Intelligent scheduling and real-time cooperative control method for HarmonyOS industrial equipment

Abstract

The invention relates to the technical field of industrial equipment scheduling, in particular to a HarmonyOS industrial equipment intelligent scheduling and real-time cooperative control method, which comprises the following steps: S1, industrial equipment autonomous behavior design and dynamic role switching based on bionic swarm intelligence; by introducing bionic swarm intelligence, multi-modal perception, transfer learning and block chain technologies, the scheduling efficiency and task execution capability of industrial equipment in a complex dynamic environment are remarkably improved, and in the actual use process, the equipment can dynamically adjust task priority and resource allocation according to the real-time operation state and scene change, so that the service life of the equipment is prolonged. And meanwhile, environment data is collected by utilizing a multi-mode sensor, prediction and reasoning are performed through a machine learning model, and the equipment can pre-judge a fault risk and optimize a task chain execution scheme, so that the fault rate and energy consumption in the production process are reduced, and the production efficiency and stability are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial equipment scheduling, and in particular to a HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method. Background Art

[0002] With the rapid development of intelligent manufacturing, the automation and collaborative control of industrial equipment have become important means to improve production efficiency, reduce costs, and ensure production quality. Traditional industrial equipment scheduling systems mostly rely on rule-driven preset plans, which are difficult to cope with complex and dynamically changing production environments. Especially in scenarios such as multi-tasking, high load, and sudden failures, existing scheduling methods are often unable to flexibly respond to changes, resulting in inefficient use of equipment resources and limited production efficiency. In addition, traditional equipment scheduling systems lack real-time fault prediction and intelligent optimization capabilities, which can easily lead to production interruptions when equipment failures occur, increasing maintenance costs and production downtime.

[0003] Although the existing industrial equipment scheduling system can achieve a certain degree of task allocation and resource management, it still has many shortcomings: 1. Traditional equipment scheduling methods often rely on manually set rules, lack flexible adaptive capabilities, and cannot be dynamically adjusted according to production environment and task changes in real time; 2. When facing complex tasks, the existing scheduling system lacks sufficient fault tolerance and fault prediction capabilities, resulting in the equipment being unable to switch tasks or self-repair in time when under high load or failure, which in turn affects the stability and efficiency of production; 3. Most scheduling systems fail to effectively integrate advanced technologies such as multimodal perception and transfer learning, resulting in insufficient intelligence level and task execution capabilities of the equipment. Therefore, we provide a HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method. Summary of the invention

[0004] In view of the above-mentioned shortcomings of the prior art, the first purpose of the present invention is to provide a HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method to solve the problems in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: A HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method, comprising the following steps: S1. Autonomous behavior design and dynamic role switching of industrial equipment based on bionic swarm intelligence; S2, Multimodal environment perception and scene state deduction and priority analysis based on graph model; S3, Task decomposition, dynamic registration and distributed execution for microservice architecture; S4. Trusted collaboration and dynamic reputation evaluation of edge devices based on blockchain technology; S5, cross-domain knowledge fusion and task reasoning and adaptive solution based on transfer learning; S6, context-aware driven real-time task scheduling and intelligent dynamic control optimization; S7, resource sharing and priority scheduling of device virtualization and multi-task parallel execution; S8, task analysis, real-time takeover and intelligent learning interaction design for enhanced human-machine collaboration.

[0006] The present invention is further configured as follows: in the step S1, autonomous behavior design and dynamic role switching of industrial equipment based on bionic swarm intelligence: S1.1. The working equipment adjusts its own strategy by sensing its own status and the signals of neighboring equipment. The rule setting requires the Bosch BME688 gas sensor and FPGA controller operating environment to support high-frequency data acquisition and real-time calculation; S1.2. In the HarmonyOS distributed system, each device periodically broadcasts feedback signals, including task completion and resource utilization, through the wireless communication module. Neighboring devices modify their behavior strategies based on the received signals, such as dynamically increasing or decreasing resource allocation, and processing feedback data in real time through low-power edge computing nodes. S1.3, the task is decomposed into multi-level subtasks, and each level of tasks is allocated according to the processing capacity of the device. HarmonyOS's lightweight containerization technology is used to support task reconstruction and generate task chain diagrams. The scene data is stored in the edge node database in JSON format for dynamic query; S1.4. Based on the real-time needs of device health status monitoring and task progress, devices can flexibly switch roles in a collaborative network. This is achieved through the task priority allocation interface of HarmonyOS and the relevant instructions for role switching are synchronized using the CAN bus protocol.

[0007] The present invention is further configured as follows: in the step S2, multimodal environment perception and scene state deduction and priority analysis based on the graph model: S2.1. A multimodal perception network consisting of industrial-grade sensors is deployed at the production site to collect visual, thermodynamic and mechanical vibration data. All data is uploaded to the edge computing node through the HarmonyOS unified transmission protocol and compressed in Protobuf format. S2.2. Use graph neural networks and time series models to make multi-dimensional predictions on scenario data, such as predicting the probability of device failure or future workload fluctuations. These models run in edge nodes optimized with TensorFlow Lite, and the training data comes from the device's historical operation logs and sensor data. S2.3. Design a hierarchical model to prioritize tasks based on the complexity of real-time scenarios, and store hierarchical data in a distributed memory database to ensure fast retrieval; S2.4. HarmonyOS is based on a distributed task scheduling module that dynamically adjusts task allocation to adapt to scene changes. In high humidity or high vibration environments, tasks are allocated to devices that are more suitable for the scene. Specific task allocation is implemented through gRPC remote procedure calls.

[0008] The present invention is further configured as follows: in the step S3, task decomposition, dynamic registration and distributed execution oriented to microservice architecture: S3.1. Complex tasks are split into multiple independently executable atomic task units, which are encapsulated in Docker lightweight containers and run on embedded platforms supported by HarmonyOS. S3.2. Each task unit is registered through the distributed service registration center for dynamic invocation by other devices. The HarmonyOS distributed file system is used to store the service directory, and the service metadata is stored in the form of a YAML file. S3.3. Design the task chain through the directed acyclic graph model to ensure that the task units are executed according to the dependency relationship. The task chain is automatically generated in the edge node to support multi-threaded parallel optimization. The task dependency is tracked through Redis streaming data records; S3.4. When a device is overloaded or fails, the task unit is automatically migrated to the backup device through the edge computing node. The unloading and migration are dynamically scheduled by the distributed task engine of HarmonyOS and resources are reallocated after the migration.

[0009] The present invention is further configured as follows: in the step S4, in the trusted collaboration and dynamic reputation evaluation of edge devices based on blockchain technology: S4.1. Deploy a lightweight version of Hyperledger Fabric on the edge computing nodes at the industrial site to record task interactions and execution logs. Each node is connected through a distributed trusted computing interface based on HarmonyOS to form a blockchain network. S4.2. The PBFT consensus algorithm is introduced to verify the credibility of the task execution results. When multiple devices complete subtasks, their execution results need to be verified and recorded through the edge blockchain. The delay of the consensus process is controlled within 50ms. S4.3. The task completion rate, failure rate and other indicators of the device are recorded through blockchain to generate real-time reputation scores. The scoring model is completed using the AI ​​reasoning module based on HarmonyOS, and the calculation results are stored locally on the device to reduce transmission overhead. S4.4, HarmonyOS distributed file system supports the generation of audit logs containing device operation records and synchronizing them with blockchain data. The logs are stored in SHA-256 encrypted format.

[0010] The present invention is further configured as follows: in the step S5, cross-domain knowledge fusion and task reasoning and adaptive solution based on transfer learning: S5.1. By collecting equipment operation data and the experience of domain experts, a multi-dimensional cross-domain knowledge graph is constructed. The knowledge graph is stored in the Neo4j database in RDF format and dynamically updated through the distributed data interface of HarmonyOS. Nodes represent tasks, equipment, and environmental characteristics represent their associations. S5.2. Introduce a transfer learning model based on BERT or GPT structure to reason about unknown tasks, train using industrial task historical data sets, and run on edge devices that support FP16 precision; S5.3. Generate a standardized task solution template based on the inference results. The template is stored in YAML format and loaded into the HarmonyOS scheduling module for real-time call and reuse; S5.4. Use distributed collaborative reasoning technology to ensure that reasoning results are shared between devices and the global reasoning model is updated in real time to adapt to similar task scenarios.

[0011] The present invention is further configured as follows: in the step S6, context-aware driven real-time task scheduling and intelligent dynamic control optimization: S6.1. Extract task context features through a multimodal sensor network. The feature extraction algorithm is based on the lightweight machine learning framework of HarmonyOS to analyze the dynamic environment characteristics of the device in real time during operation. The data is stored in the InfluxDB database in the form of time series. S6.2. Based on the collected situational data, a library containing multiple scheduling strategies is built. Each strategy is logically implemented with a rule engine to support dynamic loading and modification. S6.3. Use the event-driven mechanism of HarmonyOS to match the current situation data with the policy library. The matching logic uses a reasoning model based on fuzzy logic to ensure a smooth transition when multiple situations coexist. S6.4. Design a context transfer detection mechanism based on the Markov model to monitor the changing trend of the context in real time. The detection results are broadcast to all relevant devices through the real-time notification system of HarmonyOS.

[0012] The present invention is further configured as follows: in the step S7, resource sharing and priority scheduling of device virtualization and multi-task parallel execution: S7.1. Use Docker container technology to abstract physical devices into multiple logical units. The virtualization layer is managed by the HarmonyOS distributed scheduling module and communicates with physical devices through the gRPC interface. S7.2. Based on the distributed resource management function of HarmonyOS, a resource sharing model is designed. Task resources are allocated through a time-slice-based scheduling algorithm to avoid competition. The operating environment of the sharing model is a high-throughput heterogeneous computing platform. S7.3. Develop an AI-based task load balancer to dynamically adjust the resource allocation strategy of the device. The balancer logic is implemented through a reinforcement learning algorithm, and the training data comes from the device's operation log and historical scheduling records. S7.4. Introduce a data structure based on priority queue to manage multi-task parallel scheduling. The queue implements a memory pool management module based on HarmonyOS and supports real-time task insertion and deletion operations.

[0013] The present invention is further configured as follows: in the step S8, task analysis, real-time takeover and intelligent learning interaction design for human-machine collaboration enhancement: S8.1. Develop a task parsing module based on natural language processing technology to support operators to issue instructions in voice or text form. The parsing module runs on the language reasoning engine built into HarmonyOS and supports Chinese and multi-language input; S8.2. Use AR glasses to provide real-time task guidance for operators. Task information is rendered as a three-dimensional model through the Unity engine and projected into the field of view of the AR device. The data comes from the distributed status monitoring module of HarmonyOS, and the update frequency is maintained at more than 100Hz. S8.3. Design a collaborative learning module based on deep learning to record human-computer interaction data. The learning model runs on edge devices that support GPU acceleration and is updated once a week. S8.4. Design a manual takeover mechanism during task execution. When the device detects an irrecoverable abnormality, the operator can directly control the device through the smart terminal. The takeover operation is synchronized to the device execution layer in real time through the distributed control protocol of HarmonyOS. Beneficial Effects

[0014] Compared with the known public technology, the technical solution provided by the present invention has the following beneficial effects: 1. The present invention significantly improves the scheduling efficiency and task execution capability of industrial equipment in complex dynamic environments by introducing bionic swarm intelligence, multimodal perception, transfer learning and blockchain technology. In actual use, the equipment can dynamically adjust task priority and resource allocation according to real-time operating status and scene changes to ensure efficient use of equipment resources. At the same time, by using multimodal sensors to collect environmental data and predicting and reasoning through machine learning models, the equipment can predict fault risks and optimize task chain execution plans, thereby reducing the failure rate and energy consumption in the production process and significantly improving production efficiency and stability.

[0015] 2. The present invention adopts a microservice architecture and a distributed collaboration mechanism to enable complex tasks to be flexibly split and dynamically executed, and uses blockchain technology to ensure the credibility of collaboration between devices and the immutability of data. When a device fails or the load is too high, the task unit can automatically migrate to a backup device to avoid production interruptions. Through human-machine collaboration and intelligent learning, the equipment can also continuously optimize its own capabilities and adapt to unknown task scenarios, providing technical guarantees for efficient scheduling and stable operation in complex production environments, and bringing significant economic and operational management benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of a HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] The present invention will be further described below in conjunction with the embodiments. Example

[0019] like Figure 1 As shown, the technical solution provided by the present invention is: a method for intelligent scheduling and real-time collaborative control of HarmonyOS industrial equipment, in which step S1, autonomous behavior design and dynamic role switching of industrial equipment based on bionic swarm intelligence: S1.1, the working equipment adjusts its own strategy by sensing its own status (operating temperature, power) and the signals of neighboring equipment (virtual pheromone concentration). The rule setting requires the Bosch BME688 gas sensor and FPGA controller operating environment to support high-frequency data acquisition and real-time calculation; S1.2. In the HarmonyOS distributed system, each device periodically broadcasts feedback signals, including task completion and resource utilization, through the wireless communication module (Zigbee). Neighboring devices modify their behavior strategies based on the received signals, such as dynamically increasing or decreasing resource allocation. This process processes feedback data in real time through low-power edge computing nodes (NVIDIA Jetson Nano). S1.3, tasks are decomposed into multi-level subtasks. Tasks at each level are allocated in detail according to the processing capabilities of the device (CPU frequency, memory size). HarmonyOS's lightweight containerization technology (Huawei DevEcoService Framework) is used to support task reconstruction and generate task chain diagrams. Scenario data is stored in the edge node database (SQLite) in JSON format for dynamic query; S1.4. Based on the real-time requirements of device health status monitoring (MTBF indicator) and task progress, the device can flexibly switch roles (leader node or slave node) in the collaborative network. This function is implemented through the task priority allocation interface of HarmonyOS, and the relevant instructions for role switching are synchronized with the CAN bus protocol; In the step S2, multimodal environment perception and scene state deduction and priority analysis based on the graph model: S2.1. A multimodal perception network consisting of industrial-grade sensors (FLIR A400 thermal imaging cameras and SIKO laser rangefinders) is deployed at the production site to collect visual, thermodynamic and mechanical vibration data. All data is uploaded to the edge computing node through the HarmonyOS unified transmission protocol and compressed in Protobuf format. S2.2. Use graph neural networks (GNNs) and time series models (LSTMs) to make multi-dimensional predictions on scenario data, such as predicting the probability of device failure or future workload fluctuations. These models run in edge nodes optimized with TensorFlow Lite, and the training data comes from the device's historical operation logs and sensor data. S2.3. Design a hierarchical model to prioritize tasks based on the complexity of real-time scenarios. For example, high-priority tasks (high temperature alarms) are handled by dedicated high-performance devices (industrial PLC controllers), while low-priority tasks are postponed or assigned to backup devices. Hierarchical data is stored in a distributed memory database (Redis) to ensure fast retrieval. S2.4. HarmonyOS is based on a distributed task scheduling module to dynamically adjust task allocation to adapt to scenario changes. For example, in a high humidity or high vibration environment, tasks are allocated to devices that are more suitable for the scenario (industrial robotic arms protected by IP65 casings). Specific task allocation is achieved through gRPC remote procedure calls.

[0020] In this embodiment, a dynamically adjusted industrial equipment scheduling framework is constructed by introducing bionic swarm intelligence and multimodal environmental perception technology. The equipment makes autonomous strategy adjustments through adaptive rules (virtual pheromone mechanism based on ant colony algorithm) and obtains real-time scene data in combination with a distributed environmental perception network. The perception data is used for multi-dimensional scenario deduction and fault prediction through graph neural networks to determine task priorities and dynamically assign tasks. The equipment can also switch roles (leader node or slave node) according to task requirements to ensure the efficiency and flexibility of task execution. Example

[0021] like Figure 1 As shown, the technical solution provided by the present invention is: a method for intelligent scheduling and real-time collaborative control of HarmonyOS industrial equipment, in which step S3, task decomposition, dynamic registration and distributed execution for microservice architecture: S3.1. Complex tasks are divided into multiple independently executable atomic task units. For example, equipment maintenance tasks can be divided into "temperature monitoring", "lubrication inspection" and "mechanical adjustment" subtasks, which are completed by dedicated sensors or submodules respectively. The task units are encapsulated in Docker lightweight containers and run on embedded platforms supported by HarmonyOS. S3.2. Each task unit is registered through the distributed service registration center (Eureka) for dynamic invocation by other devices. The HarmonyOS distributed file system (DFS) is used to store the service directory, and the service metadata is stored in the form of a YAML file. S3.3. Design the task chain through the directed acyclic graph (DAG) model to ensure that the task units are executed according to the dependency relationship. The task chain is automatically generated in the edge node to support multi-threaded parallel optimization. The task dependency is tracked through Redis streaming data records; S3.4. When a device is overloaded or fails, the task unit automatically migrates to the backup device (ARM Cortex A72 platform) through the edge computing node. The unloading and migration are dynamically scheduled by the HarmonyOS distributed task engine (DTask), and resources are reallocated after the migration. In the step S4, trusted collaboration and dynamic reputation evaluation of edge devices based on blockchain technology: S4.1. Deploy a lightweight version of Hyperledger Fabric on the edge computing nodes at the industrial site to record task interactions and execution logs. Each node is connected through a distributed trusted computing interface based on HarmonyOS to form a blockchain network. S4.2. Introduce the PBFT consensus algorithm to verify the credibility of task execution results. For example, when multiple devices complete subtasks, their execution results need to be verified and recorded through the edge blockchain, and the delay of the consensus process is controlled within 50ms; S4.3. The task completion rate, failure rate and other indicators of the device are recorded through blockchain to generate real-time reputation scores. The scoring model is completed using the AI ​​reasoning module based on HarmonyOS, and the calculation results are stored locally on the device to reduce transmission overhead. S4.4, HarmonyOS distributed file system supports the generation of audit logs containing device operation records and synchronizing them with blockchain data. The logs are stored in SHA-256 encrypted format.

[0022] In this embodiment, at the task decomposition and execution level, a microservice architecture is used to split complex tasks into multiple atomic units. These units are dynamically registered to a distributed service center and run in the form of lightweight containers. During the distributed execution process, blockchain technology is used to record the collaboration process between devices to ensure that the data cannot be tampered with. At the same time, the real-time reputation value of the device is evaluated. The task chain is generated by DAG (directed acyclic graph). Task units are dynamically migrated under device load balancing and status monitoring to ensure execution efficiency and security. The blockchain mechanism (PBFT consensus algorithm) further enhances the credibility of collaboration between devices. Example

[0023] like Figure 1 As shown, the technical solution provided by the present invention is: a method for intelligent scheduling and real-time collaborative control of HarmonyOS industrial equipment, in which step S5, cross-domain knowledge fusion and task reasoning and adaptive solution based on transfer learning: S5.1. By collecting equipment operation data (equipment type, task parameters, external environment data) and the experience of domain experts, a multi-dimensional cross-domain knowledge graph is constructed. The knowledge graph is stored in the Neo4j database in RDF (Resource Description Framework) format and dynamically updated through the distributed data interface of HarmonyOS. Nodes represent tasks, equipment, environmental characteristics, etc., and represent their associations (for example, "task requirements adapt to device capabilities"); S5.2. Introduce a transfer learning model based on BERT or GPT structure to reason about unknown tasks. For example, when the device encounters an unlabeled task type, the model infers the best solution through the similarity of known tasks and environments. The training uses industrial task history datasets and runs on edge devices that support FP16 precision (NVIDIA Jetson Xavier NX). S5.3. Generate a standardized task solution template based on the inference results. For example, the template may include "task decomposition strategy", "priority rule" and "device call order". The template is stored in YAML format and loaded into the HarmonyOS scheduling module for real-time call and reuse; S5.4. Use distributed collaborative reasoning technology (Federated Learning) to ensure that reasoning results are shared and updated in real time between devices. For example, when a device completes an unknown task, the reasoning mechanism broadcasts the execution record to other devices, updates the global reasoning model, and adapts to similar task scenarios. In the step S6, context-aware driven real-time task scheduling and intelligent dynamic control optimization: S6.1. Extract task context features through a multimodal sensor network (temperature, vibration, and gas composition detection). The feature extraction algorithm is based on HarmonyOS's lightweight machine learning framework (MindSpore Lite) to analyze the dynamic environmental characteristics of the device in real time during operation. The data is stored in the InfluxDB database in the form of time series. S6.2. Based on the collected situational data, a library containing multiple scheduling strategies is built, for example, "high temperature scenario priority cooling equipment scheduling strategy" or "equipment load balancing priority strategy". Each strategy is logically implemented with a rule engine (Drools) to support dynamic loading and modification; S6.3. Use the event-driven mechanism of HarmonyOS to match the current situation data with the policy library. For example, when a high temperature vibration scenario is detected, the system automatically calls the preset cooling device priority strategy and readjusts the task allocation through the distributed task engine. The matching logic uses a fuzzy logic-based reasoning model to ensure a smooth transition when multiple scenarios coexist. S6.4. Design a context transition detection mechanism based on the Markov model to monitor the changing trend of the context in real time. For example, when the context changes from a “low-load normal environment” to a “high-load failure risk environment”, the system triggers policy adjustments in advance by predicting the transition probability to avoid interruptions. The detection results are broadcast to all relevant devices through the real-time notification system of HarmonyOS. In the step S7, resource sharing and priority scheduling of device virtualization and multi-task parallel execution: S7.1. Use Docker container technology to abstract physical devices into multiple logical units. For example, an industrial robot arm can be virtualized into a "grasping unit", "moving unit" and "sensing unit". The virtualization layer is managed by the HarmonyOS distributed scheduling module and communicates with the physical device through the gRPC interface. S7.2. Based on the distributed resource management function of HarmonyOS, a resource sharing model is designed. Task resources (CPU, memory, I / O channels) are allocated through a time slice-based scheduling algorithm to avoid competition. For example, while the device is completing a grasping task, another virtual unit can perform a sensor calibration task in parallel. The shared model runs on a high-throughput heterogeneous computing platform (ARM + FPGA architecture). S7.3. Develop an AI-based task load balancer to dynamically adjust the resource allocation strategy of devices. For example, when the load of a device reaches 80%, the balancer automatically offloads some tasks to neighboring devices. The balancer logic is implemented through a reinforcement learning algorithm, and the training data comes from the device's operation log and historical scheduling records. S7.4. Introduce a data structure based on priority queues to manage multi-task parallel scheduling. For example, high-priority tasks enter the front queue (priority response time is about 5ms), and low-priority tasks enter the back queue. The queue implements a memory pool management module based on HarmonyOS and supports real-time task insertion and deletion operations. In the step S8, task analysis, real-time takeover and intelligent learning interaction design for human-machine collaboration enhancement: S8.1. Develop a task parsing module based on natural language processing (NLP) technology to support operators to issue commands in voice or text form. For example, when an operator inputs "check the gripping status of the robot arm", the system parses the task semantics through the NLP model and maps it to specific device operation commands. The parsing module runs on the language inference engine built into HarmonyOS and supports Chinese and multi-language input; S8.2. Use AR glasses (Microsoft HoloLens 2) to provide real-time task guidance for operators. Task information (equipment status, task progress) is rendered as a three-dimensional model through the Unity engine and projected into the field of view of the AR device. The data comes from the distributed status monitoring module of HarmonyOS, and the update frequency is maintained at more than 100Hz. S8.3. Design a collaborative learning module based on deep learning to record human-machine interaction data (manual intervention operation records) to optimize the autonomous capabilities of the equipment. For example, when the operator adjusts the path planning of the robotic arm multiple times, the system automatically learns the adjustment mode and adaptively applies it in future tasks. The learning model runs on edge devices that support GPU acceleration and is updated once a week. S8.4. Design a manual takeover mechanism during task execution. When the device detects an irrecoverable abnormality, the operator can directly control the device through the smart terminal (industrial-grade tablet computer). The takeover operation is synchronized to the device execution layer in real time through the distributed control protocol of HarmonyOS.

[0024] In this embodiment, transfer learning and knowledge graph technology are used to provide the device with cross-domain task reasoning capabilities, and a dynamic scheduling strategy is designed through a context-aware mechanism. In scheduling optimization, physical devices are abstracted into logical units that can be executed in parallel through virtualization technology to achieve deep resource utilization and task priority management. Human-machine collaboration enhancement provides an interactive interface for operators through natural language processing modules and augmented reality technology, and records operating behaviors to optimize the device's autonomous capabilities. The combination of context awareness, task reasoning and human-machine collaboration enables the system to have real-time adjustment and efficient execution capabilities in complex dynamic environments.

[0025] like Figure 1 As shown, in the actual use of the present invention, each device adjusts the working strategy based on preset adaptive rules by sensing its own operating status and the signals of neighboring devices, and dynamically switches roles (such as leader node and slave node). The device broadcasts status information regularly and processes feedback data in real time through low-power edge computing nodes, thereby accurately adjusting resource allocation and task priority to ensure flexible response to various task requirements in complex production environments. This dynamic scheduling mechanism effectively improves the flexibility and efficiency of task execution.

[0026] In order to further improve the intelligence level of the system, this method combines multimodal sensor networks with advanced machine learning technology. Environmental data is collected in real time through thermal imaging cameras, laser rangefinders and other devices. These data are analyzed in multiple dimensions by graph neural networks and time series models, which can timely predict equipment failures or workload fluctuations, and perform task reasoning based on transfer learning models. When encountering unknown tasks, the system can automatically infer the best solution through knowledge graphs and historical data, and optimize task decomposition and resource allocation. This intelligent reasoning capability enables the system to respond efficiently in a rapidly changing production environment and ensure efficient operation of the equipment.

[0027] At the task execution level, the present invention adopts microservice architecture and blockchain technology to ensure the distributed management of task units and trusted collaboration between devices. Complex tasks are divided into multiple atomic task units, dynamically registered through the distributed service registration center, and automatically migrated to backup devices when the device load is high or a failure occurs, to ensure the continuity and security of task execution. Blockchain technology ensures the immutability of the task execution process and the credibility of the data. At the same time, the edge computing nodes schedule tasks in real time, further improving the operating efficiency and stability of the system. Through this series of technical integration, the system can achieve efficient and flexible task scheduling and device collaboration in a dynamic and complex environment, significantly improving production efficiency and ensuring equipment reliability.

[0028] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method, characterized in that: The following steps are involved: S1. Autonomous behavior design and dynamic role switching of industrial equipment based on bionic swarm intelligence; S2, Multimodal environment perception and scene state deduction and priority analysis based on graph model; S3, Task decomposition, dynamic registration and distributed execution for microservice architecture; S4. Trusted collaboration and dynamic reputation evaluation of edge devices based on blockchain technology; S5, cross-domain knowledge fusion and task reasoning and adaptive solution based on transfer learning; S6, context-aware driven real-time task scheduling and intelligent dynamic control optimization; S7, resource sharing and priority scheduling of device virtualization and multi-task parallel execution; S8, task analysis, real-time takeover and intelligent learning interaction design for enhanced human-machine collaboration.

2. According to claim 1, a HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method is characterized in that: In the step S1, autonomous behavior design and dynamic role switching of industrial equipment based on bionic swarm intelligence: S1.

1. The working equipment adjusts its own strategy by sensing its own status and the signals of neighboring equipment. The rule setting requires the Bosch BME688 gas sensor and FPGA controller operating environment to support high-frequency data acquisition and real-time calculation; S1.

2. In the HarmonyOS distributed system, each device periodically broadcasts feedback signals, including task completion and resource utilization, through the wireless communication module. Neighboring devices modify their behavior strategies based on the received signals, such as dynamically increasing or decreasing resource allocation, and processing feedback data in real time through low-power edge computing nodes. S1.3, the task is decomposed into multi-level subtasks, and each level of tasks is allocated according to the processing capacity of the device. HarmonyOS's lightweight containerization technology is used to support task reconstruction and generate task chain diagrams. The scene data is stored in the edge node database in JSON format for dynamic query; S1.

4. Based on the real-time needs of device health status monitoring and task progress, devices can flexibly switch roles in a collaborative network. This is achieved through the task priority allocation interface of HarmonyOS and the relevant instructions for role switching are synchronized using the CAN bus protocol.

3. According to claim 1, a HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method is characterized in that: In the step S2, multimodal environment perception and scene state deduction and priority analysis based on the graph model: S2.

1. A multimodal perception network consisting of industrial-grade sensors is deployed at the production site to collect visual, thermodynamic and mechanical vibration data. All data is uploaded to the edge computing node through the HarmonyOS unified transmission protocol and compressed in Protobuf format. S2.

2. Use graph neural networks and time series models to make multi-dimensional predictions on scenario data, such as predicting the probability of device failure or future workload fluctuations. These models run in edge nodes optimized with TensorFlow Lite, and the training data comes from the device's historical operation logs and sensor data. S2.

3. Design a hierarchical model to prioritize tasks based on the complexity of real-time scenarios, and store hierarchical data in a distributed memory database to ensure fast retrieval; S2.

4. HarmonyOS is based on a distributed task scheduling module that dynamically adjusts task allocation to adapt to scene changes. In high humidity or high vibration environments, tasks are allocated to devices that are more suitable for the scene. Specific task allocation is implemented through gRPC remote procedure calls.

4. According to claim 1, a HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method is characterized by: In the step S3, task decomposition, dynamic registration and distributed execution for microservice architecture: S3.

1. Complex tasks are split into multiple independently executable atomic task units, which are encapsulated in Docker lightweight containers and run on embedded platforms supported by HarmonyOS. S3.

2. Each task unit is registered through the distributed service registration center for dynamic invocation by other devices. The HarmonyOS distributed file system is used to store the service directory, and the service metadata is stored in the form of a YAML file. S3.

3. Design the task chain through the directed acyclic graph model to ensure that the task units are executed according to the dependency relationship. The task chain is automatically generated in the edge node to support multi-threaded parallel optimization. The task dependency is tracked through Redis streaming data records; S3.

4. When a device is overloaded or fails, the task unit is automatically migrated to the backup device through the edge computing node. The unloading and migration are dynamically scheduled by the distributed task engine of HarmonyOS and resources are reallocated after the migration.

5. According to claim 1, a HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method is characterized by: In the step S4, trusted collaboration and dynamic reputation evaluation of edge devices based on blockchain technology: S4.

1. Deploy a lightweight version of Hyperledger Fabric on the edge computing nodes at the industrial site to record task interactions and execution logs. Each node is connected through a distributed trusted computing interface based on HarmonyOS to form a blockchain network. S4.

2. The PBFT consensus algorithm is introduced to verify the credibility of the task execution results. When multiple devices complete subtasks, their execution results need to be verified and recorded through the edge blockchain. The delay of the consensus process is controlled within 50ms. S4.

3. The task completion rate, failure rate and other indicators of the device are recorded through blockchain to generate real-time reputation scores. The scoring model is completed using the AI ​​reasoning module based on HarmonyOS, and the calculation results are stored locally on the device to reduce transmission overhead. S4.4, HarmonyOS distributed file system supports the generation of audit logs containing device operation records and synchronizing them with blockchain data. The logs are stored in SHA-256 encrypted format.

6. A HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method according to claim 1, characterized in that: In the step S5, cross-domain knowledge fusion and task reasoning and adaptive solution based on transfer learning: S5.

1. By collecting equipment operation data and the experience of domain experts, a multi-dimensional cross-domain knowledge graph is constructed. The knowledge graph is stored in the Neo4j database in RDF format and dynamically updated through the distributed data interface of HarmonyOS. Nodes represent tasks, equipment, and environmental characteristics represent their associations. S5.

2. Introduce a transfer learning model based on BERT or GPT structure to reason about unknown tasks, train using industrial task historical data sets, and run on edge devices that support FP16 precision; S5.

3. Generate a standardized task solution template based on the inference results. The template is stored in YAML format and loaded into the HarmonyOS scheduling module for real-time call and reuse; S5.

4. Use distributed collaborative reasoning technology to ensure that reasoning results are shared between devices and the global reasoning model is updated in real time to adapt to similar task scenarios.

7. A HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method according to claim 1, characterized in that: In the step S6, context-aware driven real-time task scheduling and intelligent dynamic control optimization: S6.

1. Extract task context features through a multimodal sensor network. The feature extraction algorithm is based on the lightweight machine learning framework of HarmonyOS to analyze the dynamic environment characteristics of the device in real time during operation. The data is stored in the InfluxDB database in the form of time series. S6.

2. Based on the collected situational data, a library containing multiple scheduling strategies is built. Each strategy is logically implemented with a rule engine to support dynamic loading and modification. S6.

3. Use the event-driven mechanism of HarmonyOS to match the current situation data with the policy library. The matching logic uses a reasoning model based on fuzzy logic to ensure a smooth transition when multiple situations coexist. S6.

4. Design a context transfer detection mechanism based on the Markov model to monitor the changing trend of the context in real time. The detection results are broadcast to all relevant devices through the real-time notification system of HarmonyOS.

8. A HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method according to claim 1, characterized in that: In the step S7, resource sharing and priority scheduling of device virtualization and multi-task parallel execution: S7.

1. Use Docker container technology to abstract physical devices into multiple logical units. The virtualization layer is managed by the HarmonyOS distributed scheduling module and communicates with physical devices through the gRPC interface. S7.

2. Based on the distributed resource management function of HarmonyOS, a resource sharing model is designed. Task resources are allocated through a time-slice-based scheduling algorithm to avoid competition. The operating environment of the sharing model is a high-throughput heterogeneous computing platform. S7.

3. Develop an AI-based task load balancer to dynamically adjust the resource allocation strategy of the device. The balancer logic is implemented through a reinforcement learning algorithm, and the training data comes from the device's operation log and historical scheduling records. S7.

4. Introduce a data structure based on priority queue to manage multi-task parallel scheduling. The queue implements a memory pool management module based on HarmonyOS and supports real-time task insertion and deletion operations.

9. A HarmonyOS industrial equipment intelligent scheduling and real-time collaborative control method according to claim 1, characterized in that: In the step S8, task analysis, real-time takeover and intelligent learning interaction design for human-machine collaboration enhancement: S8.

1. Develop a task parsing module based on natural language processing technology to support operators to issue instructions in voice or text form. The parsing module runs on the language reasoning engine built into HarmonyOS and supports Chinese and multi-language input; S8.

2. Use AR glasses to provide real-time task guidance for operators. Task information is rendered as a three-dimensional model through the Unity engine and projected into the field of view of the AR device. The data comes from the distributed status monitoring module of HarmonyOS, and the update frequency is maintained at more than 100Hz. S8.

3. Design a collaborative learning module based on deep learning to record human-computer interaction data. The learning model runs on edge devices that support GPU acceleration and is updated once a week. S8.

4. Design a manual takeover mechanism during task execution. When the device detects an irrecoverable abnormality, the operator can directly control the device through the smart terminal. The takeover operation is synchronized to the device execution layer in real time through the distributed control protocol of HarmonyOS.

Images