A low-altitude airspace collaborative management method and system based on digital twinning

Abstract

The application provides a kind of low airspace collaborative management method and system based on digital twinning, comprising: based on the real-time state data of aircraft obtained, analyzing aircraft maneuverability, flight envelope characteristic parameters, characteristic parameters include maximum speed, maximum acceleration, minimum turning radius, utilize kinematics and dynamics model, depict different aircraft movement law;Combined with real-time flight data of aircraft, through digital twinning technology, real-time synchronization aircraft motion trajectory and attitude change in virtual environment, accurately reproduce aircraft dynamic behavior, provide high-fidelity digital model for flight conflict detection, flight path prediction;For aircraft motion characteristic difference, construct multi-level, multi-granularity digital twinning model, digital twinning model includes airspace structure model at macro level, aircraft group behavior model at mesoscopic level and single aircraft dynamic model at micro level.

Description

Technical Field

[0001] This invention belongs to the field of information technology, and in particular relates to a method and system for collaborative management of low-altitude airspace based on digital twins. Background Technology

[0002] The construction of digital twin models for low-altitude airspace faces numerous technical challenges. First, the low-altitude airspace environment is complex and ever-changing, with a wide variety of aircraft types; comprehensively sensing and accurately mapping the real-time status of these aircraft is a major challenge. Second, the dynamic characteristics of low-altitude aircraft vary greatly; accurately depicting and synchronizing their motion trajectories, attitude changes, and other dynamic behaviors in the digital twin model in real time presents another technical obstacle. Third, low-altitude airspace management involves many stakeholders; coordinating their demands, optimizing resource allocation, and improving management efficiency within the digital twin system is also a significant challenge. Furthermore, the high computational complexity and stringent real-time requirements of low-altitude digital twin systems necessitate achieving efficient parallel computing and meeting real-time response needs. Finally, the security and robustness of low-altitude digital twin systems are crucial; constructing a stable, reliable, and well-protected digital twin system presents researchers with yet another technical challenge. Summary of the Invention

[0003] This invention proposes a low-altitude airspace collaborative management method and system based on digital twins to solve the problems existing in the prior art.

[0004] To achieve the above objectives, this invention provides a low-altitude airspace collaborative management method based on digital twins, comprising the following steps:

[0005] Multimodal data of the aircraft is collected, and the multimodal data is processed by a Kalman filter data fusion algorithm to obtain the key parameters of the aircraft.

[0006] Based on the aircraft model, kinematic and dynamic models are constructed to process the key parameters and obtain the aircraft's maneuverability and flight envelope characteristic parameters;

[0007] Digital twin technology is used to synchronize key parameters, maneuver performance parameters, and flight envelope characteristic parameters of an aircraft into a virtual environment to obtain a digital model.

[0008] Based on the differences in the motion characteristics of the aircraft in different digital models, a multi-level digital twin model is constructed.

[0009] Particle swarm optimization algorithm is introduced into the multi-level digital twin model to comprehensively plan and dynamically schedule low-altitude resources;

[0010] Knowledge graph technology is introduced into the multi-level digital twin model to construct a virtual simulation environment for low-altitude airspace, and low-altitude airspace situational decision-making is carried out through semantic association of the knowledge graph.

[0011] Preferably, obtaining the key parameters of the aircraft includes:

[0012] Multi-source heterogeneous sensor networks are used for data acquisition to obtain multimodal data. The multimodal data is then cleaned and standardized to eliminate noise and outliers. Feature extraction is performed on the standardized multimodal data to obtain parameters that reflect the characteristics of different aircraft. These parameters are then input into a Kalman filter data fusion model to accurately perceive and map the real-time status of the aircraft. Based on the output of the Kalman filter data fusion model, key aircraft parameters are obtained.

[0013] Preferably, obtaining the maneuverability and flight envelope characteristic parameters of the aircraft includes:

[0014] A kinematic and dynamic model is constructed based on the aircraft model, which includes the aircraft's mass, thrust, and drag parameters. The key parameters of the aircraft are input into the kinematic and dynamic model, and the position, velocity, and attitude state variables of the aircraft at the next moment are predicted by numerical integration methods. Based on the aircraft's maneuverability parameters and prediction results, the flight envelope of the aircraft is determined, which includes characteristic parameters such as maximum velocity, maximum acceleration, and minimum turning radius.

[0015] Preferably, constructing a multi-level digital twin model includes:

[0016] By analyzing the movement trajectory and formation changes of each aircraft group, an aircraft group behavior model is constructed. Based on the aircraft group behavior model, an individual aircraft dynamic model is further established. The individual aircraft dynamic model comprehensively considers the dynamic characteristics of the aircraft and environmental factors, and provides a detailed description of the motion state of the individual aircraft. Above the individual aircraft dynamic model, a macroscopic airspace structure model is constructed. The airspace structure model describes the airspace environment in which the aircraft is located based on the airspace topography, meteorological conditions and control rules. The aircraft group behavior model, the individual aircraft dynamic model and the airspace structure model are integrated to form a multi-level digital twin model.

[0017] Preferably, the overall planning and dynamic scheduling of low-altitude resources includes:

[0018] The particle swarm optimization algorithm is used to coordinate and dynamically schedule low-altitude resources in a multi-level digital twin model, generating an optimal resource allocation scheme. During the iterative process of the particle swarm optimization algorithm, the demand weights of different stakeholders are introduced, and the interests of stakeholders are balanced by adjusting the weight coefficients. Feasibility analysis and risk assessment are performed on the generated low-altitude resource allocation scheme to determine whether the scheme meets the constraints. If not, iterative optimization is carried out until a feasible optimal scheme is obtained. The optimized low-altitude resource allocation scheme is distributed to relevant aircraft operators and air traffic control departments for flight mission planning and scheduling. At the same time, the key parameters in the scheme are fed back to the multi-level digital twin model in real time for dynamic updates and adjustments.

[0019] Preferably, low-altitude airspace situational decision-making through semantic associations of knowledge graphs includes:

[0020] Knowledge graph technology is introduced into a multi-level digital twin model to construct a knowledge base and inference engine for low-altitude airspace management. The knowledge graph contains multi-dimensional information on airspace structure, flight rules, and meteorological conditions. Through semantic association, it enables low-altitude airspace situational understanding and decision support. Flight plans and airspace applications are automatically reviewed and conflict detected according to the rules and constraints of the knowledge graph.

[0021] This invention also provides a low-altitude airspace collaborative management system based on digital twins, comprising:

[0022] The multi-source data fusion sensing module is used to collect multi-modal data of the aircraft and process the multi-modal data through a Kalman filter data fusion algorithm to obtain the key parameters of the aircraft.

[0023] The aircraft motion characteristics acquisition module is used to construct kinematic and dynamic models based on the aircraft model, process the key parameters, and obtain the aircraft's maneuverability and flight envelope characteristic parameters;

[0024] The aircraft digital twin synchronization module is used to synchronize the aircraft's key parameters, maneuver performance parameters, and flight envelope characteristic parameters into a virtual environment through digital twin technology to obtain a digital model;

[0025] The multi-level digital twin model building module is used to build multi-level digital twin models based on the differences in the motion characteristics of aircraft in different digital models.

[0026] The low-altitude resource management module based on particle swarm optimization is used to introduce the particle swarm optimization algorithm into the multi-level digital twin model to carry out overall planning and dynamic scheduling of low-altitude resources.

[0027] The knowledge graph-based intelligent airspace management module is used to introduce knowledge graph technology into the multi-level digital twin model, construct a virtual simulation environment for low-altitude airspace, and make low-altitude airspace situation decisions through semantic associations of the knowledge graph.

[0028] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method.

[0029] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method.

[0030] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the method.

[0031] Compared with the prior art, the present invention has the following advantages and technical effects:

[0032] This invention discloses a method and system for collaborative management of low-altitude airspace based on digital twins. Addressing the complex and variable nature of the low-altitude environment and the diverse types of aircraft, a multi-source heterogeneous sensor network is deployed to collect multimodal data. Kalman filtering is used for data fusion to achieve accurate real-time perception of aircraft status. Based on the acquired aircraft status data, its maneuverability and flight envelope characteristic parameters are analyzed to construct a motion model. Through digital twin technology, the aircraft's trajectory and attitude changes are synchronized in real-time in a virtual environment, constructing a multi-level, multi-granular digital twin model. A particle swarm optimization algorithm is introduced into the digital twin model for comprehensive planning and dynamic scheduling of low-altitude resources. Simultaneously, knowledge graph technology is incorporated to construct a low-altitude airspace management knowledge base and inference engine, enabling situational understanding and decision support. This invention, through the organic combination of multi-source data fusion, digital twin modeling, and knowledge graph inference technologies, achieves accurate perception, intelligent planning, and collaborative management of low-altitude airspace, improving the efficiency of low-altitude resource utilization and management level. Attached Figure Description

[0033] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0034] Figure 1 This is a flowchart of a method according to an embodiment of the present invention. Detailed Implementation

[0035] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0036] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0037] Example 1

[0038] like Figure 1 As shown, this embodiment provides a low-altitude airspace collaborative management method based on digital twins, including the following steps:

[0039] Multimodal data of the aircraft is collected, and the multimodal data is processed by a Kalman filter data fusion algorithm to obtain the key parameters of the aircraft.

[0040] Based on the aircraft model, kinematic and dynamic models are constructed, key parameters are processed, and the aircraft's maneuverability and flight envelope characteristic parameters are obtained.

[0041] Digital twin technology is used to synchronize key parameters, maneuver performance parameters, and flight envelope characteristic parameters of an aircraft into a virtual environment to obtain a digital model.

[0042] Based on the differences in the motion characteristics of the aircraft in different digital models, a multi-level digital twin model is constructed.

[0043] Particle swarm optimization algorithm is introduced into a multi-level digital twin model to comprehensively plan and dynamically schedule low-altitude resources;

[0044] Knowledge graph technology is introduced into a multi-level digital twin model to construct a virtual simulation environment for low-altitude airspace, and low-altitude airspace situational decision-making is carried out through semantic association of knowledge graph.

[0045] Specifically, the following steps are included:

[0046] S101. Addressing the complex and ever-changing nature of the low-altitude airspace environment and the diverse types of aircraft, a multi-source heterogeneous sensor network is deployed to collect radar, ADS-B, and BeiDou multimodal data. Utilizing a Kalman filter data fusion algorithm, the real-time status of the aircraft is accurately perceived and mapped, acquiring key parameters such as position, speed, and heading, thus constructing a comprehensive and accurate low-altitude situational awareness capability.

[0047] Specifically, given the complex and ever-changing nature of the low-altitude airspace environment, a multi-source heterogeneous sensor network is employed for data acquisition, obtaining multimodal data from radar, ADS-B, and BeiDou. For the acquired multimodal data, preprocessing techniques are used for data cleaning and standardization to eliminate noise and outliers, improving data quality. Considering the diverse types of aircraft, feature extraction is performed on the preprocessed multimodal data to obtain key parameters reflecting the characteristics of different aircraft. These extracted key aircraft parameters are input into a Kalman filter data fusion model. Through model training and optimization, accurate perception and mapping of the aircraft's real-time status are achieved. Based on the output of the Kalman filter data fusion model, key parameters such as the aircraft's position, velocity, and heading are obtained, forming a comprehensive description of the aircraft's status. The acquired aircraft status information is visualized, generating an intuitive and easy-to-understand situational awareness map to provide decision support for relevant personnel. Continuous monitoring of changes in the low-altitude airspace environment and aircraft status is conducted, dynamically adjusting the sensor network configuration and data fusion algorithm parameters according to the changes to maintain the real-time performance and accuracy of situational awareness.

[0048] In this embodiment, due to the complex and variable low-altitude airspace environment, a multi-source heterogeneous sensor network is required for data acquisition. For example, radar, ADS-B receivers, and BeiDou satellite receivers can be deployed to form a comprehensive sensing network. Radar can detect the position and speed of the aircraft, ADS-B can receive the aircraft's actively broadcast identity and position information, and the BeiDou system can provide a precise spatiotemporal reference. This multi-source data acquisition method can complement each other's advantages and disadvantages, improving the comprehensiveness and reliability of the sensing. Preprocessing the acquired multimodal data is a key step in improving data quality. Median filtering can be used to remove sudden noise, Kalman filtering can be used to smooth trajectory data, and interpolation algorithms can be used to repair missing data. For example, if the ADS-B position data of a UAV shows a sudden jump, interpolation correction can be performed using data from before and after the jump to make the trajectory smoother and more reasonable. Feature extraction for different types of aircraft is the foundation for accurate identification. Dynamic features such as speed, acceleration, turning radius, and climb rate, as well as electromagnetic features such as radar cross section and ADS-B message format, can be extracted. For example, fixed-wing aircraft and multi-rotor UAVs exhibit significant differences in their turning characteristics, which can serve as an important basis for differentiation. Kalman filtering is a classic data fusion algorithm that can effectively fuse observation data from different sensors. By establishing a kinematic model of the aircraft and combining it with the observation equations, an optimal estimate of the aircraft's state can be obtained. For instance, radar observation data for a UAV shows its position as (x1, y1, z1), while ADS-B data shows its position as (x2, y2, z2). Kalman filtering can provide a more reliable position estimate (x, y, z) based on the uncertainties between the two sets of data.

[0049] S102. Based on the acquired real-time status data of the aircraft, analyze the aircraft's maneuverability and flight envelope characteristic parameters, including maximum speed, maximum acceleration, and minimum turning radius. Utilize kinematic and dynamic models to characterize the motion patterns of different aircraft.

[0050] Specifically, based on the acquired key parameters, the instantaneous velocity, acceleration, and turning radius of the aircraft are calculated and compared with preset thresholds to determine whether the aircraft's maneuverability is normal. For different aircraft models, corresponding kinematic and dynamic models are established, including parameters such as the aircraft's mass, thrust, and drag, to describe the aircraft's motion. Real-time state data of the aircraft is input into the kinematic and dynamic models, and state variables such as position, velocity, and attitude at the next moment are calculated using methods such as numerical integration to predict the aircraft's motion state. Based on the aircraft's maneuverability parameters and the prediction results of the motion model, the aircraft's flight envelope is determined, including characteristic parameters such as maximum velocity, maximum acceleration, and minimum turning radius, to evaluate the aircraft's maneuverability and flight performance.

[0051] Taking a fixed-wing aircraft as an example, its kinematic model includes state variables such as position, velocity, and attitude, while its dynamic model involves forces such as thrust, lift, and drag. By substituting real-time state data into the model, the aircraft's state at the next moment can be predicted using numerical integration methods. This prediction helps to identify potential risks in advance. Determining the flight envelope allows for a comprehensive assessment of aircraft performance. Taking a certain type of fighter jet as an example, its flight envelope includes characteristic parameters such as a maximum speed of Mach 1.8, a maximum overload of 9G, and a minimum turning radius of 300 meters. These parameters reflect both the aircraft's extreme performance and provide boundary conditions for safe flight. By monitoring the relationship between flight status and the flight envelope in real time, situations exceeding the safe range can be detected promptly.

[0052] S103. Combining the real-time flight data of the aircraft, the aircraft's motion trajectory and attitude changes are synchronized in real time in a virtual environment through digital twin technology, accurately reproducing the dynamic behavior of the aircraft, and providing a high-fidelity digital model for flight conflict detection and trajectory prediction.

[0053] Specifically, a digital twin model of the aircraft is constructed in a virtual environment. This model maintains consistency with the physical and dynamic characteristics of the real aircraft, accurately reproducing its trajectory and attitude changes. Based on real-time flight data, the digital twin model is used to calculate and update the aircraft's position and attitude in the virtual environment in real time, achieving real-time synchronization of the aircraft's dynamic behavior. Based on the digital twin model, machine learning algorithms are used to predict the aircraft's future trajectory, and the prediction results are used to determine if there are potential flight conflict risks. If potential flight conflict risks are detected, different flight strategies and obstacle avoidance schemes are simulated using the digital twin model to evaluate the safety and feasibility of each scheme, and the optimal scheme is selected as the aircraft's trajectory adjustment strategy. The optimized trajectory adjustment strategy is transmitted to the real aircraft to guide it in executing corresponding flight maneuvers and avoiding actual flight conflicts. Throughout the process, the real-time flight data of the aircraft is continuously monitored, and the aircraft's motion state is constantly updated through real-time synchronization between the digital twin model and the real aircraft, ensuring a high degree of consistency between the digital twin model and the real aircraft.

[0054] In this embodiment, the collected data is transmitted to a virtual environment to construct a digital twin model. This model needs to accurately replicate the physical and dynamic characteristics of the aircraft. Taking a fixed-wing UAV as an example, its digital twin model includes geometric parameters such as wingspan, fuselage length, and center of gravity, as well as dynamic parameters such as engine thrust and lift coefficient. Through these parameters, the model can accurately simulate the aircraft's motion behavior. Based on real-time data, the digital twin model continuously updates the aircraft's state in the virtual environment. For example, when the real UAV flies north at a speed of 5 m / s, the virtual model will also synchronously update its position and attitude to maintain consistency with the actual aircraft. This real-time synchronization lays the foundation for subsequent trajectory prediction and collision detection. Machine learning algorithms, such as Long Short-Term Memory (LSTM) networks, are used to predict the future trajectory of the aircraft. The algorithm learns the aircraft's motion patterns by analyzing historical flight data, thereby predicting possible trajectories within a certain period. For example, the prediction system might detect a potential collision risk between the UAV and a nearby tall building within the next 5 minutes. Once a potential risk is detected, the system will use the digital twin model to simulate various obstacle avoidance schemes. For example, different strategies such as yawing 30 degrees to the left, yawing 30 degrees to the right, or climbing 100 meters can be simulated to evaluate the safety and feasibility of each option. Through simulation calculations, the system may find that yawing 30 degrees to the right is the optimal choice, effectively avoiding obstacles while maintaining the original flight path. After determining the optimal solution, the system transmits the trajectory adjustment command to the real aircraft. Upon receiving the command, the flight control system executes the corresponding flight maneuvers, such as changing the yaw angle or adjusting the altitude, thereby avoiding actual conflicts. Throughout the process, the digital twin model continuously synchronizes with the real aircraft, constantly updating the flight status and providing accurate basic data for the next round of prediction and decision-making. This digital twin-based flight conflict detection and obstacle avoidance system has significant advantages. It can rehearse multiple possible scenarios in a virtual environment without conducting high-risk actual flight tests. At the same time, through continuous learning and optimization, the system can continuously improve prediction accuracy and decision-making efficiency, providing strong protection for the safe operation of the aircraft.

[0055] Furthermore, based on the digital twin model, machine learning algorithms are used to predict the future trajectory of the aircraft, and the prediction results are used to determine whether there is a potential risk of flight conflict.

[0056] Based on historical flight data and real-time data collected by sensors, a digital twin model of the aircraft is constructed, incorporating information such as its physical characteristics, dynamics, and environmental factors. A Long Short-Term Memory (LSTM) neural network algorithm is used to train the historical trajectory data, establishing a trajectory prediction model. The model's input includes the aircraft's current state and environmental information, and its output is the predicted trajectory over a future period. A large number of aircraft trajectory samples are generated using Monte Carlo simulation, encompassing the aircraft's motion under different environmental conditions. This sample data is used to further train and optimize the LSTM prediction model, improving its prediction accuracy and generalization ability. Information from other aircraft is incorporated into the digital twin model. A multi-agent reinforcement learning algorithm simulates the interaction of multiple aircraft in the same airspace, analyzing parameters such as distance, speed, and direction between aircraft to determine potential conflict risks. When the distance between the predicted trajectory and the trajectories of other aircraft is less than a safety threshold, a conflict risk alarm is triggered. Appropriate measures, such as trajectory adjustment and speed control, are taken based on the risk level to ensure the aircraft's operational safety. The digital twin model continuously updates the aircraft's state and environmental information in real time. By comparing it with sensor data from the actual aircraft, the digital twin model is continuously optimized and corrected, improving its simulation accuracy and predictive capabilities. Based on the aircraft's mission requirements and risk assessment results, a genetic algorithm is used to optimize the aircraft's trajectory, generating a safe and efficient flight path. The optimized trajectory parameters are then transmitted to the flight control system to guide the aircraft's actual movement.

[0057] S104. To address the differences in the motion characteristics of the aircraft, a multi-level, multi-granularity digital twin model is constructed. The digital twin model includes a macroscopic airspace structure model, a mesoscopic group behavior model of the aircraft, and a microscopic individual dynamic model of the aircraft.

[0058] Specifically, based on the acquired aircraft motion characteristic data, clustering algorithms are used to classify the aircraft, resulting in groups of aircraft with similar motion characteristics. For each aircraft group, a group behavior model is constructed by analyzing its trajectory, formation changes, etc., to characterize the overall motion characteristics of the group. Building upon the group behavior model, a dynamic model of each individual aircraft is further established. This model comprehensively considers the aircraft's dynamic characteristics and environmental factors, providing a refined description of the motion state of each individual aircraft. Above the individual aircraft dynamic model, a macroscopic airspace structure model is constructed. This model characterizes the airspace environment in which the aircraft resides based on airspace topography, meteorological conditions, and control rules. The group behavior model, the individual aircraft dynamic model, and the airspace structure model are then integrated to form a multi-level, multi-granularity digital twin model of the aircraft. This model comprehensively reflects the motion characteristics of the aircraft at different levels. Machine learning algorithms are used to train and optimize the digital twin model, improving its accuracy and predictive ability. Recurrent neural networks from deep learning can be used to learn the temporal characteristics of aircraft motion; reinforcement learning algorithms can be used to achieve autonomous optimization of aircraft motion strategies; and transfer learning methods can be used to accelerate the construction of new aircraft models by utilizing existing aircraft model knowledge.

[0059] In this embodiment, clustering algorithms can effectively identify groups of aircraft with similar motion characteristics. For example, using the K-means clustering method, aircraft can be classified according to speed and altitude characteristics, potentially resulting in several typical groups such as "low-altitude slow speed," "medium-high altitude cruise," and "high-altitude fast speed." This classification helps in the subsequent targeted development of group behavior models. The construction of group behavior models requires analyzing the overall motion characteristics of the aircraft. Taking military formation flight as an example, by observing the relative position changes of multiple fighter jets, typical formations such as "wedges" and "echelons" can be identified, and mathematical models describing the formation transformation rules can be established. Such models can predict the future motion trend of the entire formation. Individual aircraft dynamic models are more refined and require consideration of the aircraft's dynamic characteristics. For example, when modeling a helicopter, its unique vertical takeoff and landing capabilities and hovering characteristics must be considered. The model may include parameters such as rotor lift and tail rotor anti-torque to accurately describe the helicopter's motion performance under various flight conditions. Airspace structure models provide a macroscopic description. Taking the airspace surrounding an airport as an example, the model needs to include information such as runway direction, approach routes, and no-fly zones. Simultaneously, local weather conditions, such as wind direction, wind speed, and visibility, must be considered, as these factors all affect the actual trajectory of the aircraft. The fusion of multi-level models is a complex process. Taking a drone swarm as an example, a dynamic model of a single drone can be established first, then a swarm model can be built based on group behavior rules (such as collision avoidance algorithms), and finally, it can be placed in a specific airspace environment to form a complete digital twin model. This multi-granularity model can simultaneously reflect individual drone characteristics and swarm effects. Machine learning algorithms play a crucial role in model optimization. For example, Long Short-Term Memory (LSTM) networks can effectively capture the temporal characteristics of aircraft motion and predict their future trajectories. Reinforcement learning algorithms such as Q-learning can be used to optimize flight strategies, such as finding the optimal route under complex weather conditions. Transfer learning can leverage existing passenger aircraft model knowledge to quickly build digital twin models of new cargo aircraft, greatly improving modeling efficiency. This multi-level, multi-method digital twin model construction approach can comprehensively and accurately describe the motion characteristics of aircraft, providing strong support for fields such as flight safety management and air traffic control.

[0060] S105. The optimized aircraft motion model is applied to low-altitude resource management. Particle swarm optimization algorithm is introduced into the digital twin model to plan and dynamically schedule low-altitude resources, balance the needs of different stakeholders, improve the efficiency and coordination of low-altitude airspace management, and achieve optimal allocation of low-altitude resources.

[0061] Specifically, based on predicted aircraft trajectories and state parameters, combined with real-time low-altitude resource occupancy, a particle swarm optimization (PSO) algorithm is used in a digital twin model to comprehensively plan and dynamically schedule low-altitude resources, generating an optimal resource allocation scheme. During the iterative process of the PSO algorithm, demand weights from different stakeholders, such as aircraft operators, air traffic control departments, and ground facilities, are introduced. By adjusting the weight coefficients, the interests of each party are balanced, ensuring that the generated resource allocation scheme is both fair and efficient. Feasibility analysis and risk assessment are performed on the generated low-altitude resource allocation scheme to determine whether it meets constraints such as safety and reliability. If not, the process is returned, and the optimization iteration continues until a feasible optimal scheme is obtained. The optimized low-altitude resource allocation scheme is then distributed to relevant aircraft operators and air traffic control departments to guide their flight mission planning and scheduling. Simultaneously, key parameters in the scheme are fed back to the digital twin model in real time for dynamic updates and adjustments. In the process of low-altitude airspace resource management, continuous monitoring of the real-time status of aircraft and the occupancy of low-altitude airspace resources is crucial. Machine learning algorithms are used to analyze and predict data, identify potential conflict risks and resource bottlenecks, and adjust and optimize strategies in a timely manner to improve the real-time nature and accuracy of management. Establishing a collaborative mechanism and data-sharing platform for low-altitude airspace management promotes information exchange and collaborative decision-making among different stakeholders, improves the management efficiency and resource utilization of low-altitude airspace, and achieves optimal allocation and sustainable development of low-altitude resources.

[0062] In this embodiment, the particle swarm optimization algorithm is applied in this scenario. Each particle can represent a potential resource allocation scheme, and its position and velocity correspond to the trajectory and state parameters of the aircraft. For example, a particle might represent a scheme such as "UAV A flies along a specific path at an altitude of 500 meters and a speed of 60 km / h". The algorithm searches for the optimal resource allocation scheme by iteratively adjusting the position and velocity of the particles. The design of the fitness function is crucial to the success of the algorithm. It can comprehensively consider multiple factors, such as flight safety, resource utilization efficiency, and task completion time. For example, a simple fitness function might be: f = w1 * safety + w2 * efficiency + w3 * time, where w1, w2, and w3 are weight coefficients. Safety can be measured by the minimum distance between the aircraft and obstacles, efficiency can be represented by airspace utilization, and time is directly taken as the task completion time. The convergence process of the algorithm reflects the process of scheme optimization. Initially, particles may be distributed throughout the solution space, representing various possible resource allocation schemes. As iterations proceed, particles gradually move towards the optimal solution. For example, in urban delivery scenarios, the initial plan might include dangerous routes navigating between tall buildings. However, after optimization, these routes can be adjusted to safer and more efficient flight paths. The final optimal solution needs to be applied in actual management. For instance, for an urban logistics delivery network, the optimized plan might include multiple drone flight paths, each carefully designed to avoid high-risk areas and maximize delivery efficiency. The ground control center can dynamically adjust the flight plans of each drone based on this plan, ensuring the coordinated operation of the entire network. This low-altitude resource management method based on particle swarm optimization not only improves resource utilization efficiency but also significantly enhances flight safety. Through real-time adjustment and optimization, it can flexibly respond to complex and ever-changing low-altitude environments, laying the foundation for future smart city air traffic management.

[0063] S106. In the application of the digital twin model, knowledge graph technology is introduced to construct a low-altitude airspace management knowledge base and inference engine. The knowledge graph contains multi-dimensional information such as airspace structure, flight rules, and meteorological conditions, and achieves low-altitude airspace situational understanding and decision support through semantic association. Based on the rules and constraints of the knowledge graph, flight plans and airspace applications are automatically reviewed and conflict detected, improving the level of intelligent management.

[0064] Specifically, based on multi-dimensional information such as the airspace structure, flight rules, and meteorological conditions of low-altitude airspace, a knowledge graph for low-altitude airspace management is constructed, establishing entities, attributes, and relationships to form a semantic network. Through a knowledge graph reasoning engine, the semantic association information within the graph, combined with real-time acquired airspace situation data, is used to understand and analyze the low-altitude airspace situation, providing support for management decisions. Based on the airspace management rules and constraints defined in the knowledge graph, submitted flight plans and airspace applications are automatically reviewed to determine compliance with relevant regulations and requirements. During the review process, the semantic association and reasoning capabilities of the knowledge graph are used to detect conflicts and incompatibilities between different plans and applications, ensuring the safety and rationality of airspace use. Digital twin technology is combined with the knowledge graph to construct a virtual simulation environment for low-altitude airspace. Through simulation and analysis of the virtual airspace, airspace management strategies and decision-making schemes are optimized. Machine learning algorithms, such as decision trees, support vector machines, or neural networks, are used to train an airspace management decision-making model based on the semantic association features of the knowledge graph, achieving intelligent management decision support. Through the visualization of knowledge graphs, managers can be provided with an intuitive presentation of airspace situation and correlation analysis, which can assist in decision-making and management optimization in complex airspace environments.

[0065] In this embodiment, the low-altitude airspace management knowledge graph is a semantic network structure used to represent airspace entities, attributes, and relationships. For example, airspace can be divided into different blocks, each containing attributes such as altitude, latitude and longitude range, and adjacency relationships with surrounding blocks. Flight rules can be represented as constraints between entities, such as no-fly or restricted-fly zones. Meteorological conditions can be associated with airspace blocks as dynamic attributes. The knowledge graph inference engine can perform complex reasoning using semantic relationships within the graph. For example, when severe convective weather is detected in a certain area, the inference engine can automatically infer the affected flight routes and flight plans and provide adjustment suggestions. This intelligent analysis greatly improves the efficiency and accuracy of airspace management. The knowledge graph plays a crucial role in flight plan review. For instance, if a drone applies to fly over a city, the system will automatically check the area's flight restriction rules, building heights, and other information to determine if it meets safety requirements. If potential risks exist, the system will provide specific reasons and alternative solutions. Conflict detection is a key aspect of airspace management. Through the association analysis of the knowledge graph, the system can quickly identify spatiotemporally overlapping flight plans. For example, when the expected routes of two aircraft approach each other at a certain point in time, the system automatically marks potential conflict points and provides avoidance suggestions based on factors such as flight priority. The combination of digital twin technology and knowledge graphs can create highly realistic virtual airspace environments. In this environment, managers can simulate different airspace allocation schemes and assess their impact on indicators such as traffic flow and safety. This visual simulation greatly enhances the scientific rigor and foresight of decision-making. Machine learning algorithms have broad application prospects in airspace management. For example, a neural network model can be trained using historical data to predict airspace congestion at a future time. This model can consider various factors, such as weather and holidays, providing a basis for the dynamic allocation of airspace resources. The visualization capabilities of knowledge graphs provide managers with intuitive decision support tools. For example, in handling emergencies, the system can quickly present key information such as the airspace structure, aircraft distribution, and alternative routes in the affected area, helping managers to quickly formulate response strategies. This visual analysis greatly improves the efficiency and accuracy of decision-making in complex situations.

[0066] This embodiment also provides a low-altitude airspace collaborative management system based on digital twins, including:

[0067] The multi-source data fusion sensing module is used to collect multi-modal data of the aircraft and process the multi-modal data through a Kalman filter data fusion algorithm to obtain the key parameters of the aircraft.

[0068] The aircraft motion characteristics acquisition module is used to construct kinematic and dynamic models based on the aircraft model, process the key parameters, and obtain the aircraft's maneuverability and flight envelope characteristic parameters;

[0069] The aircraft digital twin synchronization module is used to synchronize the aircraft's key parameters, maneuver performance parameters, and flight envelope characteristic parameters into a virtual environment through digital twin technology to obtain a digital model;

[0070] The multi-level digital twin model building module is used to build multi-level digital twin models based on the differences in the motion characteristics of aircraft in different digital models.

[0071] The low-altitude resource management module based on particle swarm optimization is used to introduce the particle swarm optimization algorithm into the multi-level digital twin model to carry out overall planning and dynamic scheduling of low-altitude resources.

[0072] The knowledge graph-based intelligent airspace management module is used to introduce knowledge graph technology into the multi-level digital twin model, construct a virtual simulation environment for low-altitude airspace, and make low-altitude airspace situation decisions through semantic associations of the knowledge graph.

[0073] This embodiment also provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method.

[0074] This embodiment also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method.

[0075] This embodiment also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the method.

[0076] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A low-altitude airspace collaborative management method based on digital twins, characterized in that, Includes the following steps: Multimodal data of the aircraft is collected, and the multimodal data is processed by a Kalman filter data fusion algorithm to obtain the key parameters of the aircraft. Based on the aircraft model, kinematic and dynamic models are constructed to process the key parameters and obtain the aircraft's maneuverability and flight envelope characteristic parameters; Digital twin technology is used to synchronize key parameters, maneuver performance parameters, and flight envelope characteristic parameters of an aircraft into a virtual environment to obtain a digital model. Based on the differences in the motion characteristics of the aircraft in different digital models, a multi-level digital twin model is constructed. Building a multi-level digital twin model includes: By analyzing the movement trajectory and formation changes of each aircraft group, an aircraft group behavior model is constructed. Based on the aircraft group behavior model, an individual aircraft dynamic model is further established. The individual aircraft dynamic model comprehensively considers the dynamic characteristics of the aircraft and environmental factors, and provides a detailed description of the motion state of the individual aircraft. Above the individual aircraft dynamic model, a macroscopic airspace structure model is constructed. The airspace structure model describes the airspace environment in which the aircraft is located based on the airspace topography, meteorological conditions, and control rules. The aircraft group behavior model, the individual aircraft dynamic model, and the airspace structure model are integrated to form a multi-level digital twin model. Particle swarm optimization algorithm is introduced into the multi-level digital twin model to comprehensively plan and dynamically schedule low-altitude resources; Knowledge graph technology is introduced into the multi-level digital twin model to construct a virtual simulation environment for low-altitude airspace, and low-altitude airspace situational decision-making is carried out through semantic association of the knowledge graph.

2. The method according to claim 1, characterized in that, Key parameters for obtaining the aircraft include: Multi-source heterogeneous sensor networks are used for data acquisition to obtain multimodal data. The multimodal data is then cleaned and standardized to eliminate noise and outliers. Feature extraction is performed on the standardized multimodal data to obtain parameters that reflect the characteristics of different aircraft. These parameters are then input into a Kalman filter data fusion model to perceive and map the real-time status of the aircraft. Based on the output of the Kalman filter data fusion model, key aircraft parameters are obtained.

3. The method according to claim 1, characterized in that, The parameters for obtaining the aircraft's maneuverability and flight envelope characteristics include: A kinematic and dynamic model is constructed based on the aircraft model, which includes the aircraft's mass, thrust, and drag parameters. The key parameters of the aircraft are input into the kinematic and dynamic model, and the position, velocity, and attitude state variables of the aircraft at the next moment are predicted by numerical integration methods. Based on the aircraft's maneuverability parameters and prediction results, the flight envelope of the aircraft is determined, which includes characteristic parameters such as maximum velocity, maximum acceleration, and minimum turning radius.

4. The method according to claim 1, characterized in that, The overall planning and dynamic scheduling of low-altitude resources includes: The particle swarm optimization algorithm is used to coordinate and dynamically schedule low-altitude resources in a multi-level digital twin model, generating an optimal resource allocation scheme. During the iterative process of the particle swarm optimization algorithm, the demand weights of different stakeholders are introduced, and the interests of stakeholders are balanced by adjusting the weight coefficients. Feasibility analysis and risk assessment are performed on the generated low-altitude resource allocation scheme to determine whether the scheme meets the constraints. If not, iterative optimization is carried out until a feasible optimal scheme is obtained. The optimized low-altitude resource allocation scheme is distributed to relevant aircraft operators and air traffic control departments for flight mission planning and scheduling. At the same time, the key parameters in the scheme are fed back to the multi-level digital twin model in real time for dynamic updates and adjustments.

5. The method according to claim 1, characterized in that, Low-altitude airspace situational decision-making through semantic associations in knowledge graphs includes: Knowledge graph technology is introduced into a multi-level digital twin model to construct a knowledge base and inference engine for low-altitude airspace management. The knowledge graph contains multi-dimensional information on airspace structure, flight rules, and meteorological conditions. Through semantic association, it enables low-altitude airspace situational understanding and decision support. Flight plans and airspace applications are automatically reviewed and conflict detected according to the rules and constraints of the knowledge graph.

6. A low-altitude airspace collaborative management system based on digital twins, characterized in that, include: The multi-source data fusion sensing module is used to collect multi-modal data of the aircraft and process the multi-modal data through a Kalman filter data fusion algorithm to obtain the key parameters of the aircraft. The aircraft motion characteristics acquisition module is used to construct kinematic and dynamic models based on the aircraft model, process the key parameters, and obtain the aircraft's maneuverability and flight envelope characteristic parameters; The aircraft digital twin synchronization module is used to synchronize the aircraft's key parameters, maneuver performance parameters, and flight envelope characteristic parameters into a virtual environment through digital twin technology to obtain a digital model; The multi-level digital twin model building module is used to build multi-level digital twin models based on the differences in the motion characteristics of aircraft in different digital models. Building a multi-level digital twin model includes: By analyzing the movement trajectory and formation changes of each aircraft group, an aircraft group behavior model is constructed. Based on the aircraft group behavior model, an individual aircraft dynamic model is further established. The individual aircraft dynamic model comprehensively considers the dynamic characteristics of the aircraft and environmental factors, and provides a detailed description of the motion state of the individual aircraft. Above the individual aircraft dynamic model, a macroscopic airspace structure model is constructed. The airspace structure model describes the airspace environment in which the aircraft is located based on the airspace topography, meteorological conditions, and control rules. The aircraft group behavior model, the individual aircraft dynamic model, and the airspace structure model are integrated to form a multi-level digital twin model. The low-altitude resource management module based on particle swarm optimization is used to introduce the particle swarm optimization algorithm into the multi-level digital twin model to carry out overall planning and dynamic scheduling of low-altitude resources. The knowledge graph-based intelligent airspace management module is used to introduce knowledge graph technology into the multi-level digital twin model, construct a virtual simulation environment for low-altitude airspace, and make low-altitude airspace situation decisions through semantic associations of the knowledge graph.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method according to any one of claims 1-5.

9. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the method according to any one of claims 1-5.

Images