An airspace intelligent fusion management and control method and system based on a dynamic three-dimensional grid

Abstract

This invention discloses a method and system for intelligent airspace fusion management and control based on a dynamic three-dimensional grid. It relates to the field of low-altitude intelligent management and control technology. The method includes: Step 1: Acquiring real-time environmental data and multi-source sensing data and transmitting them to a ground control center; Step 2: The ground control center divides the low-altitude airspace into a variable-granularity grid based on the real-time environmental data and assigns a dynamic management and control level to each grid; Step 3: Synchronously collecting image data and spatial location data of UAVs within the grid based on multi-source sensing data, and performing positioning and precise identification of the UAVs; Step 4: Generating intelligent management and control decisions based on the positioning and precise identification results and the dynamic management and control level; Step 5: Activating corresponding management and control measures based on the intelligent management and control decision results. This invention, through intelligent decision-making algorithms and the delineation of dynamic safe flight zones, can effectively coordinate flight conflicts between UAVs, ensure low-altitude flight safety, and promote the healthy development of the low-altitude economy.

Description

Technical Field

[0001] This invention relates to the field of low-altitude intelligent control technology, and more specifically to a method and system for intelligent airspace fusion control based on dynamic three-dimensional grids. Background Technology

[0002] With the rapid iteration of drone technology and the continuous expansion of its application scenarios, drones have been widely penetrated into various fields such as film and television shooting, agricultural plant protection, power line inspection, and emergency rescue, leading to increasingly frequent low-altitude flight activities. However, low-altitude airspace resources are limited, and the surge in the number of drones and the diversification of their flight behaviors have resulted in a significant increase in safety risks such as airspace conflicts, illegal flights, and collisions, creating an urgent need for standardized and intelligent management and control of low-altitude airspace.

[0003] Existing low-altitude airspace control methods have several limitations: First, grid division often uses a fixed granularity model, which cannot be dynamically adjusted according to airspace flight density, regional importance parameters, and meteorological conditions, resulting in insufficient control precision in key areas and waste of resources in ordinary areas. Second, UAV identification relies on single-modal data or simple fusion technology, which is greatly affected by environmental factors such as lighting and weather, and the accuracy of positioning and type identification is difficult to meet the requirements of refined control. Third, safe flight zones are mostly static circular or square areas, without taking into account dynamic parameters such as UAV flight speed, fuselage weight, and inertia, and cannot adapt to the real-time flight status of UAVs, resulting in insufficient targeted protection. Fourth, control decisions are mostly based on single rule triggers, without considering the game of flight conflicts between multiple UAVs, and the avoidance strategy lacks global optimality, making it difficult to balance individual flight needs with overall airspace safety. Fifth, the hierarchical control command system is imperfect, and the response mechanism is rigid, making it impossible to flexibly adjust the control intensity according to the compliance status of UAVs and the degree of conflict, resulting in low control efficiency or excessive intervention.

[0004] Furthermore, existing systems have weak capabilities in integrating multi-source sensing data, making it difficult to efficiently fuse data from infrared, visible light, and lidar to support accurate identification. Simultaneously, priority management of drones in special scenarios such as emergency rescue is lacking, easily leading to flight conflicts with ordinary drones. These problems make existing control methods ill-suited to the complex and ever-changing low-altitude flight environment, failing to achieve efficient utilization of airspace resources and comprehensive flight safety assurance.

[0005] Therefore, there is an urgent need for a low-altitude airspace hybrid intelligent management and control solution with dynamic grid division, accurate identification, intelligent decision-making and flexible control capabilities. Summary of the Invention

[0006] In view of this, the present invention provides a method and system for intelligent integrated airspace management and control based on dynamic three-dimensional grids. Through intelligent decision-making algorithms and the delineation of dynamic safe flight zones, it can effectively coordinate flight conflicts between UAVs, ensure low-altitude flight safety, and promote the healthy development of the low-altitude economy.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a spatial intelligent fusion management and control method based on a dynamic three-dimensional mesh, comprising: Step 1: Acquire real-time environmental data and multi-source sensing data and transmit them to the ground control center; Step 2: The ground control center divides the low-altitude airspace into a variable-granularity grid based on real-time environmental data and assigns a dynamic control level to each grid; Step 3: Based on multi-source sensing data, synchronously collect image data and spatial location data of UAVs within the grid to locate and accurately identify the UAVs; Step 4: Generate intelligent control decisions based on the positioning and accurate identification results and the dynamic control level; Step 5: Activate the corresponding control measures based on the intelligent control decision results.

[0008] Preferably, the real-time environmental data in step 1 includes: airspace flight density, meteorological conditions, and regional importance parameters; the multi-source sensing data includes infrared image data, visible light image data, and lidar spatial location data.

[0009] Preferably, the low-altitude airspace is divided into grid regions of different granularities based on the flight density and regional importance parameters in the real-time environmental data; and control levels are determined according to the different granularities of the grid regions, and authorized aircraft types are set according to the corresponding control levels.

[0010] Preferably, step 3, which involves locating and accurately identifying the drone, includes: The infrared image data and the visible light image data are processed by the feature extractor of the DenseNet network model to obtain infrared apparent feature maps and visible light apparent feature maps respectively. The infrared apparent feature map and the visible light apparent feature map are diverged along the channel dimension to obtain a set of pixel-level feature vectors. The weighted sum is calculated through the dynamic integration module to generate multimodal pixel-level feature vectors. Global context association encoding is performed on the multimodal pixel-level feature vector to obtain a multimodal global feature vector; The multimodal global feature vector is input into the classifier to obtain the type recognition result, and the UAV recognition result is output by combining the spatial position data of the lidar.

[0011] Preferably, the ground control center generates a verification tag based on the UAV identification result, and delineates an ellipsoidal dynamic safe flight zone by combining the current flight speed, fuselage size and control level of the grid area of ​​the UAV, and generates corresponding graded expulsion commands based on the verification tag to obtain intelligent control decisions.

[0012] Preferably, the verification labels include compliant, partially non-compliant, and non-compliant; When all generated verification labels are compliant, a multi-drone flight conflict game model is constructed based on the overlap of the ellipsoidal safety zones of each drone, flight trajectory and speed parameters. The Nash equilibrium solution is solved by the improved Q-learning algorithm to obtain the optimal avoidance strategy with the minimum global comprehensive cost. When all drone verification tags are found to be in violation, a mandatory disposal order is triggered directly, and the violating drones are controlled by emitting lasers or ultrasonic waves. When a verification tag shows both compliance and partial violations, the ground control center generates a unique verification key to perform secondary identity verification on some of the violating drones. The violating drones that pass the verification participate in game theory according to the compliant scenario, while the violating drones that fail the verification are included in the restricted flight scope. Based on the conflict between their safe zone and the compliant drones, targeted avoidance instructions are generated.

[0013] Preferably, the Nash equilibrium solution is obtained by solving the improved Q-learning algorithm, and the generated optimal avoidance strategy includes graded drive-away commands such as speed adjustment, heading deflection, and altitude rise and fall, and the strategy response time is set.

[0014] Preferably, the graded expulsion instructions include no instruction, avoidance instruction, warning instruction, expulsion instruction, and cooperative avoidance instruction.

[0015] Preferably, a spatial intelligent fusion control system based on a dynamic three-dimensional grid includes: Data acquisition module: used to acquire real-time environmental data and multi-source sensing data and transmit them to the ground control center; Grid division module: The ground control center divides the low-altitude airspace into variable-granularity grids based on real-time environmental data and assigns dynamic control levels to each grid; Identification module: Based on multi-source sensing data, it synchronously collects image data and spatial location data of UAVs within the grid to locate and accurately identify UAVs; Decision module: Generates intelligent control decisions based on location and precise identification results and dynamic control levels; Measure activation module: Activates corresponding control measures based on the results of intelligent control decisions.

[0016] As can be seen from the above technical solution, compared with the prior art, the present invention discloses a method and system for intelligent fusion management and control of airspace based on dynamic three-dimensional mesh, with the following beneficial effects: 1. By using variable-granularity grids based on airspace flight density, regional importance parameters, and meteorological conditions, the goal is to achieve refined grids for key areas and lightweight grids for ordinary areas.

[0017] 2. Relying on the feature extraction and pixel-level dynamic fusion technology of the DenseNet network model, it integrates multi-source data from infrared, visible light and lidar, overcomes the limitations of single-modal recognition affected by lighting and weather, reduces UAV positioning error, improves type recognition accuracy, and effectively solves the problems of fuzzy UAV recognition and inaccurate positioning in complex environments.

[0018] 3. An ellipsoidal dynamic safe flight zone, defined based on parameters such as the drone's flight speed and weight, can adjust its boundaries in real time to follow the flight trajectory. Compared to traditional static regular zones, this provides more targeted and effective safety protection, accurately avoids conflicts between multiple drones, significantly reduces the incidence of collision accidents, and provides a dynamically adaptable safety barrier for drone flight.

[0019] 4. By solving the Nash equilibrium through the improved Q-learning algorithm, the flight game relationship between multiple UAVs is fully considered. The generated avoidance strategy (speed adjustment, heading deflection, altitude rise and fall) can minimize the global comprehensive cost, which not only meets the flight needs of individual UAVs, but also ensures the overall airspace order. The efficiency of resolving multi-aircraft conflicts is improved, and the problem of local optima and global imbalance caused by single rule decision-making is avoided.

[0020] 5. Based on the verification labels and conflict levels of compliance, partial violation, and complete violation, a hierarchical system is constructed that includes no instruction, avoidance instruction, warning instruction, expulsion instruction, and cooperative avoidance instruction. Combined with the policy response time setting, the control intensity is accurately matched with the actual scenario, which avoids excessive intervention in compliant flights and can respond quickly to violations, effectively improving control efficiency. At the same time, it ensures the priority right-of-way of emergency rescue drones and meets the control needs of special scenarios. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0022] Figure 1 The method flowchart provided by the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] like Figure 1 As shown, this embodiment of the invention discloses a spatial intelligent fusion management and control method based on a dynamic three-dimensional mesh, including: Step 1: Acquire real-time environmental data and multi-source sensing data and transmit them to the ground control center; Step 2: The ground control center divides the low-altitude airspace into a variable-granularity grid based on real-time environmental data and assigns a dynamic control level to each grid; Step 3: Based on multi-source sensing data, synchronously collect image data and spatial location data of UAVs within the grid to locate and accurately identify the UAVs; Step 4: Generate intelligent control decisions based on the positioning and accurate identification results and the dynamic control level; Step 5: Activate the corresponding control measures based on the intelligent control decision results.

[0025] Specifically, the real-time environmental data in step 1 includes: airspace flight density, meteorological conditions, and regional importance parameters; the multi-source sensing data includes infrared image data, visible light image data, and lidar spatial location data.

[0026] In a specific embodiment of the present invention, the process of preprocessing the above data includes: Spatiotemporal alignment calibration: A high-precision spatiotemporal alignment algorithm is adopted to calibrate the clocks of multi-source sensors using precise GPS time signals, ensuring time synchronization of data collected by different sensors.

[0027] Data cleaning and noise reduction: Set reasonable thresholds and filtering algorithms to filter outliers and noise interference in the collected data, thereby improving data accuracy and reliability.

[0028] Image scale normalization: For infrared and visible light images, a Gaussian pyramid model is used to generate image layers of different resolutions through multi-scale downsampling and upsampling, so as to achieve image scale consistency and lay the foundation for subsequent feature extraction and fusion.

[0029] Specifically, based on the flight density and regional importance parameters in the real-time environmental data, a granularity benchmark threshold for the grid area is preset, and the low-altitude airspace is divided into grid areas of different granularities; and control levels are determined according to the different granularities of the grid areas, and authorized aircraft types are set according to the corresponding control levels.

[0030] In a specific embodiment of the present invention, in low-altitude airspace management, in order to achieve refined management of different areas, it is necessary to divide the airspace into variable-granularity grids using a scientific algorithm based on flight density and regional importance parameters. A mathematical model for the grid granularity benchmark threshold is established based on flight density and regional importance parameters: ; Where ρ is the flight density (unit: aircraft / km) 2 ω represents the regional importance weight (ranging from 0 to 1, with 1 for core areas such as airport airspace and military-sensitive areas, and 0.2 to 0.3 for open suburban areas), and α and β are adjustment coefficients (calibrated according to the airspace management needs of different regions, typically α is 500 and β is 300). After calculating the baseline threshold using this model, the density clustering algorithm (DBSCAN) is used to adaptively divide the low-altitude airspace, forming a three-level grid system: coarse-grained monitoring area (≥500m×500m), medium-grained management area (200m×200m), and fine-grained control area (50m×50m).

[0031] Flight density in core urban business districts can reach 30 aircraft / km 2 With a regional importance weight of 1, the calculated baseline threshold is approximately 66.7, corresponding to a fine-grained grid of 50m × 50m; while the flight density in rural agricultural areas is only 2-3 aircraft / km. 2 The regional importance weight is set to 0.2, the baseline threshold is approximately 256.7, and the corresponding grid is divided into a coarse-grained grid of 500m×500m.

[0032] Furthermore, after the grid granularity is divided, a risk assessment matrix is ​​constructed using the Analytic Hierarchy Process (AHP) by combining the historical accident rate of the past 30 days and the real-time meteorological risk index (such as the risk index of 1.0-0.8 for severe weather such as thunderstorms and strong winds, and 0.2-0.1 for clear weather) within the grid, and assigning control levels (L) of 1-5 to each grid.

[0033] Level 1 is the lowest control level, and Level 5 is the highest control level. The specific level classification standards are as follows: Level 1 (Open Area): Rural open areas, flight density ≤ 5 aircraft / km 2 No critical facilities are required; all compliant and registered drones are permitted to fly. Level 2 (General Control Area): Suburban industrial areas, flight density 5-15 aircraft / km 2 There are a few civilian facilities that allow small and smaller compliant drones to fly; Level 3 (Moderate Control Area): Non-core urban areas, flight density 15-25 aircraft / km 2 There are many residential areas, which allow the flight of micro and light compliant drones; Level 4 (High Altitude Control Zone): Urban core areas and areas surrounding transportation hubs, with a flight density of 25-35 aircraft / km. 2 There are important public facilities, and only designated light drones are permitted to fly; Level 5 (No-Fly Zone): Airport airspace, military-sensitive area, flight density ≥ 35 aircraft / km 2 If national security is involved, only specially authorized drones for emergency rescue and other purposes are permitted to fly.

[0034] At the same time, an authorized aircraft whitelist is established for each control level, and the UAV identification code (UAV-ID) is compared with the whitelist in real time to achieve rapid verification of aircraft type access.

[0035] Specifically, step 3, the process of locating and accurately identifying the drone, includes: The infrared image data and the visible light image data are processed by the feature extractor of the DenseNet network model to obtain infrared apparent feature maps and visible light apparent feature maps respectively. The infrared apparent feature map and the visible light apparent feature map are diverged along the channel dimension to obtain a set of pixel-level feature vectors. The weighted sum is calculated through the dynamic integration module to generate multimodal pixel-level feature vectors. Global context association encoding is performed on the multimodal pixel-level feature vector to obtain a multimodal global feature vector; The multimodal global feature vector is input into the classifier to obtain the type recognition result, and the UAV recognition result is output by combining the spatial position data of the lidar.

[0036] Specifically, the ground control center generates a verification tag based on the UAV identification result, and delineates an ellipsoidal dynamic safe flight zone by combining the current UAV's flight speed, fuselage size, and control level of the grid area. Based on the verification tag, it generates corresponding graded expulsion commands to obtain intelligent control decisions.

[0037] In a specific embodiment of the present invention, an ellipsoidal dynamic safety zone is constructed with the UAV's center of mass as the origin, combined with flight speed (v) and control level (L). Its mathematical model is as follows: (x / a) 2 + (y / b) 2 + (z / c) 2 = 1, where the major axis a = 2 × v × t (t is the response time, set according to the control level, 3s for level 1-2, 2s for level 3-4, and 1s for level 5), the minor axis b = 1.5 × fuselage width, and the high axis c = 1.2 × fuselage height.

[0038] The safe zone is updated in real time according to the flight status of the drone. When the safe zones of two drones overlap, a collision warning mechanism is triggered.

[0039] Specifically, the verification labels include compliant, partially non-compliant, and non-compliant; When all generated verification labels are compliant, a multi-drone flight conflict game model is constructed based on the overlap of the ellipsoidal safety zones of each drone, flight trajectory and speed parameters. The Nash equilibrium solution is solved by the improved Q-learning algorithm to obtain the optimal avoidance strategy with the minimum global comprehensive cost. When all drone verification tags are found to be in violation, a mandatory disposal order is triggered directly, and the violating drones are controlled by emitting lasers or ultrasonic waves. When a verification tag shows both compliance and partial violations, the ground control center generates a unique verification key to perform secondary identity verification on some of the violating drones. The violating drones that pass the verification participate in game theory according to the compliant scenario, while the violating drones that fail the verification are included in the restricted flight scope. Based on the conflict between their safe zone and the compliant drones, targeted avoidance instructions are generated.

[0040] In one specific embodiment of the present invention, compliance verification employs a triple verification mechanism to generate verification labels: First, the model is verified by comparing the UAV-ID of the drone with the authorized model whitelist corresponding to the current grid management level. If the model does not match, it is marked as a violation. Second, qualification verification. If the aircraft type matches, the pilot's electronic license (validity period, permitted aircraft type) and flight plan filing information (flight range, time period, altitude) are further checked. If all information matches, it is deemed compliant; if any information does not match, it is marked as partially non-compliant. Thirdly, special exemptions are granted. Drones with exclusive emergency rescue identification are directly deemed compliant, and their priority in the safety zone is increased by 30%, allowing them priority passage.

[0041] Specifically, the improved Q-learning algorithm is used to solve the Nash equilibrium solution, generate graded clearance commands that include speed adjustment, heading deflection, and altitude gain / loss, and set the policy response time.

[0042] Specifically, the graded expulsion commands include no command, avoidance command, warning command, expulsion command, and cooperative avoidance command.

[0043] In a specific embodiment of the present invention, a non-cooperative game model involving multiple unmanned aerial vehicles (UAVs) is established to minimize collision risk (objective function 1: overlapping area of ​​safe zones ≤ 0.1m). 2The objective functions are: 1) Maximizing flight efficiency and 2) Flight path deviation ≤ 5% (a dual objective function). An improved Q-learning algorithm is used to solve the Nash equilibrium solution. The state space is divided into three categories: "safe zone with no overlap", "partial overlap", and "complete overlap". Twelve action spaces are designed (including speed adjustment ±5% and ±10%, heading deflection ±5°, ±10°, and ±15°, and altitude rise and fall ±3m and ±5m).

[0044] Generate tiered expulsion instructions based on the game outcome: For drones that violate regulations, a mandatory avoidance command will be sent directly, requiring the speed to be reduced to 0.5v, the heading to be deflected by 15°, and the altitude to be reduced to below 100m. For some non-compliant drones, send guidance and avoidance instructions such as "speed adjustment ±5% and heading deflection 10°"; For compliant drones, only collision warning information is sent, and the pilot can make adjustments independently.

[0045] The entire decision-making process response time is less than 200ms, at 10 aircraft / km. 2 In high-density scenarios, the accuracy of collision warnings is effectively improved.

[0046] Furthermore, a four-level response mechanism is designed: early warning alert, parameter limitation, heading guidance, and forced landing. The Level 1 response sends a voice warning to the drone pilot via the air traffic control platform; Level 2 response restricts the drone's flight speed and altitude via remote commands; The Level 3 response guides the drone away from the danger zone using real-time navigation information; A Level 4 response targets uncooperative drones that violate regulations by activating electromagnetic interference equipment to cut off their communication links and guiding them to a designated forced landing area.

[0047] Specifically, a spatial intelligent fusion management and control system based on a dynamic three-dimensional grid includes: Data acquisition module: used to acquire real-time environmental data and multi-source sensing data and transmit them to the ground control center; Grid partitioning module: Divides the low-altitude airspace into variable-granularity grids based on real-time environmental data, and assigns dynamic control levels to each grid; Identification module: Based on multi-source sensing data, it synchronously collects image data and spatial location data of UAVs within the grid to locate and accurately identify UAVs; Decision module: Generates intelligent control decisions based on location and precise identification results and dynamic control levels; Measure activation module: Activates corresponding control measures based on the results of intelligent control decisions.

[0048] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0049] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligent fusion management and control of spatial domain based on dynamic three-dimensional mesh, characterized in that, include: Step 1: Acquire real-time environmental data and multi-source sensing data and transmit them to the ground control center; Step 2: The ground control center divides the low-altitude airspace into a variable-granularity grid based on real-time environmental data and assigns a dynamic control level to each grid; Step 3: Simultaneously collect image data and spatial location data of UAVs within the grid based on multi-source sensing data to locate and accurately identify the UAVs; Step 4: Generate intelligent control decisions based on the positioning and accurate identification results and the dynamic control level; Step 5: Activate the corresponding control measures based on the intelligent control decision results.

2. The spatial intelligent fusion management and control method based on dynamic three-dimensional mesh according to claim 1, characterized in that, The real-time environmental data in step 1 includes: airspace flight density, meteorological conditions, and regional importance parameters; the multi-source sensing data includes infrared image data, visible light image data, and lidar spatial location data.

3. The spatial intelligent fusion management and control method based on dynamic three-dimensional mesh according to claim 2, characterized in that, Based on the flight density and regional importance parameters in the real-time environmental data, a granularity benchmark threshold for the grid area is preset, and the low-altitude airspace is divided into grid areas of different granularities. Control levels are then determined according to the different granularities of the grid areas, and authorized aircraft types are set accordingly based on the control levels.

4. The spatial intelligent fusion management and control method based on dynamic three-dimensional mesh according to claim 2, characterized in that, Step 3, the process of locating and accurately identifying the drone, includes: The infrared image data and the visible light image data are processed by the feature extractor of the DenseNet network model to obtain infrared apparent feature maps and visible light apparent feature maps respectively. The infrared apparent feature map and the visible light apparent feature map are diverged along the channel dimension to obtain a set of pixel-level feature vectors. The weighted sum is calculated through the dynamic integration module to generate multimodal pixel-level feature vectors. Global context association encoding is performed on the multimodal pixel-level feature vectors to obtain multimodal global feature vectors; The multimodal global feature vector is input into the classifier to obtain the type recognition result, and the UAV recognition result is output by combining the spatial position data of the lidar.

5. The spatial intelligent fusion management and control method based on dynamic three-dimensional mesh according to claim 4, characterized in that, The ground control center generates a verification tag based on the UAV identification result, and delineates an ellipsoidal dynamic safe flight zone by combining the current UAV's flight speed, fuselage size, and control level of the grid area. Based on the verification tag, it generates corresponding graded expulsion commands to obtain intelligent control decisions.

6. The spatial intelligent fusion management and control method based on dynamic three-dimensional mesh according to claim 5, characterized in that, The verification labels include compliant, partially non-compliant, and non-compliant; When all generated verification labels are compliant, a multi-drone flight conflict game model is constructed based on the overlap of the ellipsoidal safety zones of each drone, flight trajectory and speed parameters. The Nash equilibrium solution is solved by the improved Q-learning algorithm to obtain the optimal avoidance strategy with the minimum global comprehensive cost. When all drone verification tags are found to be in violation, a mandatory disposal order is triggered directly, and the violating drones are controlled by emitting lasers or ultrasonic waves. When a verification tag shows both compliance and partial violations, the ground control center generates a unique verification key to perform secondary identity verification on some of the violating drones. The violating drones that pass the verification participate in game theory according to the compliant scenario, while the violating drones that fail the verification are included in the restricted flight scope. Based on the conflict between their safe zone and the compliant drones, targeted avoidance instructions are generated.

7. The spatial intelligent fusion management and control method based on dynamic three-dimensional mesh according to claim 6, characterized in that, The improved Q-learning algorithm is used to solve the Nash equilibrium solution. The generated optimal avoidance strategy includes speed adjustment, heading deflection and graded clearance commands for altitude rise and fall, and the strategy response time is set.

8. The spatial intelligent fusion management and control method based on dynamic three-dimensional mesh according to claim 7, characterized in that, The graded expulsion commands include no command, avoidance command, warning command, expulsion command, and cooperative avoidance command.

9. A spatial intelligent fusion control system based on dynamic three-dimensional mesh, characterized in that, include: Data acquisition module: used to acquire real-time environmental data and multi-source sensing data; Grid partitioning module: Divides the low-altitude airspace into variable-granularity grids based on real-time environmental data, and assigns dynamic control levels to each grid; Identification module: Based on multi-source sensing data, it synchronously collects image data and spatial location data of UAVs within the grid to locate and accurately identify UAVs; Decision module: Generates intelligent control decisions based on location and accurate identification results and dynamic control levels; Measure activation module: Activates corresponding control measures based on the results of intelligent control decisions.

Images