Traffic management system and method of unmanned aerial vehicle, electronic equipment and medium

Abstract

The invention provides a flow management system and method for an unmanned aerial vehicle, electronic equipment and a medium. A central scheduler in the system obtains current pose data of the unmanned aerial vehicle; according to the current pose data, a preset flight task and a target area are combined, and a dynamic scheduling demand is generated; according to the dynamic scheduling requirement, a task decomposition rule is adopted, and a preset flight task is divided into a plurality of sub-tasks; distributing each subtask to a corresponding edge node based on a load balancing strategy; and the edge node generates a conflict avoidance scheme based on the target subtask allocated to the edge node by adopting a double-layer conflict detection mechanism and a path re-planning algorithm, and sends the conflict avoidance scheme to the unmanned aerial vehicle. The central scheduler decomposes a scheduling task into a plurality of sub-tasks through a distributed computing architecture and distributes the sub-tasks to the edge nodes for processing, then the edge nodes generate a conflict avoidance scheme for each sub-task, dynamic scheduling and conflict avoidance of the unmanned aerial vehicle are achieved, and therefore the processing real-time performance is improved, and the single-point fault risk is avoided.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a traffic management system, method, electronic equipment and medium for an unmanned aerial vehicle (UAV). Background Art

[0002] Unmanned Traffic Management (UTM) refers to the real-time coordination of the flight paths, speeds, altitudes and priorities of multiple drones in low-altitude airspace through technical means to prevent airspace conflicts, optimize flight efficiency and ensure safety.

[0003] The current development of drone traffic management technology can be divided into three phases: Phase I (before 2015): Manual scheduling was predominant, relying on visual flight rules (VFR), and supporting only small-scale, low-density scenarios. Phase II (2015-2020): Automated scheduling was initially achieved based on static airspace divisions (such as NASA's UTM project) and preset routes, but dynamic adaptability was poor. Phase III (2020 to present): The introduction of artificial intelligence (such as reinforcement learning) and edge computing enabled real-time route replanning.

[0004] Although current drone traffic management technology has evolved from manual scheduling to intelligent collaboration, it still faces many technical challenges. The dynamic scheduling and conflict avoidance of massive drones place extremely high demands on computing power. The traditional centralized scheduling architecture is difficult to meet real-time requirements and there is a risk of single point failure. Summary of the Invention

[0005] In order to overcome the above-mentioned problems of difficulty in meeting real-time requirements and the risk of single point failure, the present invention provides a traffic management system for drones, comprising a central dispatcher and a plurality of edge nodes in communication connection; the central dispatcher and each of the edge nodes are in communication connection with the drones;

[0006] The central scheduler is configured to obtain current posture data of the UAV; generate dynamic scheduling requirements based on the current posture data, in combination with a preset flight mission and a target area; divide the preset flight mission into multiple subtasks based on the dynamic scheduling requirements using a task decomposition rule; and assign each subtask to a corresponding edge node based on a load balancing strategy;

[0007] The edge node is used to generate a conflict avoidance plan based on the target subtask assigned to the edge node and send it to the drone by adopting a double-layer conflict detection mechanism and a path replanning algorithm.

[0008] Optionally, the central dispatcher is specifically used to predict the demand density index of each airspace that the UAV passes through during its flight to the target area based on the current posture data and the target area and based on the time series prediction model; and generate the dynamic scheduling requirements of the UAV in each airspace based on the demand density index of each airspace and the mission timeliness of the preset flight mission.

[0009] Optionally, the central scheduler is specifically used to divide the target area into multiple sub-areas based on the dynamic scheduling requirements by using a Voronoi diagram spatial segmentation algorithm; and based on each of the sub-areas, divide the preset flight mission into multiple sub-tasks by using a task chain sharding technology.

[0010] Optionally, the central scheduler is specifically configured to dynamically adjust the allocation strategy of the plurality of subtasks based on the resource utilization reported by each edge node, using an improved Hungarian algorithm with the goal of minimizing the maximum load.

[0011] Optionally, the edge node is specifically used to calculate the minimum safe distance between drones based on the speed barrier method; according to the obtained global optimization strategy model and the minimum safe distance between the drones, a fast exploration random tree algorithm that introduces dynamic constraints and minimum risk entropy evaluation criteria is used to generate a conflict avoidance plan.

[0012] Optionally, the edge node is further used to obtain and store the current posture data of the drone in a local spatiotemporal database; the local spatiotemporal database supports querying using a hybrid index structure of an R-tree combined with timestamp segmentation.

[0013] On the other hand, the present invention also provides a traffic management method for a drone, comprising:

[0014] Using a central dispatcher to obtain current posture data of the UAV;

[0015] Generate dynamic scheduling requirements based on the current posture data, combined with preset flight missions and target areas;

[0016] According to the dynamic scheduling requirements, the preset flight mission is divided into multiple subtasks using task decomposition rules;

[0017] Based on the load balancing strategy, each subtask is assigned to a corresponding edge node. The edge node uses a double-layer conflict detection mechanism and a path replanning algorithm based on the target subtask assigned to the edge node to generate a conflict avoidance plan and send it to the drone.

[0018] Optionally, generating dynamic scheduling requirements based on the current posture data in combination with a preset flight mission and a target area includes: predicting the demand density index of each airspace that the UAV passes through during its flight to the target area based on a time series prediction model according to the current posture data and the target area; and generating dynamic scheduling requirements for the UAV in each airspace according to the demand density index of each airspace and the mission timeliness of the preset flight mission.

[0019] Optionally, the task decomposition rules are used to divide the preset flight mission into multiple subtasks based on the dynamic scheduling requirements, including: based on the dynamic scheduling requirements, the Voronoi diagram spatial segmentation algorithm is used to divide the target area into multiple sub-areas; based on each of the sub-areas, the task chain sharding technology is used to divide the preset flight mission into multiple sub-tasks.

[0020] Optionally, the task decomposition rule is used to divide the preset flight mission into multiple subtasks, including: based on the resource utilization reported by each edge node, an improved Hungarian algorithm is used to dynamically adjust the allocation strategy of the multiple subtasks with the goal of minimizing the maximum load.

[0021] On the other hand, the present invention also provides a traffic management method for a drone, comprising:

[0022] Utilize the edge node to obtain the target subtask assigned to the edge node; the target subtask is that the central scheduler obtains the current posture data of the UAV, generates dynamic scheduling requirements based on the current posture data, combined with the preset flight mission and target area, and divides the preset flight mission into multiple subtasks based on the dynamic scheduling requirements using task decomposition rules, and allocates them based on the load balancing strategy;

[0023] Based on the target subtask, a two-layer conflict detection mechanism and a path replanning algorithm are adopted to generate a conflict avoidance plan and send it to the UAV.

[0024] Optionally, the use of a double-layer conflict detection mechanism and a path replanning algorithm to generate a conflict avoidance plan includes: calculating the minimum safe distance between drones based on the speed barrier method; and generating a conflict avoidance plan based on the acquired global optimization strategy model and the minimum safe distance between the drones using a fast exploration random tree algorithm that introduces dynamic constraints and a minimum risk entropy evaluation standard.

[0025] Optionally, it also includes: acquiring and storing the current posture data of the UAV in a local spatiotemporal database; the local spatiotemporal database supports querying using a hybrid index structure of an R-tree combined with timestamp segmentation.

[0026] On the other hand, the present invention further provides a computer device, characterized by comprising: one or more processors;

[0027] The processor is configured to store one or more programs;

[0028] When the one or more programs are executed by the one or more processors, any one of the above-mentioned methods for managing the traffic of a drone is implemented.

[0029] On the other hand, the present invention also provides a computer-readable storage medium, characterized in that a computer program is stored thereon, and when the computer program is executed, it implements the traffic management method of the drone described in any one of the above.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention provides a traffic management system, method, electronic device, and medium for an unmanned aerial vehicle (UAV). The system includes: a central scheduler and multiple edge nodes in communication connection; the central scheduler and each edge node are in communication connection with the UAV; the central scheduler obtains the UAV's current posture data; generates dynamic scheduling requirements based on the current posture data, in combination with a preset flight mission and target area; divides the preset flight mission into multiple subtasks based on the dynamic scheduling requirements using a task decomposition rule; assigns each subtask to a corresponding edge node based on a load balancing strategy; and the edge node generates a conflict avoidance plan based on the target subtask assigned to the edge node using a two-layer conflict detection mechanism and a path replanning algorithm, and sends the plan to the UAV. The central scheduler decomposes the scheduling task into multiple subtasks through a distributed computing architecture and assigns them to the edge nodes for processing. The edge nodes then generate a conflict avoidance plan for each subtask, thereby achieving dynamic scheduling and conflict avoidance for the UAV, thereby improving the real-time processing performance and avoiding the risk of single point failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 Schematic diagram of the traffic management system of the UAV of the present invention;

[0033] Figure 2 Schematic diagram of the flow of the traffic management method of the UAV of the present invention;

[0034] Figure 3 Schematic diagram of another flow chart of a method for managing traffic of a UAV according to the present invention;

[0035] Figure 4 Schematic diagram of the structure of the electronic device of the present invention. DETAILED DESCRIPTION

[0036] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0037] Example 1:

[0038] The present invention provides a traffic management system for drones, the structure of which is as follows: Figure 1 As shown, it includes a central scheduler 1 and multiple edge nodes 2 that are communicatively connected; the central scheduler 1 and each edge node 2 are communicatively connected to the drone;

[0039] The central scheduler is used to obtain the current posture data of the UAV. Based on the current posture data, combined with the preset flight mission and target area, it generates dynamic scheduling requirements. Based on the dynamic scheduling requirements, it uses task decomposition rules to divide the preset flight mission into multiple subtasks. Based on the load balancing strategy, each subtask is assigned to the corresponding edge node.

[0040] The edge node is used to generate a conflict avoidance plan based on the target subtask assigned to the edge node and send it to the UAV using a two-layer conflict detection mechanism and path replanning algorithm.

[0041] In order to overcome the problem that traditional centralized scheduling architecture is difficult to meet real-time requirements and has the risk of single point failure, making it impossible to achieve dynamic scheduling and conflict avoidance of massive drones, the embodiment of the present invention uses a central scheduler to decompose the scheduling task into multiple subtasks through a distributed computing architecture and assign them to edge nodes for processing. The edge nodes then generate a conflict avoidance plan for each subtask, realizing dynamic scheduling and conflict avoidance of drones, thereby improving the real-time processing and avoiding the risk of single point failure.

[0042] In order to accurately and effectively generate conflict avoidance solutions for drones, the drone's traffic management system includes a central scheduler and multiple edge nodes. The central scheduler is communicated with multiple edge nodes. In order to enable the drone to implement the conflict avoidance solution, the central scheduler and each edge node are communicated with the drone.

[0043] The central dispatcher is responsible for obtaining the current posture data of each drone, where the posture data includes position, speed and heading information. The heading information may refer to the yaw angle. It can be understood that each drone can send its own posture data to the central dispatcher in real time or at preset time intervals.

[0044] Specifically, the drone can be deployed with a lightweight embedded data acquisition module, which integrates a high-precision Global Navigation Satellite System (GNSS), an Inertial Measurement Unit (IMU) and airborne sensors. Among them, GNSS can be used for real-time dynamic differential positioning (RTK), and the airborne sensor can be a laser detection and ranging system (LiDAR) point cloud. The drone uploads real-time posture data at a frequency of 20Hz, including high-precision position (±5cm), speed (±0.1m / s) and heading angle (±0.5°) information.

[0045] The central dispatcher generates dynamic scheduling requirements based on the current posture data of each drone, combined with the preset flight mission and target area. Among them, the preset flight mission is defined by JSON structured instructions, including target coordinates, priority (urgent / routine), mission timeliness (hard deadline / flexible time), and is bound to the airspace dynamic map (including temporary obstacles and weather data) in real time. The target area refers to the specific geographical space that the drone needs to reach or cover when performing the preset flight mission. This can be a specific geographical location, such as a farm, a forest or an urban area, or it can be a wider geographical range. Dynamic scheduling requirements refer to the adjustments to the flight plan and task allocation based on real-time conditions during the drone's execution of preset flight missions and target area coverage.

[0046] Among them, the preset flight missions are defined through JSON structured instructions, including target coordinates, priority (urgent / routine), mission timeliness (hard deadline / flexible time), and are bound to the dynamic map of the airspace (including temporary obstacles and weather data) in real time.

[0047] In one possible implementation, the central scheduler may input the current posture data, preset flight mission, and target area of ​​the UAV into a pre-saved scheduling model to obtain the dynamic scheduling requirements output by the scheduling model.

[0048] To further enhance the flexibility and efficiency of mission execution, the central scheduler can employ task decomposition rules to break down complex pre-set flight missions into a series of more manageable and executable subtasks based on dynamic scheduling requirements. This process ensures the rationality of task allocation and promotes optimal resource allocation. For example, if a pre-set flight mission is to fly from point A to point B, the subtasks could be: flight from point A to point C, and flight from point C to point B.

[0049] After determining each subtask, the central scheduler can assign each subtask to the corresponding edge node based on a load balancing strategy. In one possible implementation, the central scheduler can match the requirements of each subtask with the capabilities of the edge node. For example, a subtask requiring a fast response can be assigned to an edge node with strong real-time performance.

[0050] After receiving the assigned target subtask, the edge node can first perform a preliminary conflict detection. Specifically, it can quickly scan the task path and schedule to check for obvious conflicts, such as time overlap and spatial intersection. It then conducts a deeper conflict analysis, taking into account additional factors such as the drone's performance limitations, flight rules, and environmental obstacles. After detecting a conflict, the edge node uses a path replanning algorithm to generate a conflict avoidance plan. The core of the path replanning algorithm is to find a new, conflict-free flight path that meets the drone's mission requirements. The edge node transmits the conflict avoidance plan to the drone, which then executes it. It should be noted that edge computing nodes are small data centers distributed at the edge of the network, used to quickly process and analyze local data.

[0051] The present invention adopts a distributed computing architecture: the central scheduler is only responsible for global airspace division and task classification, and the specific obstacle avoidance calculation is performed by the edge nodes.

[0052] The present invention provides a traffic management system for drones. The central scheduler decomposes the scheduling task into multiple subtasks through a distributed computing architecture and assigns them to edge nodes for processing. The edge nodes then generate a conflict avoidance plan for each subtask, realizing dynamic scheduling and conflict avoidance of drones, thereby improving the real-time processing performance and avoiding the risk of single point failures.

[0053] In order to accurately and effectively generate dynamic scheduling requirements, based on the above embodiments, in an embodiment of the present invention, the central scheduler is specifically used to predict the demand density index of each airspace that the drone passes through during its flight to the target area based on the current posture data and the target area and the time series prediction model; and generate the dynamic scheduling requirements of the drone in each airspace based on the demand density index of each airspace and the task timeliness of the preset flight mission.

[0054] To accurately and efficiently generate dynamic scheduling requirements, the central scheduler locally stores a time series prediction model. The central scheduler can input the drone's current position data and the specific location of the target area into the time series prediction model, which can be a long short-term memory network (LSTM). Specifically, the time series prediction model predicts the demand density index of each airspace that the drone will pass through when flying from its current location to the target area over a period of time. The demand density index refers to the intensity of existing flight activities, including the number of other aircraft, the intersection of flight paths, and possible air traffic control requirements.

[0055] In one possible implementation, based on a time series prediction model, combined with historical traffic heat maps and real-time airspace load, the demand density index of each airspace within the next 5 minutes (for example only) is predicted.

[0056] The central dispatcher dynamically generates drone scheduling requirements for each airspace based on the predicted demand density index for each airspace and the timeliness requirements of pre-set flight missions. Timeliness refers to the time sensitivity of a flight mission, such as when a mission needs to be completed within a specific timeframe or has specific flight speed requirements.

[0057] In one possible implementation, the central dispatcher can dynamically classify dispatch levels (low, medium, or high) based on the demand density index of each airspace, triggering differentiated response mechanisms (such as initiating a distributed collaborative avoidance mode in high-density areas). By coupling task timeliness with airspace congestion status, the dispatch priority of urgent tasks in congested areas can be automatically increased, avoiding the rigidity of the traditional "first-come, first-served" policy.

[0058] It should be noted that if the demand density index of a certain airspace is high, it means that there may be more flight activities in the area, and drones need to be more cautious when passing through. They may need to adjust their flight speed, altitude or route to avoid conflicts with other drones. At the same time, if the mission timeliness requirements are strict, drones may need to give priority to passing through areas with lower demand density to ensure that the mission is completed on time.

[0059] In order to accurately divide the preset flight mission into multiple subtasks, based on the above embodiments, in an embodiment of the present invention, the central scheduler is specifically used to divide the target area into multiple sub-areas based on the dynamic scheduling requirements by using the Voronoi diagram spatial domain segmentation algorithm; based on each sub-area, the preset flight mission is divided into multiple subtasks by using the task chain sharding technology.

[0060] After determining the dynamic scheduling requirements of the drones, the central scheduler further divides the target area using a Voronoi (Thiessen polygon) spatial partitioning algorithm. Specifically, the target area is divided into multiple adjacent, non-overlapping polygonal sub-regions. Each sub-region contains a specific point, called a spawning point, and the distance from any location within the sub-region to this spawning point is smaller than the distance to other spawning points.

[0061] In the present invention, the central scheduler divides the target airspace into multiple sub-areas according to the coverage range of edge nodes based on the Voronoi diagram airspace segmentation algorithm, ensuring that the number of drones in each sub-area is no more than a preset value, where the preset value can be 50.

[0062] When applying the Voronoi diagram airspace segmentation algorithm, the central dispatcher selects a series of appropriate generation points based on the target area's geographic characteristics, air traffic conditions, flight restrictions, and the dynamic scheduling requirements of drones. The Voronoi diagram airspace segmentation algorithm then calculates the Voronoi polygons bounded by these generation points, thereby dividing the target area into multiple sub-areas.

[0063] After acquiring each sub-region, the central scheduler uses task chain sharding technology to divide the pre-set flight mission into multiple sub-tasks. Task chain sharding is a task decomposition strategy that breaks down complex missions into multiple sub-tasks based on factors such as the mission's overall objective, flight path, required resources, and time constraints. For cross-regional flight missions, task chain sharding is used to split continuous flight segments into multiple sub-segments and distribute them to edge nodes along the route for relay processing.

[0064] In order to accurately and effectively allocate subtasks, based on the above embodiments, in an embodiment of the present invention, the central scheduler is specifically used to dynamically adjust the allocation strategy of multiple subtasks based on the resource utilization reported by each edge node, using an improved Hungarian algorithm with the goal of minimizing the maximum load.

[0065] To accurately and efficiently allocate subtasks, the central scheduler collects resource utilization data from each edge node, including metrics such as the central processing unit (CPU), memory, and bandwidth. Edge nodes send their resource utilization information to the central scheduler. The central scheduler uses a modified Hungarian algorithm, taking into account the current load and task requirements of each edge node, to determine the optimal allocation strategy.

[0066] The improved Hungarian algorithm not only considers the matching cost between tasks and edge nodes, but also dynamically adjusts weights to ensure that high-priority tasks are assigned to edge nodes with sufficient resources. Furthermore, through iterative optimization, the improved Hungarian algorithm gradually approaches the goal of minimizing the maximum load, avoiding situations where some edge nodes are overloaded while others have idle resources.

[0067] The present invention reduces cross-node coordination overhead through a segment continuity-aware sharding strategy, and reduces communication delay by more than 40% compared to traditional random sharding.

[0068] In order to accurately and effectively generate a conflict avoidance plan, based on the above embodiments, in an embodiment of the present invention, the edge node is specifically used to calculate the minimum safe distance between drones based on the speed barrier method; according to the obtained global optimization strategy model and the minimum safe distance between drones, a fast exploration random tree algorithm that introduces dynamic constraints and the minimum risk entropy evaluation standard is adopted to generate a conflict avoidance plan.

[0069] To accurately and effectively generate a collision avoidance solution, in an embodiment of the present invention, the edge node calculates the minimum safe distance between drones based on the speed barrier method. Specifically, the edge node can obtain the relative position and speed information between the drones and use the speed barrier method to calculate the minimum safe distance between the drones based on the relative position and speed information between the drones.

[0070] Specifically, the edge node can calculate the minimum safety distance using the following formula:

[0071] D safe =v max ×t resp +ε

[0072] Among them, D safe is the minimum safety distance, v max is the speed information, t resp is time, t resp It can be 1.2s, ε is the preset deviation, and ε can be 10m.

[0073] To accurately and efficiently generate collision avoidance solutions, edge nodes can construct a comprehensive flat state that describes the drone's motion state, including but not limited to its position, velocity, acceleration, and possibly attitude information. The edge node also specifies the drone's target state, which is a specific location or motion state that the drone wants to achieve. The target state can be pre-saved. The edge node then initializes a Rapidly Exploring Random Tree (RRT), whose root node represents the drone's starting state.

[0074] The edge node also builds an accurate dynamic model for the drone. This model not only reflects the dynamic relationship between the drone's position, velocity, and acceleration, but also considers the drone's physical limitations, such as maximum speed and maximum acceleration.

[0075] To assess the collision risk of drones in specific conditions, a predefined risk function is used. This function combines a global optimization strategy model with the minimum safe distance between drones to quantify the likelihood of a drone colliding with another drone or obstacle during flight. Furthermore, this invention introduces the concept of risk entropy, which reflects the uncertainty of the risk distribution. A lower risk entropy indicates a more certain risk distribution, meaning a lower probability of collision.

[0076] During the expansion phase of the RRT, a state is first randomly sampled from the flat state as a candidate node. Next, the nearest existing node in the RRT is found and a new node is generated by expanding from that node a certain distance in the direction of the candidate node. This new node must not only satisfy dynamic constraints but also be evaluated using a risk function to ensure its risk entropy is within an acceptable range. If the risk entropy of the new node falls below a preset threshold, or is lower than the risk entropy of certain nodes in the current RRT, it is added to the RRT.

[0077] Once the RRT is extended close enough to the target state, a path search can begin. The present invention uses an efficient search algorithm, such as Dijkstra's algorithm, to backtrack from the target state to the starting state and generate a preliminary path. However, this path may not be perfect, as it may contain some unnecessary twists and turns or sudden turns, which will increase the control difficulty and energy consumption of the drone. Therefore, this path can be smoothed to reduce the control input variation under dynamic constraints and improve the path's executableness.

[0078] After the path is smoothed, it can be checked for collisions. This ensures that every point on the path meets the minimum safe distance requirement during the actual flight of the drone, thereby avoiding collisions with other drones or obstacles. If a potential collision risk is detected, the path can be adjusted using local optimization algorithms such as gradient descent and particle swarm optimization to ensure that the drone can safely fly along the path.

[0079] This invention introduces dynamic constraints (such as maximum turning angles and acceleration limits) to ensure path feasibility. During tree growth, nodes with the lowest risk entropy (risk entropy = collision probability × severity) are prioritized, replacing the traditional pure distance cost function. Local paths can also be updated every 200ms, dynamically adjusting to real-time environmental changes.

[0080] In order to obtain the current posture data of the drone, based on the above embodiments, in an embodiment of the present invention, the edge node is also used to obtain and store the current posture data of the drone in the local spatiotemporal database; the local spatiotemporal database supports a hybrid index structure of R-tree combined with timestamp segmentation for query.

[0081] Edge nodes not only acquire real-time drone pose data but also store it in a local spatiotemporal database. This spatiotemporal database utilizes a hybrid indexing structure, combining R-tree and timestamp segmentation techniques, to support fast and accurate spatiotemporal queries. The spatial indexing capabilities of the R-tree enable edge nodes to efficiently process relevant queries, while timestamp segmentation further optimizes retrieval within the temporal dimension, ensuring high query performance even in complex spatiotemporal scenarios.

[0082] Edge nodes deploy customized edge computing units (ECUs), integrating NVIDIA Jetson AGX Orin (32TOPS computing power), 5G communication modules (supporting the NR-U band), and local storage (≥512GB SSD). Lightweight container technologies (such as Docker+K3s) enable hot-swapping of algorithm modules (such as conflict detection and path planning) within seconds and resource isolation.

[0083] Edge nodes use Kalman filtering to remove sensor noise and Delta encoding to compress trajectory data, reducing bandwidth usage to 15% of the original data. Furthermore, edge nodes store a snapshot of airspace status (such as drone position and obstacle coordinates) over the last 10 minutes, support millisecond-level spatiotemporal range queries (such as "status of all aircraft within a 500-meter radius"), and employ a hybrid index structure combining R-tree and timestamp segmentation to improve neighborhood search efficiency in conflict detection to O(log n) complexity (compared to O(n) for traditional brute-force searches).

[0084] Example 2:

[0085] Based on the same inventive concept, the present invention also provides a method for managing the flow of UAVs, which is applied to a central dispatcher. The flow chart is as follows: Figure 2 Shown, including:

[0086] Step 201: Using the central dispatcher to obtain the current posture data of the UAV;

[0087] Step 202: Generate dynamic scheduling requirements based on the current posture data, combined with the preset flight mission and target area;

[0088] Step 203: Based on the dynamic scheduling requirements, the preset flight mission is divided into multiple subtasks using the task decomposition rule;

[0089] Step 204: Based on the load balancing strategy, each subtask is assigned to the corresponding edge node. The edge node uses a double-layer conflict detection mechanism and a path replanning algorithm based on the target subtask assigned to the edge node to generate a conflict avoidance plan and send it to the UAV.

[0090] In a specific implementation, in the above step 201, based on the current posture data and the target area, based on the time series prediction model, the demand density index of each airspace that the UAV passes through during its flight to the target area is predicted; based on the demand density index of each airspace and the mission timeliness of the preset flight mission, the dynamic scheduling requirements of the UAV in each airspace are generated.

[0091] In a specific implementation, in step 203, the target area is divided into multiple sub-areas based on the dynamic scheduling requirements by using the Voronoi diagram spatial segmentation algorithm; based on each sub-area, the preset flight mission is divided into multiple sub-tasks by using the task chain slicing technology.

[0092] In a specific implementation, in step 204 , based on the resource utilization reported by each edge node, the improved Hungarian algorithm is used to dynamically adjust the allocation strategy of multiple subtasks with the goal of minimizing the maximum load.

[0093] Based on the same inventive concept, the present invention also provides a traffic management method for UAVs, which is applied to edge nodes. The flow chart is as follows: Figure 3 Shown, including:

[0094] Step 301: Utilize the edge node to obtain the target subtask assigned to the edge node; the target subtask is the central scheduler obtaining the current posture data of the UAV, generating dynamic scheduling requirements based on the current posture data, combined with the preset flight mission and target area, and then using the task decomposition rule to divide the preset flight mission into multiple subtasks based on the dynamic scheduling requirements and allocate them based on the load balancing strategy;

[0095] Step 302: Based on the target subtask, a two-layer conflict detection mechanism and a path replanning algorithm are used to generate a conflict avoidance plan and send it to the UAV.

[0096] In a specific implementation, in step 302, the minimum safe distance between UAVs is calculated based on the speed barrier method; based on the obtained global optimization strategy model and the minimum safe distance between UAVs, a fast exploration random tree algorithm that introduces dynamic constraints and the minimum risk entropy evaluation standard is used to generate a conflict avoidance plan.

[0097] In a specific implementation, it also includes: obtaining and storing the current posture data of the drone in a local spatiotemporal database; the local spatiotemporal database supports querying using a hybrid index structure of an R-tree combined with timestamp segmentation.

[0098] Example 3:

[0099] like Figure 4 As shown, the present invention also provides an electronic device, which may be a computer, a single-chip microcomputer, a smart mobile device, or the like. The electronic device in this embodiment may include a processor, a memory, a transceiver component, and the like. The memory, processor, and transceiver component are connected via a bus; the memory may be used to store an execution program, which may include instructions; and the processor may be used to execute the instructions stored in the memory. The memory may also be used to store data, which may be accessed and / or modified during the execution of the instructions.

[0100] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the storage medium to implement the corresponding method flow or corresponding function, so as to implement the steps of the traffic management method of a drone in the above embodiment.

[0101] Example 4:

[0102] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory). The electronic device readable storage medium is a memory device in the electronic device, used to store programs and data. It is understandable that the storage medium here can include both the built-in storage medium in the electronic device and, of course, the extended storage medium supported by the electronic device. The storage medium provides storage space, which stores the terminal's operating system. In addition, the storage space also stores one or more instructions suitable for being loaded and executed by the processor. These instructions can be one or more executable programs (including program code). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory (non-volatile memory), such as at least one disk storage. The processor loads and executes one or more instructions stored in the storage medium to implement the steps of the traffic management method for a drone in the above embodiment.

[0103] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0104] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0105] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0106] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0107] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit its scope of protection. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that after reading the present invention, those skilled in the art may still make various changes, modifications or equivalent substitutions to the specific implementation methods of the invention, but these changes, modifications or equivalent substitutions are all within the scope of protection of the pending claims of the invention.

Claims

1. A traffic management system for drones, characterized in that: A central scheduler and a plurality of edge nodes are communicatively connected; the central scheduler and each edge node are communicatively connected to a drone; The central dispatcher is used to obtain the current posture data of the UAV; Generate dynamic scheduling requirements based on the current posture data, combined with preset flight missions and target areas; According to the dynamic scheduling requirements, the preset flight mission is divided into multiple subtasks using task decomposition rules; Based on the load balancing strategy, each of the subtasks is assigned to the corresponding edge node; The edge node is used to generate a conflict avoidance plan based on the target subtask assigned to the edge node and send it to the drone by adopting a double-layer conflict detection mechanism and a path replanning algorithm.

2. The system according to claim 1, wherein The central dispatcher is specifically used to predict the demand density index of each airspace that the UAV passes through during its flight to the target area based on the current posture data and the target area and the time series prediction model; and generate the dynamic scheduling requirements of the UAV in each airspace based on the demand density index of each airspace and the mission timeliness of the preset flight mission.

3. The system according to claim 1, wherein: The central scheduler is specifically used to divide the target area into multiple sub-areas based on the dynamic scheduling requirements by using the Voronoi diagram spatial domain segmentation algorithm; based on each of the sub-areas, the preset flight mission is divided into multiple sub-tasks by using the task chain slicing technology.

4. The system according to claim 3, wherein: The central scheduler is specifically configured to dynamically adjust the allocation strategy of the plurality of subtasks based on the resource utilization reported by each edge node and using an improved Hungarian algorithm with the goal of minimizing the maximum load.

5. The system according to claim 1, wherein: The edge node is specifically used to calculate the minimum safe distance between drones based on the speed barrier method; based on the obtained global optimization strategy model and the minimum safe distance between drones, a fast exploration random tree algorithm that introduces dynamic constraints and risk entropy minimum evaluation criteria is used to generate a conflict avoidance plan.

6. The system according to claim 5, wherein: The edge node is further used to obtain and store the current posture data of the drone in a local spatiotemporal database; the local spatiotemporal database supports querying using a hybrid index structure that combines R-tree with timestamp segmentation.

7. A method for managing the flow of UAVs, characterized in that: include: Using a central dispatcher to obtain current posture data of the UAV; Generate dynamic scheduling requirements based on the current posture data, combined with preset flight missions and target areas; According to the dynamic scheduling requirements, the preset flight mission is divided into multiple subtasks using task decomposition rules; Based on the load balancing strategy, each subtask is assigned to a corresponding edge node. The edge node uses a double-layer conflict detection mechanism and a path replanning algorithm based on the target subtask assigned to the edge node to generate a conflict avoidance plan and send it to the drone.

8. A method for managing the flow of UAVs, characterized in that: include: Utilizing the edge node to obtain the target subtask assigned to the edge node; The target subtask is that the central scheduler obtains the current posture data of the UAV, generates dynamic scheduling requirements based on the current posture data, combined with the preset flight mission and target area, and divides the preset flight mission into multiple subtasks based on the dynamic scheduling requirements and the task decomposition rules, and allocates them based on the load balancing strategy; Based on the target subtask, a two-layer conflict detection mechanism and a path replanning algorithm are adopted to generate a conflict avoidance plan and send it to the UAV.

9. An electronic device, characterized in that: include: at least one processor and memory; The memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, the traffic management method for a drone as claimed in claim 7 is implemented, or the traffic management method for a drone as claimed in claim 8 is implemented.

10. A readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, the traffic management method for the drone as claimed in claim 7 is implemented, or the traffic management method for the drone as claimed in claim 8 is implemented.

Images