Cloud edge cooperative computing method and system for low-altitude unmanned aerial vehicle supervision space-time stream data, terminal and storage medium

Abstract

The invention discloses a cloud-edge cooperative computing method and system for low-altitude unmanned aerial vehicle supervision space-time stream data, a terminal and a storage medium, and the method comprises the steps: building a cloud-edge integrated cooperative mode of an unmanned aerial vehicle supervision task based on a cloud-edge network structure, and converting a cloud-edge computing resource topological relation into a graph network structure; based on node sequence features of a cloud edge integrated cooperation mode, constructing a cloud edge cooperation allocation strategy of a cross-edge node global task adaptive to a graph network structure, and realizing cross-edge node global task cloud edge cooperation allocation; and establishing a cloud-side resource scheduling adaptive optimization mechanism considering global storage, calculation and transmission low delay cost, and realizing cloud-side resource scheduling adaptive optimization of a space-time flow calculation task on a global level. According to the method, a cloud edge collaborative optimization scheduling mechanism for large-scale unmanned aerial vehicle supervision space-time flow data calculation in a three-dimensional space is established, so that the efficient scheduling requirement of real-time calculation of unmanned aerial vehicle supervision tasks in a cloud edge collaborative mode is met.

Description

Technical Field

[0001] The present invention relates to the field of drone data processing technology, and in particular to a cloud-edge collaborative computing method, system, terminal and computer-readable storage medium for low-altitude drone monitoring spatiotemporal flow data. Background Art

[0002] Low-altitude drone supervision requires real-time acquisition of drone flight information such as position, attitude, speed and heading. This information constitutes spatiotemporal flow data in three-dimensional space with characteristics such as continuous arrival and infinite growth. How to perform real-time computation and analysis on the spatiotemporal flow data generated during the flight of large-scale low-altitude drones to support real-time response to tasks such as drone anomaly detection and conflict warning has become a major requirement for ensuring the safe operation of large-scale drones.

[0003] The use of cloud-edge collaborative computing architecture to efficiently calculate drone supervision tasks is an important guarantee for the real-time calculation of various supervision tasks during large-scale drone flights. The cloud-edge collaborative computing architecture realizes the complementary advantages between cloud computing and edge computing. The cloud provides massive computing and storage resources, and the edge provides real-time task processing through widely distributed nodes. Under the cloud-edge collaborative computing architecture, reasonable task scheduling is used to coordinate available resources such as computing, storage, network, and virtualization resources to achieve the best scheduling strategy, so as to efficiently complete computing tasks. This is the key to unleashing the performance of cloud-edge collaboration. For example, mathematical programming algorithms, heuristic algorithms, evolutionary algorithms, game algorithms, reinforcement learning algorithms, machine learning algorithms and other methods are used to achieve optimal task scheduling.

[0004] In recent years, there have also been studies on lighter scheduling algorithms and collaborative frameworks, such as lightweight genetic algorithms and space-ground integrated cloud-edge collaborative computing frameworks. Overall, the existing research on task scheduling algorithms under cloud-edge collaboration provides theoretical and practical support for the performance of cloud-edge collaborative computing. However, most studies do not consider the difficulties in cloud-edge task division and resource collaborative scheduling caused by the high spatiotemporal dependency of tasks.

[0005] Therefore, the prior art still needs to be improved and developed. Summary of the invention

[0006] The main purpose of the present invention is to provide a cloud-edge collaborative computing method, system, terminal and computer-readable storage medium for spatiotemporal flow data of low-altitude UAV supervision, aiming to solve the problem that when the cloud-edge collaborative computing architecture is used to calculate UAV supervision tasks in the prior art, the high spatiotemporal dependency of the tasks is not considered, which leads to difficulty in cloud-edge task division and resource collaborative scheduling.

[0007] To achieve the above object, the present invention provides a cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data, and the cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data comprises the following steps: Establish a cloud-edge integrated collaboration model for drone monitoring tasks based on a cloud-edge network structure, and convert the topological relationship of cloud-edge computing resources into a graph network structure; Based on the node sequence characteristics of the cloud-edge integrated collaboration mode, a cloud-edge collaborative allocation strategy for global tasks across edge nodes that adapts to the graph network structure is constructed to achieve cloud-edge collaborative allocation of global tasks across edge nodes; Based on the cloud-edge collaborative allocation strategy of global tasks across edge nodes, an adaptive optimization mechanism for cloud-edge resource scheduling that takes into account the global storage, computing and transmission low latency costs is established, and adaptive optimization of cloud-edge resource scheduling for spatiotemporal stream computing tasks is achieved at the global level.

[0008] Optionally, the cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data, wherein the cloud-edge integrated collaborative mode for UAV supervision tasks based on the cloud-edge network structure is established, and the cloud-edge computing resource topological relationship is converted into a graph network structure, specifically including: Map the topological relationship of cloud-edge resources for large-scale drone supervision and build a cloud-edge integrated network model. = (Uc, Ue, V), where Uc represents the central node of the cloud-edge structure, Ue represents the edge node of the cloud-edge structure, and V represents the homogeneous attribute edge set in the cloud-edge integrated network model; Define a set of heterogeneous points (Uc, Ue), Uc = {nc1, nc2, ..., nc n}, represents the cloud center node and related attribute mapping in the cloud edge structure, where nc1 to nc n Represents multiple attribute mappings in the central node, Ue={ne1, ne2, ..., ne n}, indicating edge nodes and related attribute mapping, where ne1 to ne n Represents multiple attribute mappings in edge nodes; node attributes are represented as: ; in, Represents node attributes, including central nodes and edge nodes. Considering the influencing factors comprehensively, Indicates node computing resources, Represents node storage resources, Indicates the node function type, Indicates the current load status or task deployment of the node; Define a homogeneous attribute edge set V: ; in, represents the homogeneity property of the edge set between edge nodes, It represents the transmission cost incurred by the nodes to interact, including the delay in the communication or information transmission process. Indicates the amount of computing tasks migrated by the node during task scheduling. represents the efficiency of information transmission between nodes, Indicates the type of communication or edge connectivity between nodes; Cloud-edge integrated network model The expression is: ; in, Indicates the number of cloud nodes, Expressed as the number of edge nodes, Represents the cloud-cloud subnetwork, which is used to describe the communication status between cloud nodes. represents the edge-edge subnetwork, which is used to characterize the cross-edge task computing process. and Representing cloud-edge heterogeneous bipartite networks with different numbers of nodes, used to characterize the communication process between cloud and edge; Based on the cloud-edge integrated network model, a cloud-edge integrated collaborative mode for drone supervision tasks is established. The structural walk and feature walk analysis methods of heterogeneous node networks are used to integrate the semantic associations of different node types into the network structure for network representation learning, thereby capturing the structural similarity and attribute characteristics of the network.

[0009] Optionally, in the cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data, the process of the cloud-edge collaborative allocation strategy includes: allocating local real-time computing tasks to edge clusters and controlling global tasks through the cloud to achieve cloud-edge collaboration; The node sequence characteristics of the cloud-edge integrated collaborative mode are used to construct a cloud-edge collaborative allocation strategy for global tasks across edge nodes that adapts to the graph network structure to achieve cloud-edge collaborative allocation of global tasks across edge nodes, specifically including: Assign local real-time computing tasks to the nearby edge for computing. If the computing resources of the nearby edge cannot meet the requirements of low-latency processing, the cloud will be used to coordinate the tasks and offload the local real-time computing tasks to another nearby edge or cloud for takeover and processing, so as to achieve cross-edge communication and global collaborative optimization. When the local real-time computing tasks transmitted to the edge are processed, the edge transmits the real-time instructions and data to the terminal, and transmits the spatiotemporal data to the cloud, and uses the situation awareness model to perform state awareness and classification processing on the spatiotemporal data; The cloud is used to store big data and train intelligent supervision models for processing various supervision tasks, and regularly distribute the trained intelligent supervision models to each edge node; The cost loss and efficiency generated in the cross-edge communication and computing process are dynamically estimated and transmitted to the cloud-edge integrated network, multiple independent real-time computing units of different granularities are constructed, and multiple independent real-time computing units are dynamically combined in the cloud-edge collaborative scheduling service chain to build a scheduling service chain for cloud-edge scheduling; Based on the workflow, the scheduling service chain within the preset time window is automatically executed. According to the resource status, network conditions and task priority of each node in each time window, the computing task units are intelligently allocated and orchestrated to ensure that the nodes are efficiently executed in parallel within each time window.

[0010] Optionally, the cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data, wherein the establishment of a cloud-edge resource scheduling adaptive optimization mechanism that takes into account the global storage, computing and transmission low latency cost specifically includes: Establish a balanced cloud-edge resource allocation strategy that takes into account the global storage, computing, and transmission low latency costs; Establish a global task cloud-edge resource dynamic allocation mechanism based on network structure; Establish a dynamic and iterative cloud-edge resource scheduling adaptive optimization mechanism.

[0011] Optionally, the cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data, wherein the establishment of a cloud-edge resource balanced allocation strategy that takes into account the global storage, computing and transmission low latency cost specifically includes: The correlation of real-time computing tasks of spatiotemporal streams is formally described based on computing task units formed by information at different time points and different interaction modes. The dependency of computing tasks in different time windows on cloud-edge computing resources and the storage, computing, and transmission delay costs generated under the current task are mapped to nodes and edges in the network, and dynamic calculation of feature attributes is performed. Construct a network-structure-based load balancing allocation algorithm driven by latency cost, measure the global resource allocation and system load status of the network through balance evaluation indicators, simulate cloud-edge computing task allocation and execution with dynamic configuration of node attributes and edge attributes, and converge the task allocation process to the optimal state with the minimum latency cost of a single task and maximum resource utilization as the objective function; The process of load balancing based on network structure includes: Obtain computing resources of edge nodes and computing task requirements carried by them, and coordinate and define the current load state of the nodes through the difference in computing task requirements, wherein the load state includes high load, low load and balanced state; Evaluate the cost of communication and information transmission between nodes, obtain edge weights in the network to assist algorithm decision-making, and define edge states; Based on node attributes and edge attributes, the maximum weight sum matching is used to optimally match cloud-edge resources and delay costs in the network to obtain a balanced allocation solution that satisfies minimum delay and maximum resource utilization.

[0012] Optionally, the cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data, wherein the establishment of a global task cloud-edge resource dynamic allocation mechanism based on a network structure specifically includes: Based on the global perspective of the cloud-edge integrated network, the cloud center systematically collects and organizes the target key information of all global tasks to be executed, including the computing intensity of the task, the amount of data, the delay cost, and the dependencies between tasks; Use mapping relationships and node attributes to obtain all edge nodes with available resources, comprehensively consider the hardware configuration, network status and service capabilities of each edge node, conduct real-time evaluation of resource volume and computing efficiency, save global tasks in the corresponding edge cluster, and locate the most suitable edge node for each global task; Based on the optimal scheduling strategy, global tasks are assigned to the cloud center and related edge nodes for processing one by one, so as to complete the full-link management from global task feature analysis, edge node adaptation, edge node screening to specific task allocation.

[0013] Optionally, the cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data, wherein the establishment of a dynamic iterative cloud-edge resource scheduling adaptive optimization mechanism specifically includes: Capture and update the target core information in the global scope of the cloud-edge integrated network in real time, and track and record the status of all global computing tasks in real time, including the progress of task initiation, execution, and completion, as well as the interdependencies between tasks; Based on the cloud-edge network topology and task characteristics, a delay cost prediction model is established to calculate the execution cost of tasks on different edge nodes, and the minimization of global delay cost is taken as the main optimization goal; Based on the process-driven dynamic iterative scheduling algorithm, it makes adaptive adjustments and optimizations according to the real-time changing task requirements and system resource status, and adaptively adjusts the task scheduling plan according to the real-time requirements of different tasks, network transmission delays and computing resource constraints.

[0014] In addition, to achieve the above-mentioned purpose, the present invention also provides a cloud-edge collaborative computing system for low-altitude UAV monitoring spatiotemporal flow data, wherein the cloud-edge collaborative computing system for low-altitude UAV monitoring spatiotemporal flow data includes: The cloud-edge integrated collaborative model establishment module is used to establish a cloud-edge integrated collaborative model for drone supervision tasks based on the cloud-edge network structure, and convert the topological relationship of cloud-edge computing resources into a graph network structure; A cloud-edge collaborative allocation strategy construction module is used to construct a cloud-edge collaborative allocation strategy for global tasks across edge nodes that is adapted to the graph network structure based on the node sequence characteristics of the cloud-edge integrated collaborative mode, so as to realize cloud-edge collaborative allocation of global tasks across edge nodes; The cloud-edge resource scheduling optimization module is used to establish a cloud-edge resource scheduling adaptive optimization mechanism that takes into account the global storage, computing and transmission low latency costs based on the cloud-edge collaborative allocation strategy of the global tasks across edge nodes, and realize the cloud-edge resource scheduling adaptive optimization of spatiotemporal stream computing tasks at the global level.

[0015] In addition, to achieve the above-mentioned purpose, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a cloud-edge collaborative computing program for low-altitude UAV supervision spatiotemporal stream data stored on the memory and run on the processor, and when the cloud-edge collaborative computing program for low-altitude UAV supervision spatiotemporal stream data is executed by the processor, the steps of the cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal stream data as described above are implemented.

[0016] In addition, to achieve the above-mentioned objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a cloud-edge collaborative computing program for low-altitude UAV supervision spatiotemporal flow data, and when the cloud-edge collaborative computing program for low-altitude UAV supervision spatiotemporal flow data is executed by a processor, the steps of the cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data as described above are implemented.

[0017] In the present invention, a cloud-edge integrated collaborative mode for UAV supervision tasks based on a cloud-edge network structure is established, and the cloud-edge computing resource topological relationship is converted into a graph network structure; based on the node sequence characteristics of the cloud-edge integrated collaborative mode, a cloud-edge collaborative allocation strategy for global tasks across edge nodes that adapts to the graph network structure is constructed to achieve cloud-edge collaborative allocation of global tasks across edge nodes; based on the cloud-edge collaborative allocation strategy for global tasks across edge nodes, a cloud-edge resource scheduling adaptive optimization mechanism that takes into account the global storage, computing and transmission low latency costs is established, and cloud-edge resource scheduling adaptive optimization of spatiotemporal stream computing tasks is achieved at the global level. The present invention establishes a cloud-edge collaborative optimization scheduling mechanism for large-scale UAV supervision spatiotemporal stream data calculations in three-dimensional space to support the efficient scheduling requirements of real-time computing of UAV supervision tasks under the cloud-edge collaborative mode. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of a preferred embodiment of the cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data of the present invention; Figure 2 It is a schematic diagram of the overall process in a preferred embodiment of the cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data of the present invention; Figure 3 It is the overall technical roadmap of the preferred embodiment of the cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data of the present invention; Figure 4 It is a schematic diagram of a cloud-edge integrated network model structure in a preferred embodiment of the cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data of the present invention; Figure 5 It is a schematic diagram of collaborative scheduling of computing tasks across edge segments in a preferred embodiment of the cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data of the present invention; Figure 6 It is a schematic diagram of a global cloud-edge resource scheduling adaptive optimization mechanism in a preferred embodiment of the cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data of the present invention; Figure 7 It is a structural diagram of a preferred embodiment of the cloud-edge collaborative computing system for monitoring spatiotemporal flow data of low-altitude UAVs of the present invention; Figure 8 It is a structural diagram of a preferred embodiment of the terminal of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the present invention clearer and more specific, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0020] The present invention discloses a cloud-edge collaborative computing method for spatiotemporal flow data of low-altitude unmanned aerial vehicle supervision. Based on cloud-edge collaborative technology, real-time and efficient computing of unmanned aerial vehicle real-time tasks and reasonable allocation of computing resources are performed, cloud-edge computing resources are abstracted into graph network relationships according to topological relationships and computing capabilities, a graph network model is established, and computing resources are quantified to provide support for subsequent task allocation. Subsequently, a cross-edge node global task cloud-edge collaborative allocation strategy that adapts to the graph network structure is constructed to realize the inclination of unmanned aerial vehicle computing tasks at the high-load edge end to computing tasks at the low-load edge end, thereby ensuring load balancing of each edge end. Finally, a cloud-edge resource scheduling adaptive optimization mechanism with low latency cost that comprehensively considers storage, computing and transmission is established, and a global network load balancing algorithm for storage, computing and transmission is implemented from unmanned aerial vehicle task queues, data stream transmission to computing resource allocation, and dynamic optimization of the load balancing algorithm is realized through unmanned aerial vehicle feedback.

[0021] The cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data described in the preferred embodiment of the present invention is as follows: Figure 1 , Figure 2(Overall process diagram) and Figure 3 As shown in the (Technical Route Schematic Diagram), the cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data includes the following steps: Step S10: Establish a cloud-edge integrated collaborative mode for drone supervision tasks based on a cloud-edge network structure, and convert the topological relationship of cloud-edge computing resources into a graph network structure.

[0022] Specifically, the cloud-edge resource topology relationship of large-scale drone supervision is network mapped, and a cloud-edge integrated network model is constructed. Resource exploration and performance evaluation are carried out on the cloud and edge of large-scale drone supervision, including key parameters such as computing power, storage capacity, communication bandwidth, and response delay. The cloud-edge integrated network model is constructed using graph theory and other related theories. In the cloud-edge integrated network model, the cloud is a core node with powerful processing power and massive storage space, while the edge distributed in different geographical locations and close to the drone flight area serves as an intermediary layer connecting the cloud and terminal devices (drones). Each edge is reflected in the network model in the form of a node according to its physical location, service capabilities, and communication link characteristics with the cloud and drones.

[0023] Cloud-edge integrated network model = (Uc, Ue, V) is mainly formed by two key definitions, including point set definition and edge set definition, where Uc represents the central node of the cloud-edge structure, Ue represents the edge node of the cloud-edge structure, and V represents the homogeneous attribute edge set in the cloud-edge integrated network model; first, define the heterogeneous type point set (Uc, Ue), Uc={nc1, nc2,…, nc n}, represents the cloud center node and related attribute mapping in the cloud edge structure, where nc1 to nc n Represents multiple attribute mappings in the central node, Ue={ne1, ne2, ..., ne n}, indicating edge nodes and related attribute mapping, where ne1 to ne n Represents multiple attribute mappings in edge nodes; node attributes are represented as: ; in, Represents node attributes, including central nodes and edge nodes. Considering the influencing factors comprehensively, Indicates node computing resources, Represents node storage resources, Indicates the node function type, Indicates the current load status or task deployment of the node.

[0024] At the same time, a homogeneous attribute edge set V is defined, whose weight or attribute represents the connection status and information transmission process between cloud and edge, and between edges. Its specific definition is expressed as: ; in, represents the homogeneity property of the edge set between edge nodes, It represents the transmission cost incurred by the nodes to interact, including the delay in the communication or information transmission process. Indicates the amount of computing tasks migrated by the node during task scheduling. represents the efficiency of information transmission between nodes, Indicates the type of communication or edge connectivity between nodes, which can be subdivided into no connection, cloud-edge connection, and edge-edge connection.

[0025] Cloud-edge integrated network model The structure is expressed through the adjacency matrix, including the "cloud-cloud" subnetwork composed of the cloud node point set Uc and its corresponding edge set , used to characterize the communication status between cloud nodes; the "edge-edge" subnetwork composed of the edge node set Ue and its corresponding edge set , used to characterize the cross-edge task computing process; the "cloud-edge" heterogeneous bimodal network composed of the cloud node point set Uc and the edge node point set Ue or , which is used to describe the communication process between cloud and edge. Therefore, the cloud-edge integrated network model The expression can be: ; in, Indicates the number of cloud nodes, Expressed as the number of edge nodes, Represents the cloud-cloud subnetwork, which is used to describe the communication status between cloud nodes. represents the edge-edge subnetwork, which is used to characterize the cross-edge task computing process. and Representing cloud-edge heterogeneous bipartite networks with different numbers of nodes, used to characterize the communication process between cloud and edge. Figure 4 , Figure 4 The cloud Uc in the example exchanges information with edge nodes Ue1, Ue2, Ue3, Ue4, and Uen respectively.

[0026] Then, based on the cloud-edge integrated network model, a cloud-edge integrated collaborative mode for drone supervision tasks is established. The structural walk and feature walk analysis methods of heterogeneous node networks are used to integrate the semantic associations of different node types into the network structure for network representation learning, thereby capturing the structural similarity and attribute characteristics of the network.

[0027] Through the node sequence of the corresponding path (ne i , nc j ,ne i+1 ,…) define the “cloud-edge” interaction (nc j ,ne i ), “edge-edge” interaction (ne i ,ne j ), “edge-cloud-edge” interaction (ne i , nc j ,ne i+1 ), “edge-cloud-edge-edge” interaction (ne i , nc j ,ne i+1 ,ne i+2 ), where ne represents different node sequences (i represents nodes) and nc represents different edge sequences (j represents edges). The information sequence containing node status and its attribute information is represented as various basic collaboration modes and their combination forms in the cloud-edge system, thereby supporting the design and implementation of collaborative optimization scheduling and adaptive resource allocation strategies for cloud-edge computing tasks.

[0028] Step S20: Based on the node sequence characteristics of the cloud-edge integrated collaborative mode, a cloud-edge collaborative allocation strategy for global tasks across edge nodes that is adapted to the graph network structure is constructed to achieve cloud-edge collaborative allocation of global tasks across edge nodes.

[0029] Specifically, firstly, based on the node sequence representation of the cloud-edge integrated collaboration mode, the node sequence representation is the attribute information displayed by different nodes, and a cross-edge node global task cloud-edge collaborative allocation strategy that adapts to the network structure is constructed. The process of the cloud-edge collaborative allocation strategy includes: allocating local real-time computing tasks to edge clusters and controlling global tasks through the cloud to achieve cloud-edge collaboration.

[0030] Local real-time computing tasks are assigned to nearby edge terminals for fast computing. If the computing resources of the edge terminal cannot meet the requirements of low-latency processing, the tasks are coordinated through the cloud, and the tasks are offloaded to another nearby edge terminal or cloud terminal for takeover and processing, thereby achieving cross-edge communication and global collaborative optimization. When the computing tasks transmitted to the edge terminal are processed, the edge terminal transmits real-time instructions and data to the terminal, and at the same time transmits the data required for applications that do not require high real-time performance to the cloud. The situation awareness model is used to perceive and classify the data status of these spatiotemporal data, providing data support for early warning. The cloud has a global control function, which is responsible for the storage of big data and the training of intelligent supervision models for processing various supervision tasks, and regularly distributes the trained models to each edge node to further shorten the computing delay at the edge terminal. Through the above method, a task-driven cloud-edge collaborative scheduling optimization strategy for parallel computing tasks across edge nodes is implemented to ensure low-latency operation of spatiotemporal flow data. See the schematic diagram of cross-edge collaborative scheduling for details. Figure 5 , Figure 5 In the formula, t1 and t2 represent different subtask scheduling, Ue represents the edge computing node (i.e., the edge node of the cloud-edge structure), and Uc represents the central computing node (i.e., the central node of the cloud-edge structure). The meaning of the t1 and t2 formulas is the scheduling calculation relationship of the subtask in the edge set between different computing nodes.

[0031] Then, a series of independent real-time computing units of different granularities are constructed and dynamically combined in the cloud-edge collaborative scheduling service chain. The cost loss and efficiency formed in the cross-edge communication and computing process are dynamically estimated and transmitted to the cloud-edge integrated network to construct a series of independent real-time computing units of different granularities. These independent real-time computing units can be dynamically combined in the cloud-edge collaborative scheduling service chain to build a scheduling service chain for cloud-edge scheduling.

[0032] For computing job J, it can be decomposed into a series of computing tasks executed in parallel , where P represents the computing task, represents multiple different computing units. Assuming that the communication cost between tasks is C, if Computational Unit With Computing Unit satisfy , represents an empty set, and the communication cost between two computing units satisfies When , the computational task J can be expressed as , represents the communication cost between two computing units, then the computing job J is called a dynamic scheduling service task chain for independent parallel real-time computing. In the cloud-edge collaborative scheduling service chain, the above-mentioned real-time computing units will be intelligently combined and dynamically adjusted in a modular way. According to the current task allocation strategy and factors such as the network status and device load status obtained in real time, the system will automatically select the most appropriate computing unit and its execution order.

[0033] Finally, the scheduling service chain within the preset time window is automatically executed based on the workflow. According to the resource status, network conditions, and task priority of each node in each time window, the computing task units are intelligently allocated and arranged to ensure that each node can execute efficiently and in parallel within its own time window, thereby maximizing the concurrent processing capabilities of the distributed system and reducing the overall response time. The fine division and optimized combination of the service chain within the time window realizes the seamless collaboration and efficient completion of computing tasks in a multi-node environment at the cloud edge.

[0034] Step S30: Based on the cloud-edge collaborative allocation strategy of global tasks across edge nodes, an adaptive optimization mechanism for cloud-edge resource scheduling that takes into account the global storage, computing and transmission low latency costs is established to achieve adaptive optimization of cloud-edge resource scheduling for spatiotemporal stream computing tasks at the global level.

[0035] Specifically, first, a cloud-edge resource balanced allocation strategy that takes into account the global storage, computing and transmission low latency costs is established. The correlation of real-time computing tasks in spatiotemporal flows is formally described based on the computing task units formed by information at different time points and different interaction modes. On this basis, the dependencies of computing tasks in different time windows on cloud-edge computing resources and the storage, computing and transmission latency costs generated under the current task are mapped to the nodes and edges in the network, and the dynamic calculation of element attributes is performed; a load balancing allocation algorithm driven by latency costs and based on network structure is constructed, and the global resource allocation and system load status of the network are measured through balance evaluation indicators. The cloud-edge computing task allocation and execution are simulated with the dynamic configuration of node attributes and edge attributes, and the task allocation process is converged to the optimal state with the minimum latency cost and maximum resource utilization of a single task as the objective function.

[0036] The load balancing algorithm based on network structure can be divided into the following processes: (1) Obtain the computing resources of the edge node and the computing task requirements it carries, and coordinate the difference between the two to define the current load state of the node, including high load, low load, and balanced state; (2) Evaluate the cost of communication and information transmission between nodes, obtain the edge weights in the network to assist algorithm decision-making, and define the state of the edges; (3) Based on the node attributes and edge attributes, the maximum weight sum matching is used to optimally match the cloud-edge resources and delay costs in the network to obtain a balanced allocation solution that satisfies the minimum delay and maximizes resource utilization.

[0037] Then, a dynamic allocation mechanism of cloud-edge resources for global tasks based on the network structure is established. From the global perspective of the cloud-edge integrated network, the cloud center systematically collects and organizes the key information of all global tasks to be executed, including key parameters such as the computing intensity of the task, the size of the data, the delay cost, and the dependencies between tasks. Subsequently, all edge nodes with available resources are obtained by using the mapping relationship and node attributes, and the hardware configuration, network status, service capabilities and other aspects of each edge node are comprehensively considered. By evaluating the resource quantity and computing efficiency in real time and saving the global tasks in the corresponding edge cluster, each global task is located at the edge node that is most suitable for carrying the task, so as to achieve an accurate match between task requirements and edge computing resources. At the same time, the total amount of resources and the actual computing efficiency of these edge nodes are analyzed and scientifically evaluated in real time to ensure that the tasks can be deployed smoothly and the resources can be optimally utilized. Finally, according to the optimal scheduling strategy, the global tasks are allocated one by one to the cloud center and related edge nodes for processing, completing the full-link management from global task feature analysis, edge node adaptation, edge node screening to specific task allocation, and effectively realizing the resource optimization allocation of global computing tasks in the cloud-edge collaborative environment.

[0038] Finally, a dynamic and iterative cloud-edge resource scheduling adaptive optimization mechanism is established. By capturing and updating various core information in the global scope of the cloud-edge integrated network in real time, including real-time tracking and recording of the status of all global computing tasks, including the progress of task initiation, execution, and completion, as well as the interdependence between tasks, etc., the resource demand changes required for drones to perform tasks at different times and spatial locations can be accurately grasped, and resources can be foreseen and prepared in advance to cope with possible resource demand peaks in the future. Then, based on the cloud-edge network topology and task characteristics, a delay cost prediction model is established to calculate the execution cost of tasks on different edge nodes, and the minimization of global delay cost is taken as the main optimization goal. Subsequently, based on the process-driven dynamic iterative scheduling algorithm, adaptive adjustment and optimization are performed according to the real-time changing task requirements and system resource status, and the task scheduling scheme is adaptively adjusted according to the real-time requirements of different tasks, network transmission delays, computing resource tension and other factors, striving to minimize the overall delay cost of task execution, thereby realizing the cloud-edge resource scheduling adaptive optimization of spatiotemporal flow computing tasks at the global level. The schematic diagram of the cloud-edge resource scheduling adaptive optimization mechanism taking into account the global storage and transmission low delay cost is referenced. Figure 6 .

[0039] The present invention proposes a cloud-edge-terminal graph network structure task calculation method, which maps the cloud-edge resource topology of large-scale UAV supervision tasks, constructs a cloud-edge integrated network model structure, quantifies the computing resources, and provides data support for subsequent collaborative optimization. Then, based on the cloud-edge integrated network model, a cloud-edge integrated collaborative mode for UAV supervision tasks is established. This enables the rapid calculation of UAV supervision spatiotemporal flow data.

[0040] The present invention proposes a cross-edge node collaborative allocation strategy based on network structure. The real-time computing tasks of drones are allocated to edge nodes, and the global tasks are controlled by the cloud. The load of each edge end is analyzed based on the network structure of computing resources. Through cloud coordination, the computing tasks of high-load nodes are unloaded to the edge end with low load, so as to achieve load balancing of edge nodes. Subsequently, efficient task computing is achieved through the dynamic combination of real-time computing units of different granularities at the edge end.

[0041] The present invention proposes an adaptive optimization mechanism for cloud-edge resource scheduling that takes into account global low latency costs, from drone task queues and data stream transmission to computing task allocation based on the cloud-edge-end graph network model, and finally to dynamic scheduling of computing resources to realize an adaptive optimization mechanism for cloud-edge resources with global storage, computing and transmission low latency costs, and realizes adaptive dynamic optimization of the cloud-edge resource scheduling algorithm through drone feedback, further improving resource allocation and computing efficiency.

[0042] Furthermore, if Figure 7 As shown, based on the above-mentioned cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data, the present invention also provides a cloud-edge collaborative computing system for low-altitude UAV supervision spatiotemporal flow data, wherein the cloud-edge collaborative computing system for low-altitude UAV supervision spatiotemporal flow data includes: A cloud-edge integrated collaborative mode establishment module 51 is used to establish a cloud-edge integrated collaborative mode for UAV monitoring tasks based on a cloud-edge network structure, and convert the cloud-edge computing resource topology relationship into a graph network structure; A cloud-edge collaborative allocation strategy construction module 52 is used to construct a cloud-edge collaborative allocation strategy for global tasks across edge nodes that is adapted to the graph network structure based on the node sequence characteristics of the cloud-edge integrated collaborative mode, so as to realize cloud-edge collaborative allocation of global tasks across edge nodes; The cloud-edge resource scheduling optimization module 53 is used to establish a cloud-edge resource scheduling adaptive optimization mechanism that takes into account the global storage, computing and transmission low latency costs based on the cloud-edge collaborative allocation strategy of the global tasks across edge nodes, and realize the cloud-edge resource scheduling adaptive optimization of spatiotemporal stream computing tasks at the global level.

[0043] Furthermore, if Figure 8As shown, based on the cloud-edge collaborative computing method and system for the spatiotemporal flow data of low-altitude UAV supervision, the present invention also provides a terminal accordingly, which includes a processor 10, a memory 20 and a display 30. Figure 8 Only some components of the terminal are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0044] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory of the terminal. In other embodiments, the memory 20 may also be an external storage device of the terminal, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the terminal. Further, the memory 20 may also include both an internal storage unit of the terminal and an external storage device. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code of the installation terminal. The memory 20 may also be used to temporarily store data that has been output or is to be output. In one embodiment, the memory 20 stores a cloud-edge collaborative computing program 40 for low-altitude drone supervision spatiotemporal flow data, and the cloud-edge collaborative computing program 40 for low-altitude drone supervision spatiotemporal flow data can be executed by the processor 10, thereby realizing the cloud-edge collaborative computing method for low-altitude drone supervision spatiotemporal flow data in the present application.

[0045] In some embodiments, the processor 10 can be a central processing unit (CPU), a microprocessor or other data processing chip, which is used to run the program code or process data stored in the memory 20, such as executing the cloud-edge collaborative computing method of the spatiotemporal flow data of the low-altitude drone supervision.

[0046] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 30 is used to display information on the terminal and to display a visual user interface. The processor 10, the memory 20, and the display 30 of the terminal communicate with each other via a system bus.

[0047] In one embodiment, when the processor 10 executes the cloud-edge collaborative computing program 40 for low-altitude UAV supervision spatiotemporal flow data in the memory 20, the steps of the cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data as described above are implemented.

[0048] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a cloud-edge collaborative computing program for low-altitude UAV supervision spatiotemporal flow data, and when the cloud-edge collaborative computing program for low-altitude UAV supervision spatiotemporal flow data is executed by a processor, the steps of the cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data as described above are implemented.

[0049] In summary, the present invention provides a cloud-edge collaborative computing method, system, terminal and computer-readable storage medium for low-altitude UAV supervision spatiotemporal flow data, the method comprising: establishing a cloud-edge integrated collaborative mode for UAV supervision tasks based on a cloud-edge network structure, and converting the cloud-edge computing resource topological relationship into a graph network structure; based on the node sequence characteristics of the cloud-edge integrated collaborative mode, constructing a cloud-edge collaborative allocation strategy for global tasks across edge nodes that adapts to the graph network structure, so as to achieve cloud-edge collaborative allocation of global tasks across edge nodes; based on the cloud-edge collaborative allocation strategy for global tasks across edge nodes, establishing a cloud-edge resource scheduling adaptive optimization mechanism that takes into account the global storage, computing and transmission low latency costs, and achieving cloud-edge resource scheduling adaptive optimization of spatiotemporal flow computing tasks at the global level. The present invention establishes a cloud-edge collaborative optimization scheduling mechanism for large-scale UAV supervision spatiotemporal flow data computing in three-dimensional space, so as to support the efficient scheduling requirements of real-time computing of UAV supervision tasks under the cloud-edge collaborative mode.

[0050] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or terminal including the element.

[0051] Of course, those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the program can be stored in a computer-readable storage medium that can be read by a computer, and the program can include the processes of the above-mentioned method embodiments when executed. The computer-readable storage medium can be a memory, a disk, an optical disk, etc.

[0052] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data, characterized in that: The cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data includes: Establish a cloud-edge integrated collaboration model for drone monitoring tasks based on a cloud-edge network structure, and convert the topological relationship of cloud-edge computing resources into a graph network structure; Based on the node sequence characteristics of the cloud-edge integrated collaboration mode, a cloud-edge collaborative allocation strategy for global tasks across edge nodes that adapts to the graph network structure is constructed to achieve cloud-edge collaborative allocation of global tasks across edge nodes; Based on the cloud-edge collaborative allocation strategy of global tasks across edge nodes, an adaptive optimization mechanism for cloud-edge resource scheduling that takes into account the global storage, computing and transmission low latency costs is established, and adaptive optimization of cloud-edge resource scheduling for spatiotemporal stream computing tasks is achieved at the global level.

2. The cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data according to claim 1 is characterized in that: The cloud-edge integrated collaborative mode for drone monitoring tasks based on the cloud-edge network structure is established, and the topological relationship of cloud-edge computing resources is converted into a graph network structure, specifically including: Map the topological relationship of cloud-edge resources for large-scale drone supervision and build a cloud-edge integrated network model. = (Uc, Ue, V), where Uc represents the central node of the cloud-edge structure, Ue represents the edge node of the cloud-edge structure, and V represents the homogeneous attribute edge set in the cloud-edge integrated network model; Define a set of heterogeneous points (Uc, Ue), Uc = {nc1, nc2, ..., nc n }, represents the cloud center node and related attribute mapping in the cloud edge structure, where nc1 to nc n Represents multiple attribute mappings in the central node, Ue={ne1, ne2, ..., ne n }, indicating edge nodes and related attribute mapping, where ne1 to ne n Represents multiple attribute mappings in edge nodes; node attributes are represented as: ； in, Represents node attributes, including central nodes and edge nodes. Considering the influencing factors comprehensively, Indicates node computing resources, Represents node storage resources, Indicates the node function type, Indicates the current load status or task deployment of the node; Define a homogeneous attribute edge set V: ； in, represents the homogeneity property of the edge set between edge nodes, It represents the transmission cost incurred by the nodes to interact, including the delay in the communication or information transmission process. Indicates the amount of computing tasks migrated by the node during task scheduling. represents the efficiency of information transmission between nodes, Indicates the type of communication or edge connectivity between nodes; Cloud-edge integrated network model The expression is: ； in, Indicates the number of cloud nodes, Expressed as the number of edge nodes, Represents the cloud-cloud subnetwork, which is used to describe the communication status between cloud nodes. represents the edge-edge subnetwork, which is used to characterize the cross-edge task computing process. and Representing cloud-edge heterogeneous bipartite networks with different numbers of nodes, used to characterize the communication process between cloud and edge; Based on the cloud-edge integrated network model, a cloud-edge integrated collaborative mode for drone supervision tasks is established. The structural walk and feature walk analysis methods of heterogeneous node networks are used to integrate the semantic associations of different node types into the network structure for network representation learning, thereby capturing the structural similarity and attribute characteristics of the network.

3. The cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data according to claim 1 is characterized in that: The process of the cloud-edge collaborative allocation strategy includes: allocating local real-time computing tasks to edge clusters and controlling global tasks through the cloud to achieve cloud-edge collaboration; The node sequence characteristics of the cloud-edge integrated collaborative mode are used to construct a cloud-edge collaborative allocation strategy for global tasks across edge nodes that adapts to the graph network structure to achieve cloud-edge collaborative allocation of global tasks across edge nodes, specifically including: Assign local real-time computing tasks to the nearby edge for computing. If the computing resources of the nearby edge cannot meet the requirements of low-latency processing, the cloud will be used to coordinate the tasks and offload the local real-time computing tasks to another nearby edge or cloud for takeover and processing, so as to achieve cross-edge communication and global collaborative optimization. When the local real-time computing tasks transmitted to the edge are processed, the edge transmits the real-time instructions and data to the terminal, and transmits the spatiotemporal data to the cloud, and uses the situation awareness model to perform state awareness and classification processing on the spatiotemporal data; The cloud is used to store big data and train intelligent supervision models for processing various supervision tasks, and regularly distribute the trained intelligent supervision models to each edge node; The cost loss and efficiency generated in the cross-edge communication and computing process are dynamically estimated and transmitted to the cloud-edge integrated network, multiple independent real-time computing units of different granularities are constructed, and multiple independent real-time computing units are dynamically combined in the cloud-edge collaborative scheduling service chain to build a scheduling service chain for cloud-edge scheduling; Based on the workflow, the scheduling service chain within the preset time window is automatically executed. According to the resource status, network conditions and task priority of each node in each time window, the computing task units are intelligently allocated and orchestrated to ensure that the nodes are efficiently executed in parallel within each time window.

4. The cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data according to claim 1 is characterized in that: The establishment of a cloud-edge resource scheduling adaptive optimization mechanism that takes into account the global storage, computing and transmission low latency cost specifically includes: Establish a balanced cloud-edge resource allocation strategy that takes into account the global storage, computing, and transmission low latency costs; Establish a global task cloud-edge resource dynamic allocation mechanism based on network structure; Establish a dynamic and iterative cloud-edge resource scheduling adaptive optimization mechanism.

5. The cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data according to claim 4 is characterized in that: The above-mentioned cloud-edge resource balanced allocation strategy that takes into account the global storage, computing and transmission low latency cost specifically includes: The correlation of real-time computing tasks of spatiotemporal streams is formally described based on computing task units formed by information at different time points and different interaction modes. The dependency of computing tasks in different time windows on cloud-edge computing resources and the storage, computing, and transmission delay costs generated under the current task are mapped to nodes and edges in the network, and dynamic calculation of feature attributes is performed. Construct a network-structure-based load balancing allocation algorithm driven by latency cost, measure the global resource allocation and system load status of the network through balance evaluation indicators, simulate cloud-edge computing task allocation and execution with dynamic configuration of node attributes and edge attributes, and converge the task allocation process to the optimal state with the minimum latency cost of a single task and maximum resource utilization as the objective function; The process of load balancing based on network structure includes: Obtain computing resources of edge nodes and computing task requirements carried by them, and coordinate and define the current load state of the nodes through the difference in computing task requirements, wherein the load state includes high load, low load and balanced state; Evaluate the cost of communication and information transmission between nodes, obtain edge weights in the network to assist algorithm decision-making, and define edge states; Based on node attributes and edge attributes, the maximum weight sum matching is used to optimally match cloud-edge resources and delay costs in the network to obtain a balanced allocation solution that satisfies minimum delay and maximum resource utilization.

6. The cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data according to claim 4 is characterized in that: The establishment of a global task cloud-edge resource dynamic allocation mechanism based on network structure specifically includes: Based on the global perspective of the cloud-edge integrated network, the cloud center systematically collects and organizes the target key information of all global tasks to be executed, including the computing intensity of the task, the amount of data, the delay cost, and the dependencies between tasks; Use mapping relationships and node attributes to obtain all edge nodes with available resources, comprehensively consider the hardware configuration, network status and service capabilities of each edge node, conduct real-time evaluation of resource volume and computing efficiency, save global tasks in the corresponding edge cluster, and locate the most suitable edge node for each global task; Based on the optimal scheduling strategy, global tasks are assigned to the cloud center and related edge nodes for processing one by one, so as to complete the full-link management from global task feature analysis, edge node adaptation, edge node screening to specific task allocation.

7. The cloud-edge collaborative computing method for low-altitude UAV monitoring spatiotemporal flow data according to claim 4 is characterized in that: The establishment of a dynamic iterative cloud-edge resource scheduling adaptive optimization mechanism specifically includes: Capture and update the target core information in the global scope of the cloud-edge integrated network in real time, and track and record the status of all global computing tasks in real time, including the progress of task initiation, execution, and completion, as well as the interdependencies between tasks; Based on the cloud-edge network topology and task characteristics, a delay cost prediction model is established to calculate the execution cost of tasks on different edge nodes, and the minimization of global delay cost is taken as the main optimization goal; Based on the process-driven dynamic iterative scheduling algorithm, it makes adaptive adjustments and optimizations according to the real-time changing task requirements and system resource status, and adaptively adjusts the task scheduling plan according to the real-time requirements of different tasks, network transmission delays and computing resource constraints.

8. A cloud-edge collaborative computing system for low-altitude UAV monitoring spatiotemporal flow data, characterized in that: The cloud-edge collaborative computing system for low-altitude UAV monitoring spatiotemporal flow data includes: The cloud-edge integrated collaborative model establishment module is used to establish a cloud-edge integrated collaborative model for drone supervision tasks based on the cloud-edge network structure, and convert the topological relationship of cloud-edge computing resources into a graph network structure; A cloud-edge collaborative allocation strategy construction module is used to construct a cloud-edge collaborative allocation strategy for global tasks across edge nodes that is adapted to the graph network structure based on the node sequence characteristics of the cloud-edge integrated collaborative mode, so as to realize cloud-edge collaborative allocation of global tasks across edge nodes; The cloud-edge resource scheduling optimization module is used to establish a cloud-edge resource scheduling adaptive optimization mechanism that takes into account the global storage, computing and transmission low latency costs based on the cloud-edge collaborative allocation strategy of the global tasks across edge nodes, and realize the cloud-edge resource scheduling adaptive optimization of spatiotemporal stream computing tasks at the global level.

9. A terminal, characterized in that: The terminal includes: a memory, a processor, and a cloud-edge collaborative computing program for low-altitude UAV supervision spatiotemporal stream data stored in the memory and run on the processor. When the cloud-edge collaborative computing program for low-altitude UAV supervision spatiotemporal stream data is executed by the processor, the steps of the cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal stream data as described in any one of claims 1-7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a cloud-edge collaborative computing program for low-altitude UAV supervision spatiotemporal flow data. When the cloud-edge collaborative computing program for low-altitude UAV supervision spatiotemporal flow data is executed by a processor, it implements the steps of the cloud-edge collaborative computing method for low-altitude UAV supervision spatiotemporal flow data as described in any one of claims 1-7.

Images