Edge task processing method and device, electronic equipment and readable storage medium

Abstract

The invention provides an edge task processing method and device, electronic equipment and a readable storage medium, and relates to the field of computers, in particular to the technical fields of cloud computing, edge computing, cloud-edge collaboration, the Internet of Things, large models and the like. According to the specific implementation scheme, in response to received prompt data of a user for a target task, based on the prompt data, a task instruction of at least one sub-task of the target task and an edge node corresponding to the sub-task are determined; and transferring the task instruction of the subtask to the corresponding edge node, and enabling the edge node to perform task processing based on the task instruction. In the scheme, the task instruction of the subtask and the edge node corresponding to the subtask are determined by analyzing the prompt data of the user, and the edge node performs task processing based on the task instruction, so that effective processing of the complex task is realized.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, in particular to technical fields such as cloud computing, edge computing, cloud-edge collaboration, the Internet of Things, and large models. Specifically, the present disclosure relates to an edge task processing method, device, electronic device, and readable storage medium. Background Art

[0002] With the rapid development of edge computing technology, the edge supports more and more services, which puts higher demands on the edge's ability to handle complex tasks.

[0003] How to enable the edge to effectively process complex tasks has become an important technical issue in this field. Summary of the Invention

[0004] In order to solve at least one of the above-mentioned deficiencies, the present disclosure provides an edge task processing method, device, electronic device and readable storage medium.

[0005] According to a first aspect of the present disclosure, a method for processing edge tasks is provided, the method comprising:

[0006] In response to receiving prompt data for a target task from a user, determining, based on the prompt data, a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask;

[0007] The task instructions of the subtask are transferred to the corresponding edge node, and the edge node performs task processing based on the task instructions.

[0008] According to a second aspect of the present disclosure, there is provided an edge task processing device, the device comprising:

[0009] a subtask splitting module, configured to, in response to receiving prompt data of a target task from a user, determine, based on the prompt data, a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask;

[0010] The subtask processing module is used to transfer the task instructions of the subtask to the corresponding edge node, and enable the edge node to perform task processing based on the task instructions.

[0011] According to a third aspect of the present disclosure, an electronic device is provided, including:

[0012] at least one processor; and

[0013] A memory communicatively connected to the at least one processor; wherein,

[0014] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the edge task processing method.

[0015] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable a computer to execute the above-mentioned edge task processing method.

[0016] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program, which implements the above-mentioned edge task processing method when executed by a processor.

[0017] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are provided to facilitate a better understanding of the present invention and do not constitute a limitation of the present disclosure.

[0019] Figure 1 This is a flow chart of an edge task processing method provided by an embodiment of the present disclosure;

[0020] Figure 2 is a schematic diagram of the structure of the edge computing system provided by an embodiment of the present disclosure;

[0021] Figure 3 is a structural diagram of an edge task processing device provided by an embodiment of the present disclosure;

[0022] Figure 4 4 is a block diagram of an electronic device used to implement the edge task processing method of an embodiment of the present disclosure. DETAILED DESCRIPTION

[0023] The following description of exemplary embodiments of the present disclosure is made in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding. These details should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0024] With the rapid development of edge computing technology, the number of services supported by the edge is increasing, placing higher demands on the edge's ability to handle complex tasks. Enabling the edge to effectively handle these complex tasks has become a key technical issue in this field.

[0025] In related technologies, when processing complex tasks at the edge, there are generally the following defects:

[0026] (1) Complex tasks at the edge generally involve a large number of task instructions. Related technologies generally require users to enter task instructions multiple times, or to manage task instructions through multiple applications, which causes inconvenience to users and requires high user skills, making it easy for operational errors to occur.

[0027] (2) With the widespread application of IoT devices in edge environments, the edge generally involves more IoT devices when performing complex tasks. Users usually need to use different applications or interfaces to control various types of IoT devices, which causes inconvenience to users. At the same time, in the edge computing environment, tasks involving multiple IoT devices may require switching between different applications, affecting the smoothness of task execution and making task failures more likely. In addition, complex tasks may require the coordinated operation of multiple IoT devices during execution. Related technologies usually require users to manually control IoT devices, and lack a mechanism for automated control of the coordinated operation of IoT devices.

[0028] (3) In related technologies, there may be cloud-edge separation scenarios. In such scenarios, when the model is deployed on the edge, the operation of updating the model is complicated, and the maintenance cost after the model is started is high. If an edge computing system with cloud-edge collaboration is introduced, the deployment and maintenance of the model can be easily realized.

[0029] (4) When processing complex tasks, the edge end in related technologies has insufficient ability to understand the user's natural language. Users usually need to operate according to specific command syntax, which affects the convenience of user use and the efficiency of interaction with users.

[0030] (5) In related technologies, the control of the edge when processing complex tasks mostly relies on the cloud. When the network quality between the edge and the cloud is poor, interaction delays may occur, affecting the effective execution of the task.

[0031] (6) In related technologies, edge computing systems are usually unable to understand the status of IoT devices and the environmental data of the environment in which the IoT devices are located in real time, which is not conducive to the effective execution of tasks.

[0032] In addition, the mainstream edge computing systems in related technologies have not been effectively combined with large language models (LLMs). If the large language models can be effectively combined with edge computing systems, more effective natural language understanding and data analysis can be performed based on the large language models, better supporting the processing of complex tasks.

[0033] The edge task processing method, device, electronic device, and readable storage medium provided by the embodiments of the present disclosure are intended to solve at least one of the above technical problems in the prior art.

[0034] Figure 1 A flow chart of an edge task processing method provided by an embodiment of the present disclosure is shown as follows: Figure 1 As shown in , the method may mainly include:

[0035] Step S110: In response to receiving prompt data for a target task from a user, determining a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask based on the prompt data;

[0036] Step S120: Transfer the task instruction of the subtask to the corresponding edge node, and enable the edge node to perform task processing based on the task instruction.

[0037] Prompt data is user-provided data. Users can submit prompt data to indicate to the edge computing system the target task they wish to perform. Prompt data can be multimodal, including but not limited to natural language text data, images, videos, and audio data. Prompt data can also include environmental data collected by IoT devices.

[0038] In the disclosed embodiment, the user may submit prompt data when initiating a task command, and the prompt data is then transferred to the edge computing system for subsequent analysis and processing.

[0039] As an example, the edge can be deployed with a user interaction module for user interaction. The user can issue a task command to the edge based on the user interaction module. As another example, the user can issue a task command to the cloud, which then sends the task command to the edge based on the edge's communication link.

[0040] In the disclosed embodiment, the prompt data can reflect the user's intention, and the target task is the task that the user wants the edge computing system to perform, as expressed in the user's intention. The target task can be a complex task composed of a series of subtasks.

[0041] In the disclosed embodiment, task analysis can be performed based on prompt data to determine multiple subtasks, and each subtask can be represented as a task instruction. These subtasks can be understood as tasks that are split from the target task and can be executed independently.

[0042] For example, the prompt data is "need to drink coffee", and the subtasks split based on the prompt data are coffee machine and user positioning, robot path planning, coffee temperature collection and other operations.

[0043] Each edge node in the edge cluster has its own functional responsibilities. While analyzing the subtasks, it is also possible to analyze the edge nodes suitable for processing the subtasks, that is, the edge nodes corresponding to the subtasks.

[0044] In an embodiment of the present disclosure, after determining the task instructions of the subtask and the edge node corresponding to the subtask, the task instructions of the subtask can be circulated within the edge cluster, so that the task instructions of the subtask are transferred to the corresponding edge node, and the edge node corresponding to the subtask performs task processing based on the task instructions.

[0045] The method provided by the disclosed embodiments responds to user prompt data regarding a target task and, based on the prompt data, determines the task instructions for at least one subtask of the target task and the corresponding edge node for the subtask. The task instructions for the subtask are then transferred to the corresponding edge node, which then processes the task based on the task instructions. This solution effectively handles complex tasks by analyzing the user prompt data, determining the task instructions for the subtask and the corresponding edge node, and then having the edge node process the task based on the task instructions.

[0046] In the embodiment of the present disclosure, after the user provides prompt data, the edge computing system can automatically split the subtasks and automatically execute the subtasks. There is no need for the user to enter task instructions multiple times, nor is there a need to use multiple applications to manage task instructions. This facilitates user use and effectively avoids operational errors caused by complex operations.

[0047] In the disclosed embodiments, when edge nodes perform task processing based on subtask instructions, they may invoke execution devices within the IoT environment. The edge nodes can interact with these execution devices, sending the subtask instructions to the execution devices, which then execute the subtask instructions, thereby completing the execution of the subtask.

[0048] In the embodiment of the present disclosure, after the execution device completes the execution of the subtask, the task execution result can be fed back to the corresponding edge node.

[0049] In the disclosed embodiments, efficient processing of target tasks is achieved by splitting complex tasks into multiple subtasks and scheduling corresponding edge nodes to drive execution devices to perform the subtasks. This solution eliminates the need for users to use different applications or interfaces to control various types of IoT devices, making it easier for users to use them. Furthermore, the lack of application switching makes the execution of complex tasks smoother and enables effective collaboration between IoT devices.

[0050] In an optional manner of the present disclosure, determining, based on prompt data, a task instruction of at least one subtask of a target task and an edge node corresponding to the subtask includes:

[0051] Based on the prompt data and the node description information of each edge node in the edge cluster, a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask are determined.

[0052] The edge computing system provided by the embodiments of the present disclosure may include a cloud, an edge end, and a device end. The edge end is deployed with an edge cluster, which is composed of multiple edge nodes. The device end may include multiple execution devices in an Internet of Things scenario.

[0053] In the embodiments of the present disclosure, the node description information may include the description information of the application module in the edge node and the software and hardware information. The description information of the application module may include the name, version, instance deployment status, etc. The software and hardware information may include the container version, node resource consumption, etc.

[0054] Based on the node description information, the edge nodes that are compatible with the subtasks can be effectively analyzed to ensure the reasonable scheduling of the subtasks.

[0055] Based on the node description information, subtasks can be scheduled to corresponding edge nodes. For example, the subtask is robot path planning. The node description information of an edge node reflects that the edge node is deployed with an application module that interacts with the robot and can control the robot. Then, the subtask of robot path planning can be scheduled to the edge node, so that the edge node controls the robot to perform corresponding path planning operations.

[0056] As an example, a node agent can be deployed on each edge node. The node agent collects the node description information of the node to which it belongs and transfers the node description information of the node to the edge computing core module in the edge computing system. The edge computing core module uniformly manages the node description information.

[0057] In an optional manner of the present disclosure, determining a task instruction of at least one subtask of a target task and an edge node corresponding to the subtask based on prompt data and node description information of each edge node in an edge cluster includes:

[0058] Call the preset artificial intelligence model to determine the task instructions of at least one subtask of the target task and the edge node corresponding to the subtask based on the prompt data and the node description information of each edge node in the edge cluster.

[0059] In the disclosed embodiment, an artificial intelligence model may be called and its semantic understanding capability may be used to analyze prompt data and node description information of each edge node in the edge cluster, generate task instructions for subtasks, and determine the edge nodes corresponding to the subtasks.

[0060] In an optional embodiment of the present disclosure, the artificial intelligence model is a large language model deployed on a target edge node in an edge cluster.

[0061] In the disclosed embodiment, the artificial intelligence model may adopt a large language model, which has better natural language understanding and data analysis capabilities, and can better support the processing of complex tasks.

[0062] In this case, a large language model can be used to analyze prompt data and effectively understand the intention reflected in the user prompt data, thereby generating task instructions for subtasks. This eliminates the need for users to manually enter task instructions according to specific command syntax, making it more convenient for users to use and improving the interaction efficiency between the edge computing system and users.

[0063] In the disclosed embodiment, by deploying a large language model at the edge, the edge has powerful decision-making capabilities, enabling the edge to support the processing of complex tasks, greatly reducing dependence on the cloud, and even when the network quality between the edge and the cloud is poor, the edge can independently process tasks, avoiding interaction delays caused by reliance on cloud decisions and ensuring the effective execution of tasks.

[0064] In an optional manner of the present disclosure, determining a task instruction of at least one subtask of a target task and an edge node corresponding to the subtask based on prompt data and node description information of each edge node in an edge cluster includes:

[0065] Calling the large language model to determine the environmental data requirement information based on the prompt data;

[0066] Acquire environmental data based on environmental data demand information;

[0067] The large language model is called to generate at least one subtask of the target task based on the environment data, prompt data, and functional description information of each node in the edge cluster, and determine the edge node corresponding to the subtask.

[0068] In the disclosed embodiment, the prompt data may reflect the need for certain environmental data. The large language model can be used to analyze the environmental data requirement information. The environmental data requirement information can indicate the environmental data required in the process of generating the subtask. The environmental data can be obtained based on the environmental data requirement information.

[0069] As an example, the prompt information is "Help me get a piece of clothing". Based on the prompt data, the large language model can analyze the environmental data requirement information as "need to obtain the temperature of the user's environment", so that the temperature data of the user's environment can be obtained based on the environmental data requirement information, and a subtask can be generated according to the temperature data. Specifically, according to the temperature data, it can be analyzed that the target clothing suitable for the temperature needs to be taken, thereby generating a subtask of obtaining the target clothing.

[0070] In the disclosed embodiment, by acquiring environmental data, the large language model can jointly analyze and process environmental data, prompt data, and functional description information of each node in the edge cluster, thereby improving the accuracy of task instructions of subtasks and edge nodes corresponding to the subtasks.

[0071] In an optional manner of the present disclosure, the subtask is executed by an execution device, and the task instruction matches the driving protocol of the execution device.

[0072] In the embodiment of the present disclosure, the task instructions of the subtask can be matched with the driving protocol of the execution device, that is, the large language model generates task instructions based on the instruction specifications defined in the driving protocol of the execution device, so that the task instructions of the subtask can be quickly executed after being transferred to the execution device.

[0073] In an optional embodiment of the present disclosure, the above method further includes:

[0074] In response to receiving the model resources corresponding to the large language model sent from the cloud, the large language model is deployed on the target edge node based on the model resources.

[0075] In the disclosed embodiment, the large model can be deployed on a target edge node in an edge cluster. The model resources corresponding to the large language model can be sent from the cloud, and then the large language model can be deployed on the target edge node.

[0076] When the large language model needs to be updated, the corresponding model resources can also be sent from the cloud, and then the large language model deployed in the target edge node can be updated.

[0077] In the disclosed embodiments, the deployment, management, and update of the model can be achieved collaboratively between the cloud and the edge, making it convenient for users to deploy and maintain the model.

[0078] In an optional manner of the present disclosure, transferring the task instructions of the subtask to the corresponding edge node includes:

[0079] Packing the task instructions based on a pre-specified first communication protocol to obtain a first communication message;

[0080] The first communication message is published, so that the edge node corresponding to the subtask subscribes to the first communication message and parses the first communication message to obtain the task instruction.

[0081] In an embodiment of the present disclosure, when a task instruction circulates within an edge cluster, the task instruction can be packaged based on a pre-specified first communication protocol to obtain a first communication message, and then the first communication message can be published. The edge node obtains the first communication message by subscription and can then parse the task instruction.

[0082] The task instructions are packaged based on a pre-specified first communication protocol, that is, the task instructions are encapsulated as a whole into a first communication message, so that the task instructions can be obtained by parsing the first communication message.

[0083] As an example, the first communication protocol may be Message Queuing Telemetry Transport (MQTT).

[0084] As an example, a message relay module and an edge soft gateway module can be deployed in the edge computing system. The message relay module can relay the above-mentioned first communication message to the edge soft gateway module, and the edge soft gateway module can publish the first communication message so that the edge node corresponding to the subtask can subscribe to the first communication message.

[0085] In the disclosed embodiment, the task instructions corresponding to the subtasks can match the driving protocol of the execution device. To facilitate the circulation of the task instructions within the edge cluster, the task instructions can be packaged into a first communication message, and the first communication message can be circulated within the edge cluster. This facilitates the edge node corresponding to the subtask to quickly subscribe to the corresponding first communication message and parse the task instructions from the first communication message, thereby achieving rapid circulation of the task instructions within the edge cluster.

[0086] In an optional manner of the present disclosure, enabling an edge node to perform task processing based on a task instruction includes:

[0087] Packing the task instructions based on a pre-specified second communication protocol to obtain a second communication message;

[0088] The second communication message is sent to a corresponding execution device, so that the execution device parses the second communication message to obtain a task instruction and executes the task instruction.

[0089] In the disclosed embodiments, when a task instruction is sent from an edge node to an execution device in an IoT environment, the edge node may package the task instruction based on a pre-specified second communication protocol to generate a second communication message, and then send the second communication message to the corresponding execution device. Upon receiving the second communication message, the execution device may parse the second communication message to obtain the task instruction and execute the task instruction.

[0090] The task instructions are packaged based on a pre-specified second communication protocol, that is, the task instructions are encapsulated as a whole into a second communication message, so that the task instructions can be obtained by parsing the second communication message.

[0091] As an example, the second communication protocol may be Google Remote Procedure Call Protocol (gRPC). Using the gRPC protocol can facilitate the execution device in the Internet of Things to quickly receive the second communication protocol sent by the edge node.

[0092] In the disclosed embodiments, the task instructions corresponding to the subtasks can be matched with the driver protocol of the execution device. To facilitate the edge node to quickly send the task instructions to the execution device in the Internet of Things, the task instructions can be packaged into a second communication message, so that the edge node sends the second communication message to the execution device. After receiving the second communication message, the execution device can quickly parse the task instructions and execute the subtask in a timely manner.

[0093] In an optional manner of the present disclosure, prompt data is submitted by the user to the cloud and sent from the cloud to the edge cluster.

[0094] In the embodiment of the present disclosure, a console can be deployed on the cloud, and users interact with the console, submitting task commands to the cloud through the console and submitting prompt data at the same time. After the cloud receives the prompt data, it can send the prompt data to the edge cluster, which enables users to remotely call the edge computing system to execute edge tasks.

[0095] Figure 2 The following is a schematic diagram of the structure of an edge computing system provided by an embodiment of the present disclosure. The edge computing system may mainly include:

[0096] The console is used to interact with users. Users can initiate requests to the cloud through the console to complete functions such as node creation, large model management, application binding, command issuance, and response data reception.

[0097] Edge computing cloud services in the cloud are used to implement service interfaces, node management, application management, configuration item management, model management, and data synchronization services.

[0098] Node management manages edge node information. Model management manages large language models. Application management manages application resources at edge nodes, including node creation, application generation, node updates, and command forwarding. Data synchronization services synchronize resource information at the edge.

[0099] The LLM model center is used to implement functions such as model training, model iteration, model distribution, model optimization and storage of large language models.

[0100] The edge includes: an edge computing core module (core), an LLM model service module, a message relay module, a user interaction module, an edge soft gateway and device driver module, a general application module (not shown), a related resource management module (not shown), and a data storage module (not shown). The edge computing core module is responsible for collecting edge system information and regularly communicating with the cloud to synchronize resource data, communicating with the resource management module to submit resources for deployment, and communicating with the data storage module to access data. The resource management module is responsible for receiving resource deployment information and deploying resources. The LLM service module is used to obtain user input information (i.e., the aforementioned prompt data), analyze and decompose it, obtain task instructions for subtasks, and then publish the task instructions for subtasks to the message relay module. The edge soft gateway interacts with devices through device drivers, reports device status upward, and assigns subtasks to specific devices for execution downward. The general application module is responsible for executing established business logic.

[0101] In the above edge computing system, the complete process from node creation to user interaction is as follows:

[0102] Step S210: The user initiates an operation through the console to create a node resource in the cloud;

[0103] Step S220: The user operates through the console to select a large model that matches the current application scenario, generates application resources for the selected large model, and associates the application resources with node resources.

[0104] Step S230: The user operates through the console and selects a suitable driver according to the actual device situation to complete the configuration of the edge soft gateway and associate it with the node resources.

[0105] Step S240: The user obtains the node installation command through console operations and executes the installation in the edge cluster.

[0106] Step S250: After executing the installation command, the edge side will start the edge computing core module (Core), and the edge computing core module will automatically establish a communication link with the cloud for data synchronization.

[0107] Step S260: The edge computing core module periodically communicates with the cloud through the established link to collect and report local data and synchronize cloud application information.

[0108] Step S270: The edge computing core module downloads relevant resources based on the data synchronized from the cloud and starts applications such as the LLM model service module, user interaction module, edge soft gateway and device driver.

[0109] Step S280: After the edge soft gateway and device driver are started, they will establish a connection with the corresponding device according to the configuration in step S2230.

[0110] Step S290: The user interacts with the user interaction module through dialogue on the edge side.

[0111] Step S2100: The user interaction module performs basic processing on the received prompt data and publishes it to the LLM model service module through the message transfer module.

[0112] Step S2110: The LLM model service module further cleans and analyzes the prompt data, splits the target tasks expressed in the prompt data according to semantics, and generates task instructions for subtasks.

[0113] Step S2120: The LLM model service module publishes the task instruction to the edge soft gateway through the message transfer module. After subscribing to the task instruction, the device driver distributes the task instruction to the specific execution device.

[0114] Step S2130: After receiving the instruction, the execution device executes the task instruction, completes the subtask and feeds back the result.

[0115] Step S2140: The edge soft gateway obtains the execution result of the subtask through the device driver module, and periodically collects and reports the status data of the device.

[0116] Step S2150: The reported data will be sent to the cloud service or other third-party data access platform according to the destination address configured by the user, and the user can view the data content on the corresponding platform.

[0117] Optionally, in step S210, when the user establishes a node in the cloud, some system modules will be included by default, such as the edge computing core module (Core), the message transfer module (Broker), the rule engine module (Rule), etc., which are used for end-cloud communication after deployment, edge-side module status acquisition, communication between modules on the device side, message routing, etc.

[0118] Optionally, in step S220, the model repository is part of the cloud-based model management service and is responsible for model construction, optimization, storage, and other functions. User management of models in the cloud includes, but is not limited to, model optimization, model acceleration, model training iterations, model compression, and operations such as adding, deleting, modifying, and querying models.

[0119] Optionally, in step S230, the driver applications available to the user include but are not limited to applications encapsulated by general protocols provided by the edge computing framework, such as Modbus drivers, OLE for Process Control Unified Architecture (OPC-UA) drivers for process control, and application modules developed and encapsulated using custom driver protocols.

[0120] Optionally, in step S240, when deploying the edge computing program on the edge end, installation and deployment may be completed by, but not limited to, online installation, offline installation package, image burning, etc.;

[0121] Optionally, in step S250, after the edge computing program device is normally deployed, the edge computing core module that will be started and the cloud service that establishes the data channel adopt the application layer protocol, including but not limited to Web Socket Protocol (WebSocket), Hypertext Transfer Protocol (HTTP), Message Queuing Telemetry Transport Protocol (MQTT), Hypertext Transfer Protocol 3 (HTTP3) and other technologies.

[0122] Optionally, in step S260, the information collected by the edge computing core module includes but is not limited to the information name, version, instance deployment status of all application modules, and the software and hardware information of all nodes in the cluster, such as container version, node resource consumption status, etc.

[0123] Optionally, in step S270, the content synchronized from the cloud by the edge computing core module includes the applications and related configuration data created in steps S210, 220, and 230, as well as subsequent resource change operations in the cloud including but not limited to adding new application modules or configurations, updating existing applications or configuration information, and deleting change data brought about by existing applications or configurations.

[0124] Optionally, in step S280, the link between the driver and the device may be established based on existing industry protocol specifications, such as Modbus protocol, Building Automation and Control Networks Data Communication Protocol (Bacnet), etc., or based on a user-defined protocol;

[0125] Optionally, in step S290, the data interacted between the user and the edge device interaction module includes but is not limited to voice, text, pictures, videos, temperature and humidity sensors, smoke sensors, etc., and supports the transmission and interaction of rich text.

[0126] Optionally, in S2100, the user interaction module can pre-process the acquired prompt data, and the processing methods include but are not limited to: text cleaning, word segmentation, stop word removal, word embedding, padding and truncation for natural language; scaling and cropping, feature extraction and normalization, data enhancement, object detection for images, frame sampling, frame processing, time series processing, optical flow analysis for videos, cleaning of sensor data, timestamp alignment, feature engineering, data interpolation, etc.

[0127] Optionally, in S2110, the large language model further processes the data in ways that include but are not limited to: syntactic and lexical analysis, sentiment analysis, topic modeling, sequence labeling, automatic task process extraction, relationship extraction, subtask identification, task priority allocation, resource allocation, process review, operation implementation generation, task report analysis, anomaly detection and recovery, etc.

[0128] Optionally, in S2120, the large language model can send task instructions to the edge soft gateway through a message relay module or directly, and can use multiple protocols such as MQTT, HTTP, WebScoket, and GRPC to complete task release. In the process of converting task instructions into protocol instructions in the edge soft gateway, the processing methods required for the task instruction data include but are not limited to: data encoding and formatting, protocol data conversion, data verification and security checking, error handling, and task context parameter transmission.

[0129] Optionally, in S2140, the edge soft gateway can periodically synchronize the execution results or status data, or actively obtain the data in response to the cloud's acquisition request to the edge side. The data reported by the edge soft gateway can be obtained using the actual acquisition equipment or sensor equipment in the system. The data reported by the edge soft gateway includes image data, video stream data, temperature, humidity, wind speed, network environment data, etc.

[0130] Optionally, in S2150, users can view the reported information directly through a third-party platform, or access the third-party platform for viewing using the cloud service of the edge computing framework, or directly view the data through the local console of the edge computing.

[0131] The system provided by the embodiments of the present disclosure can effectively guarantee the privacy of user data and the security of device control, ensuring that only authorized users can perform operations, thereby providing security protection.

[0132] The various parts of the edge computing system provided by the embodiments of the present disclosure are not limited to platforms and systems, support multiple implementations, and can be deployed on various devices across platforms.

[0133] Based on the edge computing system and edge task processing method provided by the embodiments of the present disclosure, efficient task automation and collaborative operation can be achieved in edge computing application scenarios. Specifically, the cloud service manages the iteration of large language models to ensure model updates and performance optimization. Cloud node application management can dynamically configure and monitor applications on each edge node. The end-cloud data synchronization mechanism ensures timely data transmission and synchronization to support real-time feedback and environmental perception. The edge-side LLM model service module is responsible for processing natural language input, interpreting user intentions, and breaking down tasks into executable subtasks, generating task instructions for subtasks. Finally, these task instructions are issued to specific execution devices through the edge soft gateway for execution, and the execution results are collected and returned to the cloud, achieving efficient task collaborative operation and real-time feedback. The integration of these functions will provide users with a more intelligent, convenient and secure experience of edge device control and task automation collaborative operation.

[0134] based on Figure 1 The same principle as shown in the method, Figure 3 A schematic diagram of the structure of an edge task processing device provided by an embodiment of the present disclosure is shown. Figure 3 As shown, the edge task processing device 30 may include:

[0135] The subtask splitting module 310 is configured to, in response to receiving prompt data of a target task from a user, determine a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask based on the prompt data;

[0136] The subtask processing module 320 is configured to transfer the task instructions of the subtask to the corresponding edge node, and enable the edge node to perform task processing based on the task instructions.

[0137] The device provided in the disclosed embodiments responds to user prompt data regarding a target task and, based on the prompt data, determines the task instructions for at least one subtask of the target task and the corresponding edge node for the subtask. The device then transfers the subtask task instructions to the corresponding edge node, and causes the edge node to perform task processing based on the task instructions. This solution effectively handles complex tasks by analyzing the user prompt data, determining the subtask task instructions and the corresponding edge node, and causing the edge node to perform task processing based on the task instructions.

[0138] Optionally, when determining the task instruction of at least one subtask of the target task and the edge node corresponding to the subtask based on the prompt data, the subtask splitting module is specifically configured to:

[0139] Based on the prompt data and the node description information of each edge node in the edge cluster, a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask are determined.

[0140] Optionally, when determining the task instruction of at least one subtask of the target task and the edge node corresponding to the subtask based on the prompt data and the node description information of each edge node in the edge cluster, the subtask splitting module is specifically configured to:

[0141] Call the preset artificial intelligence model to determine the task instructions of at least one subtask of the target task and the edge node corresponding to the subtask based on the prompt data and the node description information of each edge node in the edge cluster.

[0142] Optionally, the artificial intelligence model is a large language model deployed on a target edge node in an edge cluster.

[0143] Optionally, when determining the task instruction of at least one subtask of the target task and the edge node corresponding to the subtask based on the prompt data and the node description information of each edge node in the edge cluster, the subtask splitting module is specifically configured to:

[0144] Calling the large language model to determine the environmental data requirement information based on the prompt data;

[0145] Acquire environmental data based on environmental data demand information;

[0146] The large language model is called to generate at least one subtask of the target task based on the environment data, prompt data, and functional description information of each node in the edge cluster, and determine the edge node corresponding to the subtask.

[0147] Optionally, the subtask is executed by an execution device, and the task instruction matches the driving protocol of the execution device.

[0148] Optionally, the above device further includes:

[0149] The large model deployment module is used to deploy the large language model on the target edge node based on the model resources received from the cloud in response to receiving the model resources corresponding to the large language model.

[0150] Optionally, when transferring the task instruction of the subtask to the corresponding edge node, the subtask processing module is specifically configured to:

[0151] Packing the task instructions based on a pre-specified first communication protocol to obtain a first communication message;

[0152] The first communication message is published, so that the edge node corresponding to the subtask subscribes to the first communication message and parses the first communication message to obtain the task instruction.

[0153] Optionally, when the subtask processing module enables the edge node to perform task processing based on the task instruction, it is specifically used to:

[0154] Packing the task instructions based on a pre-specified second communication protocol to obtain a second communication message;

[0155] The second communication message is sent to a corresponding execution device, so that the execution device parses the second communication message to obtain a task instruction and executes the task instruction.

[0156] Optionally, the prompt data includes at least one of the following:

[0157] Natural language text data;

[0158] Image data;

[0159] Video data;

[0160] Audio data.

[0161] Optionally, the prompt data is submitted by the user to the cloud and sent from the cloud to the edge cluster.

[0162] It is understandable that the above modules of the edge task processing device in the embodiment of the present disclosure have the function of realizing Figure 1 The functions of the corresponding steps of the edge task processing method in the embodiment shown in . This function can be implemented by hardware, or by hardware executing the corresponding software implementation. The hardware or software includes one or more modules corresponding to the above functions. The above modules can be software and / or hardware, and the above modules can be implemented separately or integrated with multiple modules. For the functional description of each module of the above edge task processing device, please refer to Figure 1 The corresponding description of the edge task processing method in the embodiment shown in will not be repeated here.

[0163] In the technical solutions disclosed herein, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0164] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0165] The electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the edge task processing method provided in the embodiment of the present disclosure.

[0166] Compared to existing technologies, this electronic device, in response to receiving user prompt data regarding a target task, determines, based on the prompt data, the task instructions for at least one subtask of the target task and the corresponding edge node for the subtask; transfers the subtask task instructions to the corresponding edge node, and enables the edge node to process the task based on the task instructions. This solution effectively handles complex tasks by analyzing the user prompt data, determining the subtask task instructions and the corresponding edge node, and enabling the edge node to process the task based on the task instructions.

[0167] The readable storage medium is a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute the edge task processing method provided in the embodiment of the present disclosure.

[0168] Compared to existing technologies, this readable storage medium, in response to receiving user prompt data for a target task, determines, based on the prompt data, the task instructions for at least one subtask of the target task and the corresponding edge node for the subtask; transfers the subtask task instructions to the corresponding edge node, and causes the edge node to perform task processing based on the task instructions. This solution effectively handles complex tasks by analyzing user prompt data, determining the task instructions for the subtasks and the corresponding edge nodes, and causing the edge node to perform task processing based on the task instructions.

[0169] The computer program product includes a computer program, which, when executed by a processor, implements the edge task processing method provided in the embodiment of the present disclosure.

[0170] Compared to existing technologies, this computer program product, in response to receiving user prompt data for a target task, determines, based on the prompt data, a task instruction for at least one subtask of the target task and the corresponding edge node for the subtask; transfers the subtask task instruction to the corresponding edge node, and causes the edge node to perform task processing based on the task instruction. This solution effectively handles complex tasks by analyzing user prompt data, determining the subtask task instruction and the corresponding edge node for the subtask, and causing the edge node to perform task processing based on the task instruction.

[0171] Figure 4 A schematic block diagram of an example electronic device 40 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are provided as examples only and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0172] like Figure 4 As shown, the electronic device 40 includes a computing unit 410, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 420 or a computer program loaded from a storage unit 480 into a random access memory (RAM) 430. Various programs and data required for the operation of the device 40 can also be stored in the RAM 430. The computing unit 410, the ROM 420, and the RAM 430 are connected to each other via a bus 440. An input / output (I / O) interface 450 is also connected to the bus 440.

[0173] Various components in device 40 are connected to I / O interface 450, including an input unit 460, such as a keyboard, mouse, etc.; an output unit 470, such as various types of displays, speakers, etc.; a storage unit 480, such as a magnetic disk, optical disk, etc.; and a communication unit 490, such as a network card, modem, wireless communication transceiver, etc. Communication unit 490 allows device 40 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0174] The computing unit 410 can be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 410 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 410 performs the edge task processing method provided in the embodiments of the present disclosure. For example, in some embodiments, the execution of the edge task processing method provided in the embodiments of the present disclosure can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 480. In some embodiments, part or all of the computer program can be loaded and / or installed on the device 40 via the ROM 420 and / or the communication unit 490. When the computer program is loaded into the RAM 430 and executed by the computing unit 410, one or more steps of the edge task processing method provided in the embodiments of the present disclosure can be performed. Alternatively, in other embodiments, the computing unit 410 can be configured to perform the edge task processing method provided in the embodiments of the present disclosure by any other appropriate means (e.g., by means of firmware).

[0175] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system comprising at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0176] The program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device so that when the program code is executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0177] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in conjunction with an instruction execution system, device or equipment. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0178] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0179] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0180] A computer system may include a client and a server. The client and server are generally remote from each other and typically interact through a communication network. The client-server relationship arises through computer programs running on the respective computers and having a client-server relationship with each other. The server may be a cloud server, a server in a distributed system, or a server integrated with a blockchain.

[0181] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved. This is not limited herein.

[0182] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.

Claims

1. A method for processing edge tasks, comprising: In response to receiving prompt data for a target task from a user, determining, based on the prompt data, a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask; The task instruction of the subtask is transferred to the corresponding edge node, and the edge node is enabled to perform task processing based on the task instruction.

2. The method according to claim 1, wherein: The determining, based on the prompt data, a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask includes: Based on the prompt data and the node description information of each edge node in the edge cluster, a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask are determined.

3. The method according to claim 2, wherein: The determining, based on the prompt data and the node description information of each edge node in the edge cluster, a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask includes: A preset artificial intelligence model is called to determine, based on the prompt data and the node description information of each edge node in the edge cluster, a task instruction of at least one subtask of the target task and the edge node corresponding to the subtask.

4. The method according to claim 3, wherein: The artificial intelligence model is a large language model deployed on a target edge node in the edge cluster.

5. The method according to claim 4, wherein: The determining, based on the prompt data and the node description information of each edge node in the edge cluster, a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask includes: Calling the large language model to determine environmental data requirement information based on the prompt data; Based on the environmental data demand information, acquiring environmental data; The large language model is called to generate at least one subtask of the target task based on the environment data, the prompt data, and functional description information of each node in the edge cluster, and an edge node corresponding to the subtask is determined.

6. The method according to claim 4 or 5, wherein: The subtask is executed by an execution device, and the task instruction matches a driving protocol of the execution device.

7. The method according to any one of claims 4 to 6, wherein: Also includes: In response to receiving the model resources corresponding to the large language model sent from the cloud, the large language model is deployed on the target edge node based on the model resources.

8. The method according to any one of claims 1 to 7, wherein: The transferring the task instruction of the subtask to the corresponding edge node includes: Packing the task instruction based on a pre-specified first communication protocol to obtain a first communication message; The first communication message is published, so that the edge node corresponding to the subtask subscribes to the first communication message, and parses the first communication message to obtain the task instruction.

9. The method according to any one of claims 1 to 8, wherein: The enabling the edge node to perform task processing based on the task instruction includes: Packing the task instructions based on a pre-specified second communication protocol to obtain a second communication message; The second communication message is sent to a corresponding execution device, so that the execution device parses the second communication message to obtain the task instruction and executes the task instruction.

10. The method according to any one of claims 1 to 9, wherein: The prompt data includes at least one of the following: Natural language text data; Image data; Video data; Audio data.

11. The method according to any one of claims 1 to 10, wherein: The prompt data is submitted by the user to the cloud and sent from the cloud to the edge cluster.

12. An edge task processing device, comprising: A subtask splitting module, configured to, in response to receiving prompt data of a target task from a user, determine a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask based on the prompt data; The subtask processing module is used to transfer the task instruction of the subtask to the corresponding edge node, and enable the edge node to perform task processing based on the task instruction.

13. The device according to claim 12, wherein: When the subtask splitting module determines the task instruction of at least one subtask of the target task and the edge node corresponding to the subtask based on the prompt data, it is specifically used to: Based on the prompt data and the node description information of each edge node in the edge cluster, a task instruction of at least one subtask of the target task and an edge node corresponding to the subtask are determined.

14. The device according to claim 13, wherein: When the subtask splitting module determines the task instruction of at least one subtask of the target task and the edge node corresponding to the subtask based on the prompt data and the node description information of each edge node in the edge cluster, the subtask splitting module is specifically used to: A preset artificial intelligence model is called to determine, based on the prompt data and the node description information of each edge node in the edge cluster, a task instruction of at least one subtask of the target task and the edge node corresponding to the subtask.

15. The device according to claim 14, wherein: The artificial intelligence model is a large language model deployed on a target edge node in the edge cluster.

16. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 11.

17. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-11.

18. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 11.

Images