive cost" C(u,v,t) to measure the cost or time-consuming of the luggage passing through this edge. It can be decomposed into 4 sub-cost functions, corresponding to 4 attention heads:

[0109] Timeliness cost f time (u,v,t) reflects the estimated time required for the luggage to pass through the edge (u,v), or the impact on the transfer connection.

[0110] Device load cost f load (u,v,t) reflects the busyness of the conveyor belt and sorting equipment where this edge is located. The higher the load, the lower the efficiency of the luggage passing through.

[0111] Resource reserve cost f resource (u, v, t) reflects the availability of human resources and equipment allocation at nodes or airports. The cost is higher when there is a shortage of human resources and equipment.

[0112] Weather impact cost f weather (u, v, t) When the weather is bad, the availability of this edge or node will decrease, the transportation speed will slow down or the success rate will decrease, and the corresponding cost will increase.

[0113] Using the weight α(t) output by the multi-head attention mechanism, the "comprehensive" cost function can be obtained as follows:

[0114]

[0115] Dynamic path planning objective

[0116] Given the starting node s where the luggage is located and the target node (or airport) d, find a path P at time t to minimize the cumulative cost. If the path is represented by a sequence of edges (e1, e2....e L ), then the objective function is:

[0117]

[0118] Dynamic update: Every preset time interval (such as 5 minutes), or whenever key operation parameters change significantly (such as sudden equipment failure or weather warning), the large model will recalculate the new attention weight α(t), update the cost function of each edge, and then replan the optimal transfer path of the luggage.

[0119] Implementation process of dynamic path planning

[0120] Luggage detection and recognition

[0121] Real-time identify the luggage label through the computer vision system and obtain the current location s of the luggage.

[0122] Multi-head attention mechanism to calculate weights

[0123] The large model analyzes the latest sequence of tokens (including information such as timeliness, load, resources, weather, etc.) and outputs α(t).

[0124] Update edge cost

[0125] Recalculate C(u, v, t) for each feasible edge (u, v) to reflect the latest operating conditions and attention focus.

[0126] Optimal path search

[0127] On G=(V,E), solve the optimal path P ∗ (t) through the shortest path algorithm (such as Dijkstra, A*, or a custom multi-objective algorithm), satisfying the following formula:

[0128]

[0129] If the system adopts parallel multi-objective optimization, an algorithm based on Pareto optimal solutions can also be used for searching.

[0130] Execution and monitoring

[0131] Send the optimal path to the baggage transfer control system to achieve automatic regulation of conveyor belts, sorting equipment, etc.;

[0132] Continuously monitor the operation data and the positions of the baggage. Whenever there are significant changes in the environment, repeat steps 2 - 4 to dynamically update the baggage transfer path.

[0133] Adaptability of the multi-head attention mechanism

[0134] As time progresses and the environment changes (such as flight delays, equipment loads, weather mutations), the attention weight α(t) can be adaptively adjusted to make the cost function C(u,v,t) reflect the priority requirements of baggage transfer in real time.

[0135] Real-time performance and robustness. Through dynamic path planning, the route is automatically re-planned at fixed intervals or when key operation parameters mutate, ensuring that the baggage can still reach the destination smoothly and efficiently in the complex and changeable airport environment.

[0136] Comprehensive multi-factor decision-making. Considering multiple factors such as timeliness, equipment load, resources, and weather simultaneously can significantly improve the transfer efficiency and reduce baggage misplacement or delays.

[0137] Scalability. If more attention heads (such as cost attention head, carbon emission attention head, etc.) need to be incorporated in the future, only the corresponding sub-cost functions and attention weights need to be added to the comprehensive cost function to maintain compatibility with the existing framework.

[0138] Customize the information feedback strategy according to the passenger type, and feedback the key status information of the baggage on the blockchain to the passengers in a personalized and easy-to-understand way through mobile applications or airport display screens. For example, for business passengers, highlight the baggage connection timeliness information, such as the countdown of the remaining transit time; for passengers carrying special items, focus on displaying the environmental safety information of the items, such as whether the temperature and humidity are suitable for the preservation of special drugs.

[0139] At each transfer node, the baggage collects environmental data and image data through smart tags and cameras, and simultaneously synchronizes the flight dynamic information and the airport real-time operation data.

[0140] The query, feedback, and personalized demand information submitted by passengers through the terminal are collected and classified and coded in real time by the system.

[0141] The multi-source data collected is converted into sequence tokens through specific coding rules to ensure a unified data format for easy processing by the large model.

[0142] Various types of data (such as sensor data, image text, flight information, operation data, passenger interaction information) are coded separately to generate structured sequence tokens.

[0143] The multi-head attention mechanism built into the large model respectively focuses on the timeliness requirements of the luggage, equipment load, resource reserve and allocation capabilities, and the impact of weather changes on luggage transfer.

[0144] Each attention head calculates the optimal transfer speed and route of the luggage based on real-time data and historical data analysis to optimize the luggage transfer path.

[0145] The large model analyzes the converted sequence tokens and dynamically adjusts the operating parameters of the intelligent label sensors to adapt to the environmental changes that may occur during the luggage transportation process.

[0146] The key status information of the luggage is recorded through blockchain technology to ensure the immutability and transparency of the data.

[0147] The system personalizes and feeds back the luggage status information to passengers through multiple channels according to the passenger type to improve the user experience.

[0148] The computer vision system identifies the luggage label in real time to improve the recognition accuracy and ensure the precise tracking of the luggage at each transfer node.

[0149] The large model combines real-time data and uses the multi-head attention mechanism to generate and dynamically optimize the luggage transfer path to ensure that the luggage reaches the destination efficiently and safely.

[0150] Provide intelligent interaction services based on passenger needs, respond to passenger queries and feedback, provide an immersive experience, and enhance the emotional connection between passengers and luggage.

[0151] Real-time monitor the weather, equipment, and flight dynamics at airports around the world, establish a risk warning model. When the warning is triggered, coordinate the resources of multiple airports, formulate and execute a luggage transfer adjustment plan to ensure the stability and efficiency of the luggage transportation network.

[0152] Through the combination of intelligent tags and high-definition cameras, achieve precise identification and real-time tracking of luggage at each transfer node, reducing the risk of luggage loss or mistransportation.

[0153] Utilize the multi - head attention mechanism to focus on key parameters, combine real - time data to dynamically optimize the transfer path of luggage, ensure that the luggage reaches the destination efficiently and safely, and shorten the luggage transportation time.

[0154] Through the application of large models and multi - head attention mechanisms, realize intelligent decision - making and automated control during the luggage transfer process, reduce manual intervention, and improve operational efficiency.

[0155] Adopt blockchain technology to record the key status information of luggage, ensure the immutability and transparency of data, and enhance the trust and data security of the system.

[0156] Customize the information feedback strategy according to different passenger types, provide personalized and easy - to - understand luggage status information, and enhance the service experience and satisfaction of passengers.

[0157] Through cross - airport collaborative management and risk warning models, anticipate and respond to potential operational crises in advance, ensure the stable and efficient operation of the luggage transportation network under abnormal conditions, and reduce luggage delay losses.

[0158] Through the method described in this embodiment, it is possible to achieve full - process intelligent tracking and efficient transfer of luggage between multiple airports, improve the accuracy and efficiency of luggage management, enhance the service experience of passengers, and have significant practical value and broad application prospects.

[0159] Embodiment 2:

[0160] As Figures 1 to 2 shown, this embodiment details a multi - airport luggage full - process tracking system, covering key modules such as data acquisition, data conversion, large model processing, intelligent tag and blockchain collaborative processing, automatic identification and dynamic path planning optimization, virtual luggage companion service, and cross - airport collaborative management execution. The specific implementation methods and functions of each module will be introduced step by step below.

[0161] The data acquisition module is responsible for obtaining various types of luggage - related data from various data sources, including intelligent tag sensor data, luggage tag images, flight dynamic information, airport real - time operation data, and passenger interaction information.

[0162] Install intelligent tags on each piece of luggage. The intelligent tags are built - in with temperature sensors, humidity sensors, and acceleration sensors. The sensors regularly upload data to the central data server through low - power wireless communication technology (such as Bluetooth Low Energy BLE).

[0163] Install high - definition cameras at key transfer nodes (such as security checkpoints, sorting centers, boarding gates) in each airport to continuously capture the luggage tag images passing by. The cameras transmit the image data to the computer vision system for recognition processing in real - time through the network.

[0164] The system interfaces with the airline flight information system through a secure and efficient interface protocol (such as API) to obtain dynamic data such as flight departure and arrival times, delay situations, aircraft type information, etc. The obtained data is stored in the central database for subsequent processing.

[0165] Through the data interface with the airport operation and management system, real-time operation data including equipment status (such as the operation status of conveyor belts and sorting equipment), personnel configuration (number of on-duty personnel), and baggage flow (number of bags passing through key monitoring points per minute) is obtained. The data is updated to the central data server in real-time through encrypted transmission to ensure the timeliness and accuracy of the data.

[0166] Passengers submit baggage query, feedback, and personalized demand information through mobile applications, self-service terminals, or manned counters. The interactive information collected by the system is classified and encoded and stored in the central database for subsequent processing.

[0167] The data conversion module is responsible for encoding and embedding multi-source heterogeneous data to generate sequence tokens suitable for large model processing, ensuring unified data format and semantic fusion.

[0168] Coding rule unit:

[0169] Set different coding rules according to the sensor type: temperature data is encoded in degrees Celsius, accurate to one decimal place. Humidity data is encoded in percentage form. Acceleration data is encoded according to the three-dimensional coordinate axis direction components respectively.

[0170] The baggage label text recognized by computer vision is cleaned by natural language processing technology, removing noise characters and correcting spelling mistakes, and then converted into text tokens of a fixed dimension according to the word vector model. Flight departure and arrival times are converted into timestamp sequence tokens according to the international standard time format, the delay situation is quantified and encoded in minutes, and the aircraft type information is encoded according to the International Air Transport Association (IATA) standard code.

[0171] The equipment status in the real-time airport operation data is represented by binary encoding for the running (1) or faulty (0) state, the personnel configuration data is encoded by the actual number of on-duty personnel, and the baggage flow data is encoded according to the number of bags passing through the key monitoring points per minute. Passenger interaction information is classified and encoded according to the interaction type. Query information includes keywords and time and is converted into sequence tokens; feedback information includes feedback content and feedback emotional tendency and is converted into corresponding sequence tokens; personalized demand information is encoded according to the demand category and priority.

[0172] Embedding combination unit:

[0173] Use an efficient embedding algorithm to semantically fuse the encoded multi-source data and generate sequence tokens suitable for large model processing. Ensure that different types of data can be effectively combined semantically, reduce data redundancy, and improve data processing efficiency.

[0174] The large model processing module is built with a multi-head attention mechanism, which is responsible for deeply analyzing the transformed sequence tokens, focusing on the key parameters of the luggage, and continuously optimizing the model performance through the model training unit.

[0175] Multi-head attention mechanism: The model contains multiple attention heads, and each head focuses on different key parameters of the luggage: The timeliness attention head focuses on the timeliness requirements of the luggage. The equipment load attention head pays attention to the load situation of airport equipment. The resource attention head takes note of the resource reserves and allocation capabilities of different airports. The weather attention head monitors the impact of weather changes on airports in different regions. Through parallel computing, each attention head comprehensively processes information from multiple aspects, improving the accuracy and comprehensiveness of decision-making.

[0176] Model training unit:

[0177] Adopt federated learning and incremental learning methods, and use distributed computing resources for model training. Continuously optimize the model parameters, enhance the generalization ability of the model, and ensure its ability to adapt to the dynamically changing airport operation environment. Use historical data and real-time data for training to ensure the model's response ability to the latest situation.

[0178] The intelligent label and blockchain collaboration module is responsible for dynamically optimizing the sensor parameters of the intelligent label, and using blockchain technology to record the key status information of the luggage, ensuring the immutability and transparency of the data, and at the same time providing personalized feedback of the luggage status information to the passengers.

[0179] Environmental prediction unit:

[0180] Based on machine learning and deep learning algorithms, analyze the sequence tokens processed by the large model to predict the possible environmental changes that the luggage may encounter. For example, when predicting that the luggage will enter a high-temperature area, adjust the temperature sensor parameters of the intelligent label in advance.

[0181] Information feedback unit:

[0182] According to the passenger profile, push the luggage status information to the passengers in a personalized and easy-to-understand manner through multiple channels. Customize different information feedback strategies for different types of passengers (such as business passengers, passengers carrying special items) to improve the user experience.

[0183] Blockchain record:

[0184] Use blockchain technology to record the key status information of luggage, ensuring the immutability and transparency of data. Every change in luggage status (such as temperature and humidity changes, arrival at transfer nodes, etc.) will be recorded on the blockchain, forming an immutable luggage status log.

[0185] The automatic identification and dynamic path planning module uses a computer vision system to identify luggage tags in real time, and combines real-time data with the multi-head attention mechanism to generate and dynamically optimize the transfer path of luggage, ensuring efficient luggage transportation.

[0186] Tag recognition model:

[0187] Combining the multi-head attention mechanism with deep neural network technology to improve the recognition accuracy of luggage tags under different lighting, angles, and stacking scenarios. Adopt a regular incremental learning training method, and continuously optimize the recognition model using newly collected image data to ensure that the recognition accuracy is stable above 98%.

[0188] Path planning unit:

[0189] Based on real-time flight dynamics and airport operation data, use the multi-head attention mechanism to generate the optimal transfer path. Dynamically adjust the luggage transfer order to ensure the priority handling of high-priority luggage. Update the path planning results to the airport luggage transfer control system in real time, and automatically regulate the running direction and speed of the conveyor belt and sorting equipment to achieve automated transfer.

[0190] The virtual luggage companion service module provides intelligent interaction services and immersive experiences based on passenger needs, enhancing the emotional connection between passengers and their luggage and improving the user experience.

[0191] Intelligent interaction service:

[0192] Integrate natural language processing (NLP) and conversational AI technology to provide intelligent interaction services and respond to passengers' inquiries and feedback. Passengers interact with the virtual luggage companion through the mobile application to obtain services such as luggage status information and transfer suggestions.

[0193] Immersive experience:

[0194] Use augmented reality (AR) or virtual reality (VR) technology to generate exclusive luggage transportation scenarios in real time based on passengers' interest preferences. As the luggage status changes, dynamically update the details of the virtual scenario to enhance the emotional connection and experience of passengers.

[0195] The cross-airport collaborative management module is responsible for real-time monitoring of weather, equipment, and flight dynamics at airports around the world, establishing a risk warning model, and coordinating multi-airport resources when a warning is triggered, formulating and implementing a luggage transfer adjustment plan to ensure the stability and efficiency of the luggage transportation network.

[0196] Risk warning unit:

[0197] Utilize big data analysis technology to monitor the weather, equipment, and flight dynamics of airports worldwide in real-time.

[0198] Establish and continuously update a risk warning model to anticipate potential operational crises in advance. When the bad weather indicators of airports in a certain region exceed the set thresholds (such as heavy rain precipitation, wind speed, etc.), or there are equipment failures or large-scale flight delays, trigger a potential operational crisis warning.

[0199] Emergency response coordination unit:

[0200] When the warning is triggered, coordinate the resources of multiple airports, formulate and execute a plan for adjusting luggage transfer. Notify the standby airports around the affected airports to initiate the reception preparation, and re-plan the multi-level transfer routes of the affected luggage. According to the resource reserves of each airport, reasonably allocate the luggage transfer tasks to ensure the stability and efficiency of luggage transportation.

[0201] Execution supervision unit:

[0202] Continuously monitor the implementation of each airport during the emergency handling process to ensure the effective execution of the luggage transfer adjustment plan. If it is found that the implementation progress of a certain airport lags behind, promptly allocate additional resources for support, such as dispatching more personnel and allocating emergency equipment, to ensure the efficiency and smoothness of the entire emergency handling process.

[0203] The data acquisition module collects multi-source data through smart tags, cameras, flight information systems, airport operation management systems, and passenger terminals, and transmits the data to the data conversion module.

[0204] The encoding rule unit in the data conversion module performs specific encoding on the collected data, and the embedding combination unit performs semantic fusion on the encoded data to generate sequence tokens suitable for large model processing, and transmits them to the large model processing module.

[0205] The large model processing module is built with a multi-head attention mechanism to analyze the sequence tokens, focus on the key parameters of the luggage, generate optimized decisions, and transmit them to the smart tag and blockchain collaboration module and the automatic identification and dynamic path planning module.

[0206] The smart tag and blockchain collaboration module dynamically optimizes the parameters of smart tag sensors, uses blockchain technology to record luggage status information, and personalized feedbacks the luggage status information to passengers through the information feedback unit.

[0207] The automatic identification and dynamic path planning module real-time identifies luggage tags through a computer vision system, and combines with the multi-head attention mechanism to generate and dynamically optimize the luggage transfer path to ensure efficient luggage transportation.

[0208] The virtual luggage companion service module provides intelligent interaction services and immersive experiences according to passengers' needs, enhancing the emotional connection between passengers and their luggage.

[0209] The cross-airport collaborative management module monitors the operation dynamics of airports around the world in real time, establishes a risk warning model, coordinates the resources of multiple airports when a warning is triggered, formulates and executes a luggage transfer adjustment plan to ensure the stability and efficiency of the luggage transportation network.

[0210] At each transfer node, the luggage collects environmental data and image data through intelligent tags and cameras, and synchronizes flight dynamic information and airport operation data in real time.

[0211] The system encodes multi-source data through the encoding rule unit, embeds it into the combination unit for semantic fusion, and generates sequence tokens.

[0212] The large model processing module analyzes the sequence tokens using the multi-head attention mechanism, focuses on the key parameters of the luggage, and generates optimized decisions.

[0213] The intelligent tag and blockchain collaboration module dynamically adjusts the sensor parameters according to the analysis results of the large model, records the luggage status information, and feeds it back to the passengers.

[0214] The automatic identification and dynamic path planning module optimizes the transfer path by real-time identifying luggage tags and combining the multi-head attention mechanism to ensure efficient transportation.

[0215] The virtual luggage companion service module provides intelligent interaction and immersive experiences, enhancing the user experience.

[0216] The cross-airport collaborative management module coordinates the resources of multiple airports through real-time monitoring and risk warning to ensure the stability and efficiency of the luggage transportation network.

[0217] Through modular design, each functional module works independently and collaboratively, facilitating the expansion, maintenance, and upgrade of the system, and improving the overall reliability and flexibility of the system.

[0218] The data conversion module realizes the unified processing of multi-source data through specific encoding and embedding combination, ensures data format consistency and semantic fusion, and improves data processing efficiency.

[0219] The large model processing module is built with a multi-head attention mechanism, which can efficiently focus on the key parameters of the luggage, support intelligent decision-making, and optimize the luggage transfer path.

[0220] Using blockchain technology to record the key status information of the luggage, ensuring the immutability and transparency of the data, and enhancing the security and trust of the system.

[0221] The application of large models and multi-head attention mechanisms enables intelligent decision-making and automated control during the luggage transfer process, reducing manual intervention and improving operational efficiency.

[0222] The automatic identification and dynamic path planning module ensures the efficient and safe arrival of luggage at the destination by optimizing the luggage transfer path in real time.

[0223] The virtual luggage companion service module provides personalized services through intelligent interaction and immersive experiences, enhancing the emotional connection between passengers and their luggage and improving user satisfaction.

[0224] The system customizes information feedback strategies according to different passenger types, providing personalized and easy-to-understand luggage status information and improving service quality.

[0225] The cross-airport collaborative management module anticipates potential operational crises in advance through real-time monitoring and risk early warning, responds quickly, and coordinates multi-airport resources to ensure the stable and efficient operation of the luggage transportation network in case of anomalies, reducing the risks of luggage delays and losses.

[0226] By combining intelligent tags and high-definition cameras, precise identification and real-time tracking of luggage at each transfer node are achieved, reducing the risks of luggage loss or mistransportation.

[0227] The multi-head attention mechanism combines with real-time data to optimize the transfer path, ensuring the efficient transportation of luggage, shortening the luggage transportation time, and improving overall operational efficiency.

[0228] Through the multi-airport luggage full-process tracking system described in this embodiment, full-process intelligent tracking and efficient transfer of luggage worldwide can be realized, improving the accuracy and efficiency of luggage management, enhancing the service experience of passengers, and having significant practical value and broad application prospects.

[0229] Embodiment 3:

[0230] As Figures 1 to 2 shown, this embodiment details the advanced functions and optimization mechanisms of a multi-airport luggage full-process tracking system, covering the environmental prediction unit and information feedback unit in the intelligent tag and blockchain collaboration module, the tag recognition model and path planning unit in the automatic identification and dynamic path planning module, and the risk early warning unit, emergency linkage unit, and execution supervision unit in the cross-airport collaborative management module. The specific implementation methods and functions of each module will be introduced step by step below.

[0231] The intelligent tag and blockchain collaboration module is responsible for predicting possible environmental changes that luggage may encounter through the environmental prediction unit, and pushing luggage status information through multiple channels based on the passenger profile through the information feedback unit, enhancing data security and transparency while improving the user experience.

[0232] Environmental Prediction Unit:

[0233] Based on machine learning (ML) and deep learning (DL) algorithms, the sequence tokens processed by the large model are analyzed to predict the environmental changes (such as temperature, humidity, vibration, etc.) that the luggage may encounter during transportation.

[0234] By using historical transportation data and real-time environmental data to establish a predictive model, potential environmental risks can be identified in advance.

[0235] For example, when it is predicted that the luggage is about to enter a high-temperature area, the system automatically adjusts the temperature sensor parameters of the smart tag to increase the frequency and accuracy of data collection to ensure more accurate status monitoring of the luggage in a high-temperature environment.

[0236] Information feedback unit:

[0237] Baggage status information is pushed through multiple channels (such as mobile applications, airport displays, emails, etc.) based on the passenger's personal profile (such as passenger type, luggage nature, preference settings, etc.).

[0238] Customize feedback strategies for different types of travelers:

[0239] Business travelers: Highlight luggage timeliness information, such as the countdown to remaining transfer time, to ensure that luggage arrives on time.

[0240] Passengers carrying special items: The environmental safety information of the luggage is displayed with emphasis, such as whether the temperature and humidity are suitable for the storage of special medicines, to ensure the safety of special items.

[0241] Information feedback is presented in an easy-to-understand format with pictures and texts, which enhances passengers' perception and trust in the status of their luggage.

[0242] The automatic identification and dynamic path planning module improves the recognition accuracy of baggage tags by combining the multi-head attention mechanism and deep neural network technology, generates the optimal transfer path based on real-time data, dynamically adjusts the baggage transfer order, and ensures priority processing of high-priority baggage.

[0243] Tag recognition model:

[0244] Combining multi-head attention mechanism with deep neural network (DNN) technology, the recognition accuracy of luggage tags in different lighting, angles and stacking scenarios is improved.

[0245] Model training uses a large amount of diverse luggage tag image data, including complex scenes such as different lighting conditions, shooting at different angles, partial occlusion, etc., to ensure the robustness and accuracy of the model.

[0246] Adopt a regular incremental learning training method, and continuously optimize the recognition model using newly collected image data to ensure that the recognition accuracy rate is stably above 98%.

[0247] Path planning unit:

[0248] Based on real-time flight dynamics and airport operation data, use the multi-head attention mechanism to generate the optimal transfer path.

[0249] Dynamically adjust the luggage transfer order to ensure the priority handling of high-priority luggage (such as business class luggage, special item luggage).

[0250] Update the path planning results to the airport luggage transfer control system in real time, and automatically regulate the running direction and speed of the conveyor belt and sorting equipment to achieve automated transfer.

[0251] The path planning takes into account various factors, such as equipment load, luggage time limit requirements, resource reserve conditions, etc., to ensure the optimization and efficiency of the transfer path.

[0252] The cross-airport collaborative management module is responsible for real-time monitoring of the weather, equipment, and flight dynamics of airports around the world, establishing a risk warning model, and coordinating multi-airport resources when a warning is triggered, formulating and implementing a luggage transfer adjustment plan to ensure the stability and efficiency of the luggage transportation network.

[0253] Risk warning unit:

[0254] Use big data analysis technology to real-time monitor the weather, equipment, and flight dynamics of airports around the world.

[0255] Establish and continuously update a risk warning model to be able to anticipate potential operation crises in advance.

[0256] The monitoring indicators include severe weather (such as rainfall during heavy rain, wind speed during strong wind), equipment failures (such as conveyor belt downtime), flight delays (such as the proportion of large-scale flight delays), etc.

[0257] When the monitoring indicators of airports in a certain region exceed the set threshold, automatically trigger a potential operation crisis warning.

[0258] Emergency linkage unit:

[0259] When a warning is triggered, coordinate multi-airport resources, formulate and implement a luggage transfer adjustment plan.

[0260] Notify the standby airports around the affected airports to start receiving preparations to ensure that the luggage can be transferred in a timely manner.

[0261] Re-plan the multi-level transfer routes of the affected luggage, and reasonably allocate the luggage transfer tasks according to the resource reserve conditions of each airport to ensure the stability and efficiency of the luggage transportation.

[0262] Execution Supervision Unit:

[0263] Continuously monitor the implementation status of each airport during the emergency handling process to ensure the effective implementation of the baggage transfer adjustment plan.

[0264] If it is found that the implementation progress of a certain airport lags behind, promptly allocate additional resources for support, such as dispatching more personnel and allocating emergency equipment, to ensure the efficiency and smoothness of the entire emergency handling process.

[0265] Through real-time data monitoring and feedback mechanisms, ensure the transparency and traceability of the emergency handling process, and enhance the overall response ability and reliability of the system.

[0266] Working Process

[0267] The intelligent tag and blockchain collaboration module analyzes the possible environmental changes during the baggage transportation process through the environmental prediction unit, dynamically optimizes the intelligent tag sensor parameters, uses blockchain technology to record the baggage status information, and personalized feedbacks the baggage status information to passengers through the information feedback unit.

[0268] The automatic identification and dynamic path planning module real-time identifies the baggage tags through the tag identification model, generates and dynamically optimizes the baggage transfer path in combination with the multi-head attention mechanism to ensure efficient transportation.

[0269] The cross-airport collaborative management module real-time monitors the operation dynamics of airports around the world through the risk warning unit, establishes a risk warning model, coordinates the resources of multiple airports when the warning is triggered, formulates and implements the baggage transfer adjustment plan to ensure the stability and efficiency of the baggage transportation network.

[0270] At the transfer nodes of baggage at airports around the world, environmental data and image data are collected through intelligent tags and high-definition cameras. The system encodes and fuses the multi-source data through the data conversion module to generate sequence tokens.

[0271] The large model processing module analyzes the sequence tokens using the multi-head attention mechanism, focuses on the key parameters of the baggage, and generates optimization decisions.

[0272] The intelligent tag and blockchain collaboration module dynamically adjusts the sensor parameters according to the analysis results of the large model, records the baggage status information, and personalized pushes the information to passengers through the information feedback unit.

[0273] The automatic identification and dynamic path planning module optimizes the transfer path by real-time identifying the baggage tags and combining the multi-head attention mechanism to ensure the efficient transportation of the baggage.

[0274] The cross-airport collaborative management module monitors the operation dynamics of airports around the world in real time through the risk warning unit, establishes a risk warning model, coordinates multi-airport resources when a warning is triggered, formulates and executes a luggage transfer adjustment plan to ensure the stability and efficiency of the luggage transportation network.

[0275] Through the environmental prediction unit, the system can predict in advance the environmental changes that luggage may encounter during transportation, dynamically adjust the parameters of intelligent label sensors to ensure the accuracy and real-time nature of luggage status monitoring.

[0276] The information feedback unit pushes personalized luggage status information through multiple channels (such as mobile applications, display screens) according to the passenger profile, enhancing the service experience and satisfaction of passengers.

[0277] The label recognition model combining the multi-head attention mechanism and deep neural network technology significantly improves the recognition accuracy of luggage labels in complex environments, ensuring the precise tracking of luggage at each transfer node.

[0278] The path planning unit generates the optimal transfer path based on real-time data and the multi-head attention mechanism, dynamically adjusts the luggage transfer sequence to ensure the priority handling of high-priority luggage, significantly improving the luggage transportation efficiency.

[0279] Monitor the operation dynamics of airports around the world, establish and update the risk warning model, and gain insight into potential operation crises in advance. When a warning is triggered, the emergency linkage unit can quickly coordinate multi-airport resources, formulate and execute a luggage transfer adjustment plan to ensure the stable and efficient operation of the luggage transportation network under abnormal circumstances. The execution supervision unit continuously monitors the implementation of each airport during the emergency handling process, allocates resources in a timely manner to ensure the efficiency and smoothness of the emergency handling process, and reduces the risk of luggage delay and loss.

[0280] Use blockchain technology to record the key status information of luggage to ensure the immutability and transparency of data, enhancing the trust and data security of the system.

[0281] Provide personalized luggage status information and emotional connection through intelligent interaction services and immersive experiences, enhancing the service experience and satisfaction of passengers.

[0282] Through the advanced functions and optimization mechanisms of the multi-airport luggage full-process tracking system described in this embodiment, intelligent prediction, personalized information feedback, precise identification and efficient transfer of luggage on a global scale, as well as efficient risk warning and emergency response can be achieved, significantly enhancing the intelligence, automation and security of luggage management, with significant practical value and broad application prospects.

[0283] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made in form and details without departing from the scope defined by the claims of the present invention.

Claims

1. A method for tracking baggage throughout the entire process at multiple airports, characterized by: Includes the following step: S1: Data collection, including: Collect sensor data from smart tags and baggage tag images; collect flight dynamic information, covering real-time operation data of multiple airports; collect baggage query, feedback and personalized demand information submitted by terminals; S2: Data conversion, converting the multi-source data including the sensor data, baggage tag images, flight dynamic information, real-time operation data and passenger interaction information collected in step S1 into sequence tokens that can be processed by the subsequent large model processing module according to specific coding rules, including: converting the baggage tag text identified by computer vision into a fixed-dimensional text token through natural language processing technology; converting the flight take-off and landing time into a timestamp sequence token, encoding the equipment status in the airport operation real-time data in binary, and encoding the baggage flow according to the number of baggage passing through key monitoring points; classifying and encoding passenger interaction information according to the interaction type, and encoding the query information into a sequence token containing keywords and time; S3: The large model processing module processes the sequence tokens converted in step S2 using a multi-head attention mechanism to focus on key parameters of the luggage. The multi-head attention mechanism includes: Use the timeliness focus to focus on the timeliness requirements of baggage, use the equipment load focus to pay attention to the airport equipment load, use the resource focus to pay attention to the resource reserves and deployment capabilities of different airports, and use the weather focus to monitor the impact of weather changes on airports in different regions; S4: Smart label and blockchain collaborative processing, including: According to the analysis result of the sequence token converted by the large model processing module in step S2, the operating parameters of the smart tag sensor are dynamically optimized; the key status information of the luggage is recorded by using blockchain technology to ensure the immutability and transparency of the data; the luggage status information is fed back to the passenger in a personalized and easy-to-understand manner through the blockchain; S5: Automatic identification and dynamic path planning optimization, including: The computer vision system is used to identify baggage tags in real time to improve recognition accuracy. The multi-head attention mechanism is combined with the real-time data collected in the S1 step to generate and dynamically optimize the baggage transfer path to ensure efficient transportation. S6: Virtual luggage accompaniment service, including: Provide intelligent interactive services based on passenger needs, respond to passenger inquiries and feedback, and provide an immersive experience; S7: Collaborative management execution across multiple airports, including: Real-time monitoring of weather, equipment and flight dynamics at airports around the world to establish a risk warning model; when the risk warning model is triggered, coordinate the resources of multiple airports, formulate and implement baggage transfer adjustment plans to ensure the stability and efficiency of the baggage transportation network.

2. A multi-airport baggage full-process tracking method according to claim 1, characterized in that: In the S2 data conversion step, the data encoding of the smart tag includes encoding the temperature data in Celsius with one decimal place, encoding the humidity data in percentage form, and encoding the acceleration data according to the directional components of the three-dimensional coordinate axes.

3. The method for tracking baggage throughout the entire process at multiple airports according to claim 1, characterized in that: The S4 smart tag and blockchain collaborative processing steps include: dynamically adjusting the collection frequency and accuracy of the smart tag sensor according to the transportation stage of the luggage; recording the key status information of the luggage through blockchain technology, and customizing the information feedback strategy according to the passenger type.

4. The method for tracking baggage throughout the entire process at multiple airports according to claim 1, characterized in that: The S5 automatic identification and dynamic path planning optimization step also includes: using the multi-head attention mechanism to improve the recognition accuracy of baggage tags under different lighting, angles and stacking scenarios; based on real-time flight dynamics and airport operation data, dynamically generating and optimizing the baggage transfer path.

5. A multi-airport baggage full-process tracking system for executing the method of claim 1, characterized in that: include: A data acquisition module, used to execute the S1 step; A data conversion module, used to execute the step S2; The multi-source data collected by the data collection module, including sensor data, baggage tag images, flight dynamic information, real-time operation data, and passenger interaction information, are encoded and converted into sequence tokens according to specific encoding rules; A large model processing module, with a built-in multi-head attention mechanism, for executing the S3 step; the multi-head attention mechanism processes the sequence token output by the data conversion module, and includes a timeliness attention head, a device load attention head, a resource attention head, and a weather attention head; A smart label and blockchain collaboration module, used to execute the S4 step; including logic for dynamically optimizing the smart label operating parameters according to the analysis results of the large model processing module; An automatic identification and dynamic path planning module, used to execute the S5 step; configured to generate and optimize the transfer path using the multi-head attention mechanism combined with the real-time data of flight dynamic information and real-time operation data collected by the data collection module; A virtual baggage companion service module, used to execute the step S6; The cross-airport collaborative management module is used to execute the S7 step, and is configured to establish a risk warning model and coordinate the resources of multiple airports to implement adjustment plans when the warning is triggered.

6. A multi-airport baggage full-process tracking system according to claim 5, characterized in that: The data conversion module includes: an encoding rule unit, which is used to execute the specific encoding rule to encode multi-source data; an embedding combination unit, which is used to semantically fuse the encoded data to generate a sequence token suitable for processing by the large model processing module.

7. The multi-airport baggage full-process tracking system according to claim 5, characterized in that: The large model processing module also includes: a model training unit, which uses federated learning and incremental learning methods to continuously optimize model performance to adapt to the dynamically changing airport operating environment.

8. The multi-airport baggage full-process tracking system according to claim 5, characterized in that: The smart tag and blockchain collaborative module includes: an environmental prediction unit, which predicts environmental changes that luggage may encounter based on machine learning and deep learning algorithms; an information feedback unit, which pushes luggage status information through multiple channels based on passenger portraits.

9. The multi-airport baggage full-process tracking system according to claim 5, characterized in that: The automatic identification and dynamic path planning module also includes: a tag recognition model, which combines multi-head attention and deep neural network technology to improve the recognition accuracy of luggage tags under different lighting, angles and stacking scenarios, and the tag recognition model is trained using a large amount of diverse luggage tag image data; a path planning unit, which is configured to generate an optimal transfer path based on real-time data and dynamically adjust the luggage transfer order to ensure priority processing of high-priority luggage.

10. The multi-airport baggage full-process tracking system according to claim 5, characterized in that: The cross-airport collaborative management module includes: a risk warning unit, which is used to monitor the weather, equipment and flight dynamics of airports around the world in real time, and establish and update the risk warning model; an emergency linkage unit, which is used to coordinate multi-airport resources when the warning is triggered, formulate and implement the baggage transfer adjustment plan; an execution supervision unit, which is used to continuously monitor the execution status of each airport during the emergency handling process, and Deploy resources in a timely manner to ensure the efficiency and smoothness of the emergency response process.

Images