Intelligent AI-driven digital twin low-carbon dispatching system for airport luggage flow group

Abstract

The invention relates to the technical field of intelligent airport logistics, in particular to an airport luggage flow group intelligent AI-driven digital twinning low-carbon scheduling system, which comprises a physical sensing layer, a digital twinning engine layer, a group intelligent decision-making layer, a block chain evidence storage layer and a dynamic scheduling execution layer, and forms a'sensing-modeling-decision-evidence storage-execution 'closed loop; the physical sensing layer collects equipment and luggage state data; the digital twinborn engine layer constructs a total-factor twinborn body to realize synchronous mapping and carbon accounting of a physical system; the group intelligent decision-making layer adopts an improved carbon sensitive ant colony algorithm to generate a multi-objective optimization strategy of total energy consumption, residence time and load balance; the block chain evidence storage layer ensures credibility and traceability of the carbon data; and the dynamic scheduling execution layer converts the decision into a control signal and corrects simulation and actual deviation. According to the method, low-carbon, high-efficiency and reliable luggage scheduling is realized, and the method is suitable for green operation of a smart airport.

Description

Technical Field

[0001] The present invention relates to the field of intelligent airport logistics technology, and specifically to a digital twin low-carbon scheduling system for airport baggage flow swarm intelligence AI-driven. Background Art

[0002] Current airport baggage handling systems rely heavily on manual experience or simple automated scheduling, which presents significant limitations:

[0003] Inefficient operations: Traditional scheduling strategies are mostly based on fixed route planning, making it difficult to respond in real time to dynamic scenarios such as baggage flow fluctuations and sudden equipment failures. This can easily lead to problems such as chute blockage and empty conveyor belts, increasing baggage hold time and affecting flight connection efficiency.

[0004] Lack of carbon emission control: Energy consumption and carbon emissions from equipment operations (such as conveyor belts and robots) are not included in the core objectives of scheduling decisions. Most equipment operates at full capacity, resulting in inefficient energy consumption (such as idling and overloading). Furthermore, there is a lack of an accurate carbon footprint accounting mechanism, making it difficult to support the airport's carbon neutrality goals.

[0005] Insufficient equipment coordination: Each processing unit (sorting area, transfer area, robot operation area) mostly operates independently, lacking global coordinated optimization. This leads to uneven equipment load, shortened service life of some equipment due to long-term overload, and high maintenance costs.

[0006] Furthermore, the limitations of existing technologies in terms of intelligence and low carbonization are as follows:

[0007] Insufficient application of digital twin technology: Although some airports have introduced digital twin technology, most of them remain at the level of equipment status monitoring and have not achieved deep integration with scheduling decision-making. They are unable to predict the energy consumption and carbon emission impact of scheduling strategies through virtual simulation, resulting in a "disconnect between simulation and execution";

[0008] Insufficient adaptability of swarm intelligence algorithms: Existing swarm intelligence algorithms (such as traditional ant colonies and genetic algorithms) prioritize efficiency in baggage scheduling and lack a carbon-sensitive decision-making mechanism, making it difficult to balance efficiency and carbon reduction.

[0009] Poor credibility and traceability of carbon data: Carbon emissions data often relies on manual statistics or single-device records, which poses a risk of data tampering. Furthermore, there is a lack of a full-process evidence storage and audit mechanism, making it unable to meet the needs of carbon quota management and third-party compliance audits.

[0010] Weak dynamic response and fault tolerance: Traditional systems lack real-time correction mechanisms for "simulation-actual deviations" caused by sensor errors, equipment aging, etc. When a fault occurs, redundant switching relies on manual intervention, which can easily cause interruptions in the processing process.

[0011] Therefore, to address the above issues, a digital twin low-carbon scheduling system for airport baggage flow swarm intelligence AI-driven is proposed. Summary of the Invention

[0012] The purpose of the present invention is to provide a digital twin low-carbon scheduling system driven by swarm intelligence AI for airport baggage flow to solve the problems raised in the above background technology.

[0013] To achieve the above object, the present invention provides the following technical solutions:

[0014] The digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI includes:

[0015] The physical sensing layer, consisting of RFID scanners, current sensors, and laser rangefinders deployed in the baggage handling area, collects real-time information on conveyor belt speed, robot joint angles, chute congestion rates, and 3D baggage coordinates.

[0016] The digital twin engine layer builds a full-factor twin that includes equipment physical models, baggage flow network models, and carbon accounting models, and achieves synchronous mapping of physical systems through real-time data-driven implementation.

[0017] The swarm intelligence decision-making layer is composed of intelligent agent modules distributed across various processing units, and uses an improved carbon-sensitive ant colony algorithm to generate scheduling strategies;

[0018] The blockchain evidence layer, based on the Hyperledger architecture, stores device energy consumption and luggage carbon footprint data;

[0019] The dynamic scheduling execution layer converts decision instructions into device control signals and dynamically modifies strategies based on the twin simulation results.

[0020] As a preferred solution, the digital twin engine layer includes:

[0021] Device-level mirror module: Device-level mirror module: Establish a conveyor belt power calculation model ,in, is the real-time power of the conveyor belt, are the air resistance and mechanical friction coefficient, is the load friction coefficient, is the conveyor belt speed, is the total mass of luggage on the conveyor belt;

[0022] Carbon accounting module: calculate the carbon footprint of a single piece of luggage ,in, Carbon footprint of a single piece of luggage, The total number of equipment handling the baggage. For devices The average power, For devices How long it takes to process the baggage, is the carbon intensity factor of the power grid.

[0023] As a preferred solution, the swarm intelligence decision-making layer adopts a global optimization objective function: ,in, 、 、 is an adjustable weight coefficient , is the total number of devices, For devices The total energy consumption points, For luggage The residence time, is the standard deviation function, For devices workload.

[0024] As a preferred solution, the improved carbon-sensitive ant colony algorithm includes:

[0025] Path selection probability calculation: ,in, For intelligent agents Select Path The probability of For path The pheromone concentration, is the pheromone importance factor, is the path heuristic value ( is the path distance, For nodes The congestion level, is the heuristic information importance factor, is the carbon adaptation factor, is the carbon sensitivity coefficient, For path of carbon emissions, is the carbon emission baseline value of the pathway, is the dynamic carbon sensitivity coefficient, For nodes The set of adjacent nodes of For path pheromone concentration;

[0026] Pheromone update rules: ,in, For path The pheromone concentration, is the pheromone volatility factor and 0< , is the total number of agents, For intelligent agents On the path The increase in pheromone released is the carbon learning rate factor, is the total amount constant of pheromone, is the optimal carbon emission threshold of the system.

[0027] As a preferred option, the dynamic carbon sensitivity coefficient is adjusted through a feedback mechanism: ,in, for Carbon sensitivity coefficient at the moment, To adjust the step size coefficient, for Average carbon intensity during the period, is the airport carbon emission target value, It is the tolerance range of carbon emissions.

[0028] As a preferred solution, the blockchain evidence layer implements:

[0029] Construct a Merkle tree to store carbon data, with its root node generated by the hash concatenation of all device energy consumption data and luggage carbon footprint data;

[0030] Using zero-knowledge proof technology, it verifies whether the total carbon emissions of luggage meet the preset limit without leaking detailed data.

[0031] As a preferred solution, the dynamic scheduling execution layer includes a policy modification module:

[0032] The deviation of the comprehensive optimization index of the system is calculated in real time, and its value is the absolute value of the difference between the twin prediction value and the actual value divided by the prediction value;

[0033] When the deviation exceeds the threshold, adjust the energy consumption weight coefficient α and the initial carbon sensitivity coefficient , where α is adjusted proportionally with the sign of the energy consumption deviation, Adjustments will be made proportionally based on the carbon emission deviation rate.

[0034] As a preferred solution, the swarm intelligence decision-making layer implements a collaborative optimization mechanism:

[0035] Design an incentive function based on carbon credits: ,in, For intelligent agents of carbon credits, is the carbon emission reduction per unit time, is the time decay factor;

[0036] The Byzantine fault-tolerant consensus algorithm is used, which requires that the decision vector of each agent obtains the consent of at least twice the number of faulty nodes plus one node among its neighboring nodes.

[0037] As a preferred solution, the robustness optimization module is embedded in the digital twin engine layer:

[0038] Construct a disturbance-resistant scheduling model with the goal of minimizing the expected value of the scheduling strategy under disturbance scenarios plus the risk factor multiplied by the 95% confidence level;

[0039] Generates a Pareto optimal solution set consisting of strategies that satisfy the following conditions: there is no other strategy that is not inferior to this strategy in all optimization objectives and is strictly better in at least one objective.

[0040] As a preferred solution, the improved carbon-sensitive ant colony algorithm uses a convergence guarantee mechanism:

[0041] Define the potential game model, ,in, is the potential function, is the decision set of the agent, is the device number, For devices Energy consumption, For luggage The residence time, 、 is the path node number, For path pheromone concentration;

[0042] Prove that the cross-partial derivatives of the potential function with respect to any two agents’ decisions are non-negative;

[0043] Set the convergence conditions: ,in, is the two-norm, for The pheromone concentration vector at time , for The pheromone concentration vector at time , is the convergence accuracy constant, is the convergence rate factor.

[0044] From the technical solutions provided by the present invention, it can be seen that the digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI provided by the present invention has the following beneficial effects:

[0045] 1. Strengthen low-carbon emission reduction capabilities and contribute to the dual carbon goals:

[0046] Accurately manage carbon footprints: Leveraging the digital twin engine's carbon accounting module and blockchain evidence storage technology, the carbon footprint of each piece of baggage can be tracked and verified throughout the entire process, ensuring the credibility of carbon data and providing a reliable basis for airport carbon quota management.

[0047] Carbon-sensitive decision-making drives emission reductions: The carbon-sensitive ant colony algorithm in the swarm intelligence decision-making layer prioritizes low-carbon paths through carbon fitness factors and dynamic adjustment mechanisms, effectively reducing system carbon emissions and directly supporting the airport's carbon neutrality goals.

[0048] Collaborative optimization of energy consumption and carbon emissions: The global optimization goal balances energy consumption and efficiency through dynamic weighting, avoiding ineffective carbon emissions caused by equipment idling or overloading, and achieving coordinated low-carbon and efficient operation;

[0049] 2. Improve processing efficiency and reduce operating costs:

[0050] Shorten baggage detention time: Combining an improved ant colony algorithm with congestion simulation using digital twins, we can avoid route congestion in advance, improve baggage punctuality, and reduce problems caused by delays.

[0051] Balanced equipment load: Through load balancing optimization and dynamic correction mechanism, local equipment overload is avoided, equipment service life is extended, and maintenance costs are reduced;

[0052] Improved decision-making efficiency: The distributed intelligent collaborative architecture and fault-tolerant consensus mechanism accelerate decision-making response speed and can adapt to the baggage handling needs during peak hours;

[0053] 3. Enhance system reliability and adapt to complex working conditions:

[0054] Dynamically correct system deviations: The deviation calculation and parameter calibration mechanism of the dynamic scheduling execution layer can compensate for the deviation between the simulation and physical systems in real time to ensure the accuracy of strategy execution;

[0055] Enhanced anti-interference and emergency response capabilities: Robustness optimization models and redundant switching mechanisms reduce performance losses in the event of equipment failures or traffic fluctuations, quickly restore processing flows, and ensure continuity;

[0056] Achieve multi-scenario adaptation: Through dynamic weight adjustment and Pareto optimal solution sets, it adapts to different scenario requirements (such as prioritizing efficiency during peak hours and low-carbon development during off-peak hours), improving system adaptability;

[0057] 4. Build a trusted data foundation to support end-to-end management:

[0058] Ensure carbon data is credible and auditable: The Merkle tree and zero-knowledge proof technology of the blockchain evidence storage layer ensure that carbon data cannot be tampered with throughout the entire process, meeting compliance audit requirements;

[0059] Full-process traceability and responsibility definition: Combining RFID tags with blockchain ledgers enables data traceability throughout the entire baggage handling process, providing a basis for problem definition and reducing dispute resolution costs.

[0060] 5. Promote technological innovation and set industry benchmarks:

[0061] Integration of swarm intelligence and digital twins: The pioneering closed-loop mechanism of "twin rehearsal - intelligent decision-making - physical feedback" solves the lag problem of traditional scheduling and provides a paradigm for the intelligentization of complex logistics systems;

[0062] Modular architecture facilitates expansion: The five-layer loosely coupled architecture supports independent upgrades of each module, reducing upgrade costs and facilitating rapid deployment at airports of different sizes. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 This is a schematic diagram of the overall structure of the digital twin low-carbon scheduling system driven by swarm intelligence AI for airport baggage flow of the present invention. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0065] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0066] like Figure 1 As shown, the embodiment of the present invention provides a digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI, including:

[0067] The physical sensing layer, consisting of RFID scanners, current sensors, and laser rangefinders deployed in the baggage handling area, collects real-time information on conveyor belt speed, robot joint angles, chute congestion rates, and 3D baggage coordinates.

[0068] The digital twin engine layer builds a full-factor twin that includes equipment physical models, baggage flow network models, and carbon accounting models, and achieves synchronous mapping of physical systems through real-time data-driven implementation.

[0069] The swarm intelligence decision-making layer is composed of intelligent agent modules distributed across various processing units, and uses an improved carbon-sensitive ant colony algorithm to generate scheduling strategies;

[0070] The blockchain evidence layer, based on the Hyperledger architecture, stores device energy consumption and luggage carbon footprint data;

[0071] The dynamic scheduling execution layer converts decision instructions into device control signals and dynamically modifies strategies based on the twin simulation results.

[0072] In this embodiment, the physical perception layer is the data collection core of the airport baggage low-carbon dispatch system based on swarm intelligence and digital twins. Like the system's "nerve endings," it uses various sensor devices deployed in the baggage handling area to capture key physical parameters of baggage flow and equipment operation in real time, providing accurate and comprehensive raw data support for digital twin modeling and intelligent decision-making. The following describes this layer in detail, from the overall perspective:

[0073] 1. Overview of overall functions:

[0074] The physical perception layer is primarily responsible for comprehensive, high-precision, and real-time monitoring of the physical state of the airport's baggage handling area. This includes individual baggage information (such as identification, three-dimensional dimensions, and weight), equipment operating parameters (such as conveyor belt speed, motor current, and robot joint angles), and environmental and operating condition data (such as chute blockage status and area temperature). Through the collaborative work of multiple types of sensors, the layer converts dynamic changes in the physical world into quantifiable digital signals, which are then pre-processed and transmitted to the digital twin engine layer to establish a real-time mapping foundation between the physical system and the virtual model. This layer also provides the swarm intelligence decision-making layer with raw data for energy consumption and carbon footprint calculations, forming the primary link in achieving a closed loop of "perception-modeling-decision-making-execution."

[0075] 2. Submodule composition and functions:

[0076] (1) RFID scanner unit:

[0077] Baggage identification and information collection: RFID scanners deployed at baggage sorting lanes, conveyor belt entrances, and key nodes read the unique ID, flight information, destination code, and other data contained in baggage tags in real time. Using ultra-high frequency RFID technology (operating at 860-960MHz), it supports simultaneous multi-tag identification with a range of up to 3-5 meters, ensuring that baggage can be quickly located and tracked once it enters the handling area.

[0078] Data association and synchronization: The baggage identification read by RFID is associated with the three-dimensional coordinates collected by the laser rangefinder and the quality data obtained by the weight sensor to form a complete baggage file containing "ID-location-size-weight-flight information". (in, For luggage The comprehensive information tag, For luggage for Luggage at time The three-dimensional coordinates of For luggage The three-dimensional dimensions (length Width high), For luggage weight, For luggage Flight information) to achieve data structured storage;

[0079] Abnormal identification and alarm: When the RFID scanner fails to read the baggage tag in a certain area for three consecutive times (exceeding the preset threshold), ) triggers a tag missing alarm, judging that the baggage tag has fallen off or is damaged, and uses the linked camera for image-assisted recognition to ensure baggage traceability;

[0080] (2) Current sensor unit:

[0081] Equipment energy consumption parameter collection: Hall current sensors are connected in series in the power supply circuits of electrical equipment such as conveyor belt motors, robot drive modules, and chute control motors to monitor the working current in real time. , combined with the device rated voltage , through the formula (in, For devices exist The instantaneous power at the moment, For devices Rated operating voltage, for Time equipment The working current, For devices The power factor is used to calculate the instantaneous energy consumption and provide basic data for the carbon accounting module;

[0082] Equipment failure prediction: By analyzing the current change curve, when it is detected that the current fluctuation amplitude exceeds the normal range (If the peak value of the motor starting current exceeds 1.5 times the rated current and the duration is > ) is determined to be an abnormal operating state of the equipment, and an early warning signal is generated and transmitted to the dynamic scheduling execution layer, prompting maintenance inspection;

[0083] Data calibration and compensation: Considering the influence of sensor temperature drift, the temperature compensation formula is used (in, is the calibrated current value, is the raw current reading of the sensor, is the temperature coefficient, is the current ambient temperature, The current measurement accuracy is within ±0.5% at the standard calibration temperature (25°C).

[0084] (3) Laser rangefinder unit:

[0085] Baggage 3D coordinate and size measurement: Multiple laser rangefinders are deployed on both sides and above the conveyor belt (measurement accuracy mm, sampling frequency 100Hz), the distance data of each point on the luggage surface is obtained by triangulation principle, and the three-dimensional coordinates of the luggage are obtained by coordinate conversion calculation and length ( ),Width( ),high ) size parameters, providing a spatial basis for baggage sorting path planning;

[0086] Chute blocking status monitoring: Install laser rangefinders at the entrance and bend of the baggage chute to measure the distance change in the chute in real time; when the distance value is less than the preset threshold value within 5 consecutive sampling periods, the (such as 1 / 3 of the chute diameter), it is determined that the chute is blocked and the blockage rate parameter is generated. (in, is the chute blocking rate, is the duration of the blockage, is the total duration of the monitoring cycle), which is transmitted to the dynamic scheduling execution layer to trigger the dredging instruction;

[0087] Robot workspace perception: LiDAR (scanning frequency 10Hz, angular resolution 0.5°) is deployed at the end of the baggage handling robot's robotic arm and around its work area to construct a point cloud map of the robot's workspace. This allows for real-time obstacle detection (including unrecognized luggage and unusually protruding parts of equipment), providing obstacle avoidance data for robot joint angle adjustment to ensure operational safety.

[0088] (IV) Data preprocessing and fusion unit:

[0089] Noise filtering and smoothing: Filter the raw sensor data, using Kalman filtering for laser ranging data. (in, for The filtered value at the moment, for The predicted value at time, is the Kalman gain, for The measured value at the moment, is the measurement matrix); the current sensor data is filtered using a sliding average filter (in, for The smoothed current value at the moment, is the sliding window size (5-10), For the The original current value at the moment is obtained to remove environmental interference and measurement noise;

[0090] Time-space synchronization and data alignment: Based on a unified system clock (synchronization accuracy ≤ 1ms), the collected data of RFID scanners, current sensors, and laser rangefinders are aligned by timestamp to solve the time deviation caused by the sampling frequency differences of different devices; for spatial coordinate data, the preset coordinate system conversion matrix is ​​used (in, is the global coordinate system coordinate, is a 3×4 transformation matrix, The local coordinates of each device are converted to the global coordinate system to achieve spatial consistency of data;

[0091] Abnormal data elimination and completion: By setting data rationality thresholds (such as luggage weight range 5-50kg, conveyor speed 0-2m / s), outliers that exceed the range are eliminated; for short-term data loss (≤3 sampling cycles), linear interpolation is used (in, To complete the time data, 、 for Valid data adjacent to each other, and ) to complete and ensure data continuity;

[0092] 3. Key technical principles:

[0093] (1) Principle of multi-sensor collaborative perception:

[0094] The physical perception layer utilizes a multi-sensor fusion architecture comprised of RFID, current sensors, and laser rangefinders, achieving comprehensive monitoring based on the principle of information complementarity. RFID technology addresses the issue of linking individual baggage identification and information, while current sensors focus on monitoring equipment energy consumption and operating status. Laser rangefinders provide spatial dimensions and location information. The three sensor data, aligned in time and space, form a three-dimensional data structure: identity, energy consumption, and space. Data fusion enhances the system's ability to perceive complex operating conditions, overcoming the limitations of single sensors in scenarios such as occlusion, interference, and insufficient information.

[0095] (2) Principle of real-time data collection and transmission:

[0096] Sensors are deployed in a distributed manner, connected to the edge computing gateway via industrial Ethernet (transmission rate of 100Mbps), and use a publish-subscribe model (MQTT protocol) for data transmission. This ensures that collected real-time data (such as conveyor belt speed and current value) is uploaded to the digital twin engine layer in a 50ms cycle. For high-frequency collected data (such as laser point clouds, 10Hz), lightweight processing (such as downsampling) is performed on edge nodes to reduce data transmission bandwidth pressure. At the same time, a timestamp and verification mechanism is used to ensure the timing accuracy of the data.

[0097] (3) Data preprocessing optimization principle:

[0098] To address issues such as noise, bias, and missing data in sensor data, differentiated preprocessing strategies are adopted based on the characteristics of different data types: Kalman filtering is suitable for continuous data containing Gaussian noise, such as laser ranging, and achieves dynamic noise suppression through prediction-update iteration; sliding average filtering is suitable for high-frequency fluctuating data such as current, and retains trend characteristics through smoothing; linear interpolation rules are used for scenarios with short-term data loss, achieving efficient data completion based on the assumption of data continuity, providing high-quality data input for subsequent digital twin modeling and carbon accounting;

[0099] 4. Module workflow:

[0100] (1) Initialization phase:

[0101] After the physical sensing layer is started, all sensors are self-tested: the RFID scanner performs a tag reading test (identifying the preset test tag), the current sensor detects zero-point drift (≤5mA is normal), and the laser rangefinder measures the reference distance (deviation from the preset value is ≤2mm is normal) to ensure that the hardware equipment is working properly;

[0102] Complete the clock synchronization of the sensor network (based on the NTP protocol), load the calibration parameters of each sensor (such as the conversion matrix , temperature coefficient ), data threshold (such as congestion determination distance , abnormal current range), establish a communication connection with the edge gateway, and enter the standby state;

[0103] (2) Data collection stage:

[0104] When a piece of baggage enters the processing area and triggers the entrance photoelectric sensor, the system activates the RFID scanner and laser rangefinder in that area. The RFID reads the baggage tag information, the laser rangefinder simultaneously collects the three-dimensional dimensions and coordinates, and the weight sensor (integrated in the conveyor belt) obtains the baggage mass, thus forming an initial baggage file.

[0105] Current sensors continuously monitor the operating current of conveyor belt motors, robots, and other equipment, recording the current value every 50ms. Laser rangefinders periodically scan the chutes and robot operating areas (10Hz) to update blockage status and spatial obstacle information in real time.

[0106] (3) Data preprocessing and fusion stage:

[0107] After receiving the data from each sensor, the edge gateway sorts it by timestamp, performs Kalman filtering on the laser ranging data, and performs sliding average filtering on the current data to remove outliers that are outside the reasonable range;

[0108] The local coordinates of the laser ranging are converted into global coordinates through coordinate system conversion, and the RFID baggage ID is associated with the laser size and weight data to generate a structured data frame. ,in, is the timestamp, is the RFID tag data, is the data processed by laser ranging, is the smoothed current data, It is the working status of the sensor;

[0109] (IV) Data transmission and exception response phase:

[0110] The pre-processed structured data is transmitted to the digital twin engine layer periodically (50ms), and the anomaly detection mechanism is triggered at the same time: if an abnormal state such as missing label, current exceeding the limit, or chute blockage is detected, an abnormal signal is generated and transmitted to the dynamic scheduling execution layer first;

[0111] When the sensor's communication is interrupted, the local cache mechanism is activated to temporarily store the collected data in the edge node (storage capacity ≥ 1GB). After the communication is restored, the data will be re-uploaded in chronological order to ensure that the data is not lost.

[0112] (V) Shutdown phase:

[0113] When the system receives a shutdown command, the physical perception layer stops data collection, executes the sensor shutdown process (such as the laser rangefinder enters low-power mode), uploads all cached untransmitted data, records the sensor working status of this operation (such as operating time and number of exceptions), and closes the communication connection after completing data archiving.

[0114] In this embodiment, the digital twin engine layer includes:

[0115] Device-level mirror module: Establishing a conveyor belt power calculation model ,in, is the real-time power of the conveyor belt, are the air resistance and mechanical friction coefficient, is the load friction coefficient, is the conveyor belt speed, is the total mass of luggage on the conveyor belt;

[0116] Carbon accounting module: calculate the carbon footprint of a single piece of luggage ,in, Carbon footprint of a single piece of luggage, The total number of equipment handling the baggage. For devices The average power, For devices How long it takes to process the baggage, is the grid carbon intensity factor;

[0117] Furthermore, the digital twin engine layer serves as the "virtual hub" of the airport baggage low-carbon dispatch system based on swarm intelligence and digital twins. By building a fully digital mirror of the physical system, it enables real-time mapping, dynamic simulation, and optimization deduction between physical states and virtual models, providing a precise virtual testing ground for swarm intelligent decision-making. It is the core link between physical perception and intelligent decision-making. The following is a detailed explanation from the overall perspective:

[0118] 1. Overview of overall functions:

[0119] Driven by real-time data collected by the physical perception layer, the digital twin engine layer constructs a full-factor twin consisting of equipment physical models, baggage flow network models, carbon accounting models, and robustness optimization models. By dynamically updating virtual model parameters, it achieves real-time simulation of the baggage handling process's "equipment operation-baggage flow-energy consumption and carbon emissions," accurately replicating the spatiotemporal dynamic characteristics of the physical system. Furthermore, this layer can perform multiple scenario simulations based on the current state (such as equipment failures and baggage peaks), generate optimization target predictions, and provide simulation verification data for scheduling strategies for the swarm intelligence decision-making layer. This supports the closed-loop mechanism of "physical world operation-virtual model feedback-decision-making strategy optimization," and serves as the core modeling and simulation platform for achieving low-carbon scheduling and efficient operation.

[0120] 2. Submodule composition and functions:

[0121] (1) Device-level mirroring module:

[0122] Multi-device physical modeling: Build high-precision physical models for core equipment such as conveyor belts, robots, and chutes;

[0123] Conveyor motor power model: ,in, is the real-time power of the conveyor belt, are the air resistance and mechanical friction coefficient, is the load friction coefficient, is the conveyor belt speed, is the total mass of luggage on the conveyor belt. This model comprehensively considers the belt speed cube term (dominated by air resistance) and the load-speed product term (dominated by friction resistance), and is more consistent with the energy consumption characteristics under high loads.

[0124] Robot joint energy consumption model: ,in, is the total power of the robot, is the number of joints, For the Joint output torque, For the Joint angular velocity, For the Joint resistivity, For the Joint current, accurately depicting the relationship between robot motion energy consumption and joint status;

[0125] Real-time mapping of device status: Based on the current, speed, joint angle and other data of the physical perception layer, the device model parameters are updated every 100ms, and the deviation correction algorithm is used to (in, are the modified model parameters, are the theoretical parameters of the model, is the correction factor (0.1-0.3), is the measured value of the physical perception layer, (predicted value for the model) to achieve dynamic alignment between the virtual model and the physical device to ensure the mirroring accuracy error ;

[0126] Equipment fault simulation: built-in common fault model library (such as conveyor belt slippage, robot joint jamming), by injecting fault parameters (such as conveyor belt friction coefficient Sudden 30% surge to simulate slippage), simulating equipment energy consumption and operating characteristics under fault conditions, providing fault scenario data for robust decision-making;

[0127] (2) Baggage flow network modeling unit:

[0128] Spatial network topology construction: abstracting the baggage handling area into a "node-edge" network model (in, It is a collection of nodes, including key locations such as conveyor belt starting / ending points, chute entrances, and sorters; is a set of edges, representing the connection relationship between devices, such as conveyor belt segments and chute channels), and each edge is associated with physical parameters (length , Maximum load capacity , energy consumption coefficient );

[0129] Baggage flow dynamic simulation: Based on the baggage coordinate and speed data of the physical perception layer, the discrete event simulation method is used to track the movement trajectory of a single piece of baggage in the network. (in, For luggage exist The location at the moment, For the edge The device speed, is the simulation step size, For the edge The direction coefficient) calculates the next moment position and simultaneously records the luggage's detention time at each node ;

[0130] Congestion diffusion simulation: When an edge Real-time carrying capacity When the congestion model is triggered, (in, is the speed in congestion, Simulate speed decay for the congestion decay coefficient (2-5) and calculate the diffusion time of congestion to adjacent nodes to provide prediction data for path optimization;

[0131] (3) Carbon accounting module:

[0132] Carbon footprint of a single piece of luggage throughout the entire process: ,in, Carbon footprint of a single piece of luggage, The total number of equipment handling the baggage. For devices The average power, For devices How long it takes to process the baggage, is the grid carbon intensity factor, which is updated with real-time data of the regional grid;

[0133] Real-time statistics of regional total carbon emissions: aggregate carbon footprint data by processing area (such as departure sorting area, transfer area) to generate ,in, For the region Total carbon emissions at any moment, For luggage within the area The cumulative carbon footprint of For devices No-load power, providing regional carbon emission data for carbon-sensitive decision-making;

[0134] Carbon emission trend forecast: Based on historical carbon emission data and current baggage flow, LSTM neural network is used to predict the carbon emission curve for the next 30 minutes. ,in, for Predicted carbon emissions at each moment, To predict the duration, is the current baggage flow, 、 is the historical time step, the prediction error ;

[0135] (IV) Robustness Optimization Module:

[0136] Anti-interference scheduling model: For disturbance scenarios such as equipment failure and baggage peak, a robust model is built that optimizes the dual indicators of target expectation and risk value: ,in, For the disturbance scene set The expected value operator on , For disturbance scenarios (such as equipment failure and baggage peak), is the set of all possible disturbance scenarios, For the scheduling strategy vector x and the disturbance scenario The optimization objective function value under To optimize the objective function value Value at risk at a 95% confidence level;

[0137] Pareto optimal solution set generation: A multi-objective genetic algorithm is used to find a balance between low carbon and high efficiency, generating a Pareto optimal solution set. ,in, is the Pareto optimal solution set, is the scheduling policy vector, for dimensional real space, is another scheduling policy vector, For the The objective function, Indicates that for all , Indicates that it does not exist. Index of a specific objective function for the decision-making layer to select based on real-time needs;

[0138] (V) Data synchronization and mapping unit:

[0139] Real-time data fusion: Receive the structured data frame of the physical perception layer and align it with the timestamp (synchronization accuracy) ms) and spatial coordinate conversion, mapping sensor data to the corresponding components of the virtual model, updating model parameters (such as conveyor belt speed ), ensuring that the state deviation between the virtual model and the physical system ;

[0140] Model lightweighting and acceleration: Use model reduction technology for highly complex models (such as robot joint dynamics) (in, For the reduced-order model, PCA is the principal component analysis. For the original model, is the model parameter) to increase the simulation speed by 5-10 times, meeting the real-time requirements (single-step simulation time ;

[0141] Data feedback mechanism: Package the simulation results of the virtual model (such as predicted carbon emissions, optimized target values) into feedback data frames (in, To predict carbon emissions, To optimize the target prediction value, is the congestion risk level), which is transmitted to the swarm intelligence decision-making layer and the dynamic scheduling execution layer;

[0142] 3. Module workflow:

[0143] (1) Initialization phase:

[0144] Load the physical layout parameters of the airport baggage handling system (such as equipment location, network topology), equipment physical model coefficients (such as 、 ), initial carbon intensity factor , build a basic virtual twin framework;

[0145] Receive self-test data from the physical perception layer to verify the mapping relationship between the sensor and the model (such as the deviation between the laser ranging coordinates and the virtual node). cm), complete the model calibration;

[0146] (2) Real-time mapping stage:

[0147] Receive real-time data from the physical sensing layer (device current, baggage coordinates, speed, etc.) in a 50ms cycle, and update the state parameters of the virtual model (such as conveyor belt power) through the data synchronization unit. , luggage location );

[0148] Run baggage flow network simulation, track the virtual trajectory of all bags, calculate real-time indicators such as detention time and regional carbon emissions, and trigger model correction when the deviation from physical data exceeds 10%;

[0149] (III) Simulation and prediction stage:

[0150] Receive the scheduling strategy of the swarm intelligence decision layer (such as conveyor belt speed adjustment, path planning), simulate the execution effect of the strategy in the virtual twin, and output the optimization target value and carbon emission forecasts ;

[0151] For high-priority strategies (such as emergency dispatch to deal with flight delays), the robustness optimization module is activated to simulate the strategy performance under 3-5 disturbance scenarios and generate a risk assessment report;

[0152] (IV) Feedback and update stage:

[0153] The simulation results ( , , risk assessment) is fed back to the group intelligent decision-making layer to support strategy iteration; real-time carbon emission data is transmitted to the blockchain evidence layer for carbon footprint evidence;

[0154] Update model coefficients every hour (such as based on historical data optimization 、 ), daily updated grid carbon intensity factor , ensuring model adaptability;

[0155] (V) Shutdown phase:

[0156] Save the virtual twin's operating data for the day (such as energy consumption simulation curves and carbon emission statistics) and generate a model accuracy report (average deviation, maximum deviation);

[0157] Shut down the simulation engine, release computing resources, and disconnect the communication with the physical perception layer and decision-making layer.

[0158] In this embodiment, the swarm intelligence decision layer adopts a global optimization objective function: ,in, 、 、 is an adjustable weight coefficient , is the total number of devices, For devices The total energy consumption points, For luggage The residence time, is the standard deviation function, For devices workload;

[0159] The improved carbon-sensitive ant colony algorithm includes:

[0160] Path selection probability calculation: ,in, For intelligent agents Select Path The probability of For path The pheromone concentration, is the pheromone importance factor, is the path heuristic value ( is the path distance, For nodes The congestion level, is the heuristic information importance factor, is the carbon adaptation factor, is the carbon sensitivity coefficient, For path of carbon emissions, is the carbon emission baseline value of the pathway, is the dynamic carbon sensitivity coefficient, For nodes The set of adjacent nodes of For path pheromone concentration;

[0161] Pheromone update rules: ,in, For path The pheromone concentration, is the pheromone volatility factor and 0< , is the total number of agents, For intelligent agents On the path The increase in pheromone released is the carbon learning rate factor, is the total amount constant of pheromone, is the optimal carbon emission threshold of the system;

[0162] The dynamic carbon sensitivity coefficient is adjusted through a feedback mechanism: ,in, for Carbon sensitivity coefficient at the moment, To adjust the step size coefficient, for Average carbon intensity during the period, is the airport carbon emission target value, The tolerance range for carbon emissions;

[0163] The improved carbon-sensitive ant colony algorithm uses a convergence guarantee mechanism:

[0164] Define the potential game model, ,in, is the potential function, is the decision set of the agent, is the device number, For devices Energy consumption, For luggage The residence time, 、 is the path node number, For path pheromone concentration;

[0165] Prove that the cross-partial derivatives of the potential function with respect to any two agents’ decisions are non-negative;

[0166] Set the convergence conditions: ,in, is the two-norm, for The pheromone concentration vector at time , for The pheromone concentration vector at time , is the convergence accuracy constant, is the convergence rate factor;

[0167] Furthermore, the swarm intelligence decision-making layer is the "intelligent brain" of the airport baggage low-carbon scheduling system based on swarm intelligence and digital twins. It is composed of intelligent agent modules distributed across various processing units. It achieves global optimization decisions through an improved carbon-sensitive ant colony algorithm, minimizing carbon emissions while ensuring baggage handling efficiency. It is the key decision-making hub connecting digital twin simulation and dynamic execution. The following is a detailed explanation from the overall perspective:

[0168] 1. Overview of overall functions:

[0169] The swarm intelligence decision-making layer uses real-time simulation data (such as equipment energy consumption, baggage location, and carbon emission forecasts) output by the digital twin engine layer as input. Through the collaborative computing of distributed intelligent agents, it constructs a "local perception-global optimization" decision-making mechanism. The core task of this layer is to generate the optimal scheduling strategy through an improved carbon-sensitive ant colony algorithm, while meeting the timeliness of baggage transfer (such as the maximum detention time constraint). This strategy includes conveyor belt speed adjustment, baggage path planning, and robot operation allocation, achieving a multi-objective balance between total energy consumption, carbon emissions, and processing efficiency. At the same time, a dynamic feedback mechanism is used to adapt to the deviation correction of the digital twin simulation to ensure the robustness and adaptability of the decision-making strategy in the physical system. This is the core algorithm carrier for the system to achieve the dual goals of "low carbon and high efficiency".

[0170] 2. Submodule composition and functions:

[0171] (1) Distributed Agent Module:

[0172] Agent deployment and division of labor: Agents are distributed and deployed according to baggage handling areas (such as sorting areas, transfer areas, and chute groups). Each agent is responsible for equipment scheduling and baggage routing decisions within its jurisdiction, forming a "regional autonomy + global collaboration" architecture. Agents have the ability to receive data (simulation data from digital twins), perform local computing (route evaluation), and exchange information (sharing decisions with neighboring agents). Millisecond-level information synchronization is achieved through lightweight communication protocols (such as MQTT-SN).

[0173] Local decision-making and global coordination: A single agent generates a local scheduling plan based on the real-time status of the jurisdiction area (such as chute congestion rate, equipment load), and uses the formula (in, For intelligent agents Local solutions, is the regional status data, is the regional carbon emission data, is the regional residence time); at the same time, the decision vectors of surrounding intelligent agents are obtained through neighborhood communication, and a consensus mechanism is used to aggregate them into a global strategy to avoid local optimal traps;

[0174] (2) Carbon-sensitive ant colony algorithm unit:

[0175] Path selection probability calculation: The agent uses an improved path selection probability formula and introduces a carbon fitness factor based on the traditional ant colony algorithm to give priority to low-carbon and high-efficiency paths: ,in, For intelligent agents Select Path The probability of For path The pheromone concentration, is the pheromone importance factor, is the path heuristic value ( is the path distance, For nodes congestion level), is the heuristic information importance factor, is the carbon adaptation factor ( is the carbon sensitivity coefficient, For path of carbon emissions, is the carbon emission baseline value of the pathway), is the dynamic carbon sensitivity coefficient, For nodes The set of adjacent nodes of For path pheromone concentration;

[0176] Pheromone update rules: To strengthen the accumulation of pheromones in the low-carbon path, an update formula integrating carbon emission optimization is adopted: ,in, For path The pheromone concentration, is the pheromone volatility factor (0< , is the total number of agents, For intelligent agents On the path The increase in pheromone released is the carbon learning rate factor, is the total amount constant of pheromone, is the optimal carbon emission threshold of the system); when the path carbon emission near When , the pheromone increment is significantly improved, guiding more agents to choose this path;

[0177] Dynamic carbon sensitivity coefficient adjustment: real-time adjustment through feedback mechanism , the formula is: (in, for Carbon sensitivity coefficient at the moment, To adjust the step size coefficient, for Average carbon intensity during the period, is the airport carbon emission target value, is the tolerance range of carbon emissions); when the actual carbon emissions are higher than the target value, Increase, strengthen the algorithm's preference for low-carbon paths; conversely, reduce sensitivity and balance efficiency needs;

[0178] (3) Global optimization target unit:

[0179] Multi-objective optimization function construction: Taking total energy consumption, maximum residence time, and load imbalance as optimization targets, a weighted sum global objective function is constructed: ,in, 、 、 is an adjustable weight coefficient , is the total number of devices, For devices The total energy consumption points, For luggage The residence time, is the standard deviation function, For devices workload); by dynamically adjusting weights (such as increasing flight peak hours , the off-peak period increases ) Adapt to scenario requirements;

[0180] Constraint handling: The decision-making process must meet hard constraints, including: baggage transfer deadline , Maximum load of equipment ), the total carbon emission limit ( ); Penalty function method is used to integrate constraints into the objective function, and penalty values ​​are imposed on solutions that violate the constraints to ensure the feasibility of the strategy;

[0181] (IV) Collaborative Optimization Mechanism Unit:

[0182] Carbon credit incentive function: To promote collaborative low-carbon decision-making among intelligent agents, an incentive mechanism based on carbon emission reduction is designed: ,in, For intelligent agents of carbon credits, is the carbon emission reduction per unit time, It is a time decay factor, emphasizing the high incentive of immediate emission reduction. The integral value is linked to the decision-making authority of the intelligent agent. The intelligent agent with higher integral value has greater weight in the global consensus, which stimulates the motivation of collaborative emission reduction.

[0183] Byzantine Fault Tolerant Consensus: To avoid the impact of local failures on global decision-making, the Byzantine Fault Tolerant (BFT) consensus algorithm is adopted. The formula is: ,in, For intelligent agents The decision vector of is the set of decision vectors of the neighborhood agents, is the upper limit of the number of faulty nodes); when it exceeds When all agents reach a consensus, the decision takes effect, ensuring that reliable strategies can still be generated when some agents fail;

[0184] 3. Key technical principles:

[0185] (1) Principles of swarm intelligence collaborative decision-making:

[0186] Based on a hybrid architecture of "distributed perception and centralized optimization," each intelligent agent generates preliminary decisions based on local perception data, and then achieves global coordination through pheromone sharing (ant colony algorithm) and consensus mechanisms. Pheromones serve as an indirect communication medium, transforming individual experience into group knowledge and avoiding delays caused by large amounts of data interaction. The carbon fitness factor and dynamic sensitivity coefficient give the group the ability to adapt to carbon emission targets, enabling dynamic switching of decision-making from "efficiency first" to "low carbon-efficiency balance."

[0187] (2) Adaptive optimization principle of carbon-sensitive algorithm:

[0188] The carbon-sensitive ant colony algorithm achieves self-adaptation through a dual feedback mechanism: one is pheromone update and the other is carbon emission threshold. linkage, strengthen the guidance of low-carbon path; second, carbon sensitivity coefficient Dynamic adjustments are made based on the system's actual carbon emissions, forming a closed loop of "carbon emission deviation, coefficient adjustment, path preference change, and carbon emission optimization." Compared to traditional ant colony algorithms, its core innovation lies in deeply integrating environmental goals (carbon emissions) into the decision-making model, rather than simply optimizing time or distance.

[0189] (3) Multi-objective trade-off principle:

[0190] The global optimization objective is determined by the weight coefficient 、 、 Achieving dynamic trade-offs among multiple objectives: Increase (such as off-peak period), the system gives priority to reducing energy consumption and carbon emissions; when Increase (such as during busy flight periods), give priority to ensuring that luggage is transferred on time; This ensures that the equipment load is balanced and avoids energy consumption surges caused by local overloads; through the simulation data feedback of digital twins, the weight coefficient can be calibrated in real time (such as the load imbalance in a certain area). When the threshold is exceeded, the ), improve decision-making flexibility;

[0191] 4. Module workflow:

[0192] (1) Initialization phase:

[0193] Load agent deployment parameters (jurisdiction area, neighboring agent list), algorithm initial parameters (pheromone concentration , weight 、 、 , carbon emission threshold ), complete the communication connection of the intelligent agent network;

[0194] Receive the initial simulation data (initial state of the equipment, no-load energy consumption) from the digital twin engine layer and initialize the baseline value of the global optimization objective function;

[0195] (2) Decision-making stage:

[0196] The intelligent agent receives real-time data from the digital twin (baggage location, equipment power, carbon emission data) at a cycle of 100ms and calculates the carbon fitness factor of each path. and heuristic value ;

[0197] Based on the path selection probability formula Generate a set of candidate paths, evaluate the comprehensive cost of each path based on the global optimization objective function, and select the local optimal path;

[0198] (III) Global coordination stage:

[0199] Agents release pheromones , and receives the pheromone data of the neighboring agents, and updates the global pheromone matrix according to the update rules ;

[0200] The BFT consensus mechanism is used to vote on the local optimal path of each agent, exceeding The agreed paths are incorporated into the global dispatch strategy and carbon credits are calculated at the same time And update the agent permissions;

[0201] (IV) Dynamic optimization stage:

[0202] Receive feedback data from the dynamic scheduling execution layer (actual carbon emissions, residence time and simulation value deviation), through Formula adjusts carbon sensitivity coefficient and corrects it through weight calibration mechanism 、 、 ;

[0203] If the digital twin simulation shows that the carbon emissions of a certain path exceed the threshold , then through the pheromone penalty mechanism (reducing the path ) Reduce agent selection and achieve dynamic optimization;

[0204] (V) Strategy output stage:

[0205] Convert the global scheduling strategy into device control instructions (such as conveyor belt speed v, robot joint angle adjustment value), package and send them to the dynamic scheduling execution layer;

[0206] Record the optimization target values ​​of this decision (total energy consumption, maximum residence time, total carbon emissions) and upload them to the blockchain evidence layer for traceability and evaluation.

[0207] In this embodiment, the blockchain evidence storage layer implements:

[0208] Construct a Merkle tree to store carbon data, with its root node generated by the hash concatenation of all device energy consumption data and luggage carbon footprint data;

[0209] Using zero-knowledge proof technology, the system verifies whether the total carbon emissions of luggage meet the preset limit without revealing detailed data.

[0210] Furthermore, the blockchain evidence layer is the "trusted ledger" of the airport baggage low-carbon dispatch system based on swarm intelligence and digital twins. Built on the Hyperledger architecture, it constructs a distributed trusted storage system specifically responsible for the full-process evidence storage, verification, and traceability of equipment energy consumption and baggage carbon footprint data, providing underlying technical support for the credibility of carbon accounting results and the auditability of system decisions. The following is a detailed explanation from the overall perspective:

[0211] 1. Overview of overall functions:

[0212] The blockchain evidence layer, centered on distributed ledger technology, receives scheduling strategy data from the swarm intelligence decision-making layer, actual operating data (such as equipment energy consumption and total carbon emissions) from the dynamic scheduling execution layer, and carbon accounting results from the digital twin engine layer. Cryptographic algorithms are used to achieve tamper-proof storage and traceable verification of data. The core tasks of this layer include: constructing a Merkle tree structure for carbon data to ensure integrity, implementing zero-knowledge proofs for carbon emissions compliance verification, and automatically executing data evidence rules through smart contracts. This provides a trusted data foundation for airport carbon management, while supporting full-process traceability and third-party audits of carbon footprints. It serves as the key trust infrastructure for the system's closed loop of "low-carbon scheduling, data trustworthiness, and compliance verification."

[0213] 2. Submodule composition and functions:

[0214] (1) Distributed ledger unit:

[0215] Ledger structure and data partitioning: Using Hyperledger Fabric's multi-channel architecture, channels are divided according to data types: device energy consumption channel (storing current and power data), carbon footprint channel (storing single piece of luggage), and regional total carbon emissions ), scheduling strategy channel (storing the optimization objectives and path selection data of swarm intelligence decision-making); each channel corresponds to an independent ledger, and node permissions are restricted by access control lists (ACLs) (for example, the airport operator has full channel access, while the auditor can only access the carbon footprint channel);

[0216] Block generation and consensus mechanism: Kafka sorting service is used to achieve orderly block generation, and data is uploaded to the chain through the "endorsement-sorting-submission" process: the intelligent node acts as an endorsement node to verify the legitimacy of the data (such as the deviation between energy consumption data and digital twin simulation values). The identification formula is ,in, for The blocks generated at the moment, For transaction data, is the endorsement verification function, It is a sorting function that ensures the consistency and immutability of block data;

[0217] (2) Merkle tree storage unit:

[0218] Merkle tree construction for carbon data: Aggregate device energy consumption and carbon footprint data by time slice (e.g., 5 minutes) and construct a Merkle tree to verify data integrity. The formula is: ,in, is the Carbon Data Merkle Tree, is the tree root hash value, is the SHA-256 hash function, For devices Energy consumption data, For luggage carbon footprint, It is a string concatenation operation; the leaf nodes of the tree are single data hashes, and the non-leaf nodes are the concatenated hashes of the child node hashes. The root hash is stored in the blockchain header to ensure that any data tampering can be discovered through root hash comparison;

[0219] Data retrieval and verification: When querying the carbon footprint of a piece of luggage, the system returns And its path proof in the Merkle tree (the hash chain from the leaf node to the root node), through the formula Verify (in, is the hash set of path proof) to verify the authenticity of the data. The verification time complexity is ( is the total amount of data) to meet real-time query needs;

[0220] (3) Zero-knowledge proof unit:

[0221] Carbon emission compliance verification: To prove that total carbon emissions do not exceed the standard without disclosing detailed data, zero-knowledge proof (ZK-SNARK) technology is used. The verification formula is: ,in, is the total carbon footprint of all luggage, is the total carbon emission threshold of the airport, is the set of equipment energy consumption data, | represents the conditional proof based on energy consumption data; the proof process converts the carbon emission calculation logic into ) into a constraint system, generating a short proof that the prover (system) can submit to the verifier (regulator), who can confirm total carbon emissions compliance without having to look at the original data;

[0222] Privacy protection and data desensitization: Desensitize sensitive data (such as the carbon footprint of luggage on a specific flight) through hash mapping (in, To anonymize luggage identification, A random salt value is used to remove personal information and retain only carbon emission-related attributes, supporting data statistical analysis while complying with privacy protection regulations (such as GDPR);

[0223] (IV) Smart Contract Unit:

[0224] Automatic execution of evidence rules: deploy carbon data evidence contract, define the format and frequency of data on-chain: equipment energy consumption data is uploaded to the chain every 30 seconds, the carbon footprint of a single piece of luggage is uploaded to the chain within 10 seconds after processing, and the total regional carbon emissions are uploaded to the chain every hour; the contract is triggered by the conditions (in, For the timing winding interval, Automatically execute the evidence storage process for baggage handling completion events to avoid data omissions caused by human intervention;

[0225] Carbon credits liquidation and auditing: Carbon credits based on swarm intelligence decision-making layer Deploy a liquidation contract to automatically redeem rewards (such as enhanced decision-making authority) when the carbon points of the intelligent body accumulate to a threshold; at the same time, deploy an audit contract to allow authorized nodes (such as third-party audit agencies) to call Function to obtain the carbon data Merkle tree and compliance proof for a specified time period to achieve automated auditing

[0226] 3. Module workflow:

[0227] (1) Initialization phase:

[0228] Deploy a Hyperledger Fabric network, including three endorsement nodes (corresponding to different processing zones), one ordering node (Kafka cluster), and multiple peer nodes (airport operator, auditor, and regulator), and configure channels and access control policies.

[0229] Deploy smart contracts (deposit contract, liquidation contract, audit contract), initialize the root hash of the Merkle tree (empty tree state), and set the total carbon emission threshold Interval between data upload ;

[0230] (2) Data storage stage:

[0231] Receive real-time device energy consumption data from the physical perception layer Carbon footprint data with the digital twin engine layer , the format and rationality are verified by the endorsement node (such as , Deviation from simulation value ;

[0232] Aggregate data by time slice and calculate hash value , update the Merkle tree and calculate the new root hash , automatically package data and root hash into transactions through smart contract trigger conditions ;

[0233] The transaction generates a block through the sorting service , sent to all peer nodes for verification and then written into the ledger to complete the evidence storage;

[0234] (3) Verification and query stage:

[0235] When regulators need to verify total carbon emissions compliance, the system calls the zero-knowledge proof unit to generate ,The verifier confirms the validity of the proof through a verification algorithm without having to check the carbon footprint data of the specific baggage;

[0236] When querying the carbon footprint of a single piece of luggage, the system returns , the corresponding Merkle tree path proof and block information, users confirm data integrity by verifying whether the path hash matches the root hash;

[0237] Automatic execution via smart contracts every hour Function, writes the newly generated Merkle tree root hash to the latest block to ensure that the ledger is synchronized with real-time carbon data;

[0238] The third-party auditing agency authorizes the call of the audit contract to obtain the carbon data Merkle tree, zero-knowledge proof, and block metadata for a specified time period, completes the offline audit, and generates a compliance report.

[0239] (V) Archiving stage:

[0240] When the baggage exceeds the tracing period (such as 30 days after the flight), the historical data is migrated from the active ledger to archive storage (such as the IPFS distributed file system) through a smart contract, and only the Merkle tree root hash and index information are retained in the ledger, freeing up storage space while ensuring that the archived data is traceable.

[0241] In this embodiment, the dynamic scheduling execution layer includes a policy modification module:

[0242] The deviation of the comprehensive optimization index of the system is calculated in real time, and its value is the absolute value of the difference between the twin prediction value and the actual value divided by the prediction value;

[0243] When the deviation exceeds the threshold, adjust the energy consumption weight coefficient α and the initial carbon sensitivity coefficient , where α is adjusted proportionally with the sign of the energy consumption deviation, Adjustments are made proportionally based on the carbon emission deviation rate;

[0244] Furthermore, the dynamic scheduling execution layer is the "execution center" of the airport baggage low-carbon scheduling system based on swarm intelligence and digital twins. It is responsible for converting the optimization strategies generated by the swarm intelligence decision-making layer into equipment control signals. It corrects the deviations between the physical system and the digital twin simulation through a real-time feedback mechanism to ensure the precise implementation of scheduling strategies under complex working conditions. It is the key execution link connecting intelligent decision-making and physical operations. The following is a detailed explanation from the overall perspective:

[0245] 1. Overview of overall functions:

[0246] The dynamic scheduling execution layer takes the scheduling strategies (such as conveyor speed adjustment values, robot path instructions, and chute allocation plans) output by the swarm intelligence decision-making layer as input, and forms a "decision-execution-correction" closed loop through command conversion, equipment control, and real-time feedback. Its core tasks are to convert abstract optimization goals (such as minimizing total energy consumption and complying with carbon emissions) into specific equipment motion parameters. It monitors the effectiveness of strategy execution through real-time data from the physical perception layer and calculates deviations from digital twin simulation results. When deviations exceed the limit, a parameter recalibration mechanism is triggered, dynamically adjusting control signals to ensure that the actual operating status matches the optimization goals. Furthermore, this layer has emergency response capabilities, rapidly executing redundant switching in the event of sudden equipment failures to ensure the continuity of the baggage handling process. It is the ultimate executor of the system from "virtual decision-making" to "physical implementation."

[0247] 2. Submodule composition and functions:

[0248] (1) Command conversion and equipment control unit:

[0249] Strategy parsing and signal generation: Receives the scheduling strategy of the swarm intelligence decision-making layer and parses it into control parameters that can be executed by the device; for example:

[0250] Conveyor belt speed instruction: The belt speed optimization value output by the decision layer Converted into motor control signals , where PWM is the duty cycle of the pulse width modulation signal, is the speed-voltage conversion coefficient, is the zero point compensation value, and the motor speed is adjusted by the variable frequency drive;

[0251] Robot path instructions: the path points planned by the decision layer Convert to joint angle sequence ,in, For the The joint angle corresponding to each path point is Solve the inverse kinematics function to control the robot arm to move along the planned path;

[0252] Multi-device collaborative control: For cross-regional baggage handling processes (such as from the sorting area to the transfer area), the timing of multi-device actions is coordinated through time synchronization protocols (such as IEEE1588) to ensure seamless baggage handover. For example, when a bag is about to reach the end of the conveyor belt, an opening instruction is sent to the target chute 500ms in advance. (in, is the instruction trigger time, The estimated arrival time of the baggage. (Due to device response delay) to avoid handover jams;

[0253] Control signal safety check: Perform safety threshold check on the generated control signal to ensure that the equipment moves within a safe range (such as the conveyor belt speed does not exceed the mechanical limit m / s, the robot joint angle does not exceed the limit ); using the formula (in, is the original signal output by the decision layer, is the safety threshold, Limiting function) prevents equipment overload and ensures safe operation;

[0254] (2) Real-time feedback and deviation calculation unit:

[0255] Execution effect monitoring: Real-time collection of equipment operating parameters and baggage status data through the physical perception layer, including: actual conveyor belt speed , motor real-time current 2. Actual luggage detention time , actual carbon emissions etc., as quantitative indicators of the effectiveness of strategy execution;

[0256] Simulation-actual deviation calculation: Compare monitoring data with the simulation prediction value of the digital twin engine layer and calculate the deviation index:

[0257] Optimization target deviation: (in, is the optimization target value of digital twin simulation (total energy consumption + maximum residence time + load imbalance), is the optimization target value of actual operation (total energy consumption + maximum residence time + load imbalance);

[0258] Deviation of single indicator: such as energy consumption deviation , carbon emission deviation rate ,provide segmentation basis for strategy revision;

[0259] Deviation warning mechanism: ( When the deviation threshold (usually set to 15%) or a single deviation exceeds the limit, an early warning signal is triggered, marked as "strategy execution abnormality", and the deviation data is packaged and sent to the strategy correction module;

[0260] (3) Strategy Modification Module:

[0261] Dynamic parameter calibration: Based on the deviation calculation results, the control parameters and decision weights are automatically adjusted. The formula is: ,in, is the energy consumption deviation adjustment factor (0.05-0.2), is a sign function (increases for positive deviations , decreases for negative deviations), is the initial carbon sensitivity coefficient of the carbon-sensitive ant colony algorithm, The carbon emission deviation rate; by calibrating the weights and algorithm parameters, the decision-making strategy is made more consistent with the characteristics of the physical system;

[0262] Control signal correction: for the deviation between the actual state of the equipment and the instruction (such as the actual speed of the conveyor belt ), using the proportional-integral-derivative (PID) control algorithm to correct the output signal in real time: ,in, is the speed deviation, 、 、 PID controller parameters, ensuring that the actual operating parameters of the equipment converge to the decision target value;

[0263] Strategy iteration optimization: When the deviation of 3 consecutive sampling cycles (each cycle is 100ms) When all the time limits are exceeded, the deviation data will be fed back to the swarm intelligent decision-making layer, triggering the strategy to regenerate the process, and generate a more suitable scheduling strategy based on the updated physical parameters (such as the actual energy consumption coefficient of the equipment), realizing continuous optimization of "feedback-iteration".

[0264] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. The digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI features: include: The physical sensing layer, consisting of RFID scanners, current sensors, and laser rangefinders deployed in the baggage handling area, collects real-time information on conveyor belt speed, robot joint angles, chute congestion rates, and 3D baggage coordinates. The digital twin engine layer builds a full-factor twin that includes equipment physical models, baggage flow network models, and carbon accounting models, and achieves synchronous mapping of physical systems through real-time data-driven implementation. The swarm intelligence decision-making layer is composed of intelligent agent modules distributed across various processing units, and uses an improved carbon-sensitive ant colony algorithm to generate scheduling strategies; The blockchain evidence layer, based on the Hyperledger architecture, stores device energy consumption and luggage carbon footprint data; The dynamic scheduling execution layer converts decision instructions into device control signals and dynamically modifies strategies based on the twin simulation results.

2. The digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI according to claim 1 is characterized by: The digital twin engine layer includes: Device-level mirror module: Establishing a conveyor belt power calculation model ,in, is the real-time power of the conveyor belt, are the air resistance and mechanical friction coefficient, is the load friction coefficient, is the conveyor belt speed, is the total mass of luggage on the conveyor belt; Carbon accounting module: calculate the carbon footprint of a single piece of luggage ,in, Carbon footprint of a single piece of luggage, The total number of equipment handling the baggage. For devices The average power, For devices How long it takes to process the baggage, is the carbon intensity factor of the power grid.

3. The digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI according to claim 1 is characterized by: The swarm intelligence decision layer adopts a global optimization objective function: ,in, 、 、 is an adjustable weight coefficient, and , is the total number of devices, For devices The total energy consumption points, For luggage The residence time, is the standard deviation function, For devices workload.

4. The digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI according to claim 3 is characterized by: The improved carbon-sensitive ant colony algorithm includes: Path selection probability calculation: ,in, For intelligent agents Select Path The probability of For path The pheromone concentration, is the pheromone importance factor, is the path heuristic value, is the path distance, For nodes The congestion level, is the heuristic information importance factor, is the carbon adaptation factor, is the carbon sensitivity coefficient, For path of carbon emissions, is the carbon emission baseline value of the pathway, is the dynamic carbon sensitivity coefficient, For nodes The set of adjacent nodes of For path pheromone concentration; Pheromone update rules: ,in, For path The pheromone concentration, is the pheromone volatility factor and 0< , is the total number of agents, For intelligent agents On the path The increase in pheromone released is the carbon learning rate factor, is the total amount constant of pheromone, is the optimal carbon emission threshold of the system.

5. The digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI according to claim 4 is characterized by: The dynamic carbon sensitivity coefficient is adjusted through a feedback mechanism: ,in, for The carbon sensitivity coefficient at the moment, To adjust the step size coefficient, for Average carbon intensity during the period, is the airport carbon emission target value, It is the tolerance range of carbon emissions.

6. The digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI according to claim 1 is characterized by: The blockchain evidence layer implements: Construct a Merkle tree to store carbon data, with its root node generated by the hash concatenation of all device energy consumption data and luggage carbon footprint data; Using zero-knowledge proof technology, it verifies whether the total carbon emissions of luggage meet the preset limit without leaking detailed data.

7. The digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI according to claim 1 is characterized by: The dynamic scheduling execution layer includes a strategy modification module: The deviation of the comprehensive optimization index of the system is calculated in real time, and its value is the absolute value of the difference between the twin prediction value and the actual value divided by the prediction value; When the deviation exceeds the threshold, adjust the energy consumption weight coefficient α and the initial carbon sensitivity coefficient , where α is adjusted proportionally with the sign of the energy consumption deviation, Adjustments will be made proportionally based on the carbon emission deviation rate.

8. The digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI according to claim 1 is characterized by: The swarm intelligence decision-making layer implements a collaborative optimization mechanism: Design an incentive function based on carbon credits: ,in, For intelligent agents of carbon credits, is the carbon emission reduction per unit time, is the time decay factor; The Byzantine fault-tolerant consensus algorithm is used, which requires that the decision vector of each agent obtains the consent of at least twice the number of faulty nodes plus one node among its neighboring nodes.

9. The digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI according to claim 1 is characterized by: The digital twin engine layer embeds a robustness optimization module: Construct a disturbance-resistant scheduling model with the goal of minimizing the expected value of the scheduling strategy under disturbance scenarios plus the risk factor multiplied by the 95% confidence level; Generates a Pareto optimal solution set consisting of strategies that satisfy the following conditions: there is no other strategy that is not inferior to this strategy in all optimization objectives and is strictly better in at least one objective.

10. The digital twin low-carbon scheduling system for airport baggage flow driven by swarm intelligence AI according to claim 4 is characterized by: The improved carbon-sensitive ant colony algorithm adopts a convergence guarantee mechanism: Define the potential game model, ,in, is the potential function, is the decision set of the agent, For devices Energy consumption, For luggage The residence time, 、 is the path node number, For path pheromone concentration; Prove that the cross-partial derivatives of the potential function with respect to any two agents’ decisions are non-negative; Set the convergence conditions: ,in, is the two-norm, for The pheromone concentration vector at time , for The pheromone concentration vector at time , is the convergence accuracy constant, is the convergence rate factor.

Images