Internet of Things equipment collaborative management system based on intelligent edge

Abstract

The invention relates to the technical field of Internet of Things, in particular to an Internet of Things equipment collaborative management system based on an intelligent edge, which is composed of an equipment abstraction layer, a converged communication layer, a service logic layer and a firmware upgrading layer, the equipment abstraction layer realizes standardized access and full life cycle management of equipment through modeling, interface abstraction and resource pool management; the converged communication layer comprises protocol adaptation, intelligent routing and QoS guarantee to ensure smooth multi-protocol communication; the business logic layer adopts a micro-service architecture, and flexible business processing is realized through flow arrangement, intelligent decision making and event driving; the firmware upgrading layer comprises state perception, intelligent analysis, intelligent execution and firmware management so as to adaptively select an optimal upgrading strategy according to network state prediction and evaluation. According to the invention, modules at all levels cooperate, intelligent and collaborative management of Internet of Things equipment is realized, and the equipment management efficiency and the operation stability are improved.

Description

Technical Field

[0001] The present invention relates to the field of Internet of Things technology, and in particular to an Internet of Things device collaborative management system based on intelligent edge. Background Art

[0002] With the deep integration of IoT and AI technologies, intelligent upgrades in multi-source, heterogeneous equipment collaboration, such as the Industrial Internet of Things, face a core conflict between efficient management and seamless control. Existing systems have significant pain points, necessitating technological innovation to overcome these bottlenecks.

[0003] Currently, there's a lack of unified IoT device communication standards. Numerous device manufacturers utilize different communication protocols, requiring protocol conversions that lead to high latency and increased failure rates. The diverse device interface types require traditional IoT devices to connect to multiple external conversion modules, increasing system complexity and latency. Existing solutions lack unified device state modeling and event-driven mechanisms, making it impossible to achieve cross-device collaboration and real-time dynamic device adjustments. Summary of the Invention

[0004] To this end, the present invention provides an IoT device collaborative management system based on intelligent edge, thereby solving the aforementioned problems existing in the prior art.

[0005] To achieve the above objectives, the present invention provides an IoT device collaborative management system based on intelligent edge, comprising:

[0006] Device abstraction layer, used to connect to various peripheral devices to achieve abstract management and resource pooling of the peripheral devices;

[0007] A fusion communication layer connected to the device abstraction layer to achieve adaptation and communication of different protocols;

[0008] The business logic layer is connected to the integrated communication layer to implement specific business logic and functions;

[0009] The firmware upgrade layer is connected to the device abstraction layer and the fusion communication layer respectively to implement the upgrade of device firmware;

[0010] The device abstraction layer includes a device modeling module, an interface abstraction module, a resource pool management module, a device description module and a cycle management module;

[0011] The converged communication layer includes a plug-in management module, a protocol adaptation module, a routing guarantee module, an intelligent routing module, a QoS guarantee module and a queue mapping module;

[0012] The business logic layer includes a microservices partitioning module, an architecture building module, a process orchestration module, an intelligent decision-making module, and an event-driven module;

[0013] The firmware upgrade layer includes a state perception module, a policy decision module, a differential execution module and a firmware management module.

[0014] Furthermore, the device abstraction layer includes:

[0015] The device modeling module uses the device model to predefine various peripheral devices that can be connected to the system in the form of metaclasses, and creates a standardized model for each device;

[0016] an interface abstraction module, connected to the device modeling module, and performing standardization processing on the device driver interface according to the standardization model to obtain a standardized interface;

[0017] A resource pool management module, connected to the interface abstraction module, for centrally managing the resources of the device through the fusion resource pool according to the standardized interface to obtain a management result;

[0018] The device description module is used to describe the device information using a unified device description language to obtain a description result;

[0019] The cycle management module is connected to the device modeling module, the interface abstraction module, the resource pool management module, and the device description module respectively, and performs full life cycle management of the device according to the standardized model, the standardized interface, the management results, and the description results.

[0020] Furthermore, the cycle management module includes:

[0021] The device connection unit is used to record the connection status and related basic information of the device into the standardized model to obtain the connection result;

[0022] an initialization configuration unit connected to the device connection unit, performing initial configuration on the device according to the standardized model and device description information to obtain a configuration result;

[0023] An operation monitoring unit, for monitoring the operation status of the equipment in real time, and dynamically adjusting the monitoring frequency and monitoring items according to the historical operation data recorded in the management result to obtain the monitoring result;

[0024] a status evaluation unit, connected to the operation monitoring unit, for evaluating the health status and performance level of the equipment according to the monitoring results to obtain an evaluation result;

[0025] The fault repair unit is connected to the status evaluation unit and is used to perform corresponding repair or recovery operations according to the evaluation result and the fault handling strategy in the management result, and update the device status information in the management result.

[0026] Furthermore, the fusion communication layer includes:

[0027] Plug-in management module, used to load and manage independent plug-ins of various communication protocols;

[0028] The protocol adaptation module is connected to the protocol plug-in loading management module, and adapts, diverts and analyzes data of different protocols according to the loaded protocol plug-in to obtain the adaptation result;

[0029] A message conversion module, connected to the protocol adaptation module, converts the adaptation result into a standardized message format, and defines a standardized data packet structure to encapsulate and convert the data to obtain a conversion result;

[0030] an intelligent routing module, connected to the message conversion module, for using the conversion result to dynamically select an optimal transmission path according to node status and link quality to obtain a routing result;

[0031] QoS guarantee module, which is used to classify and schedule data according to the type and priority of business data to obtain guarantee results;

[0032] The queue mapping module is connected to the intelligent routing module and the QoS guarantee module respectively, and maps different service flow characteristics to different queues according to the guarantee result and the routing result to obtain mapping results.

[0033] Furthermore, the intelligent routing module includes:

[0034] A path evaluation unit, configured to collect and analyze in real time the status information of each node in the network and the quality indicators of the links to obtain collection results, and to evaluate the collection results to obtain evaluation results;

[0035] a routing decision unit, connected to the path evaluation unit, configured to use a routing algorithm to select an optimal transmission path according to the evaluation result to obtain a decision result;

[0036] The dynamic adjustment unit is connected to the routing decision unit and is used to perform dynamic adjustment according to the decision result when the network conditions change.

[0037] Furthermore, the business logic layer includes:

[0038] The microservice division module divides business rules into several microservices according to business needs, with each microservice corresponding to each domain boundary;

[0039] An architecture building module, connected to the microservice partitioning module, for building a microservice architecture according to the domain boundaries;

[0040] A process orchestration module is connected to the architecture building module and introduces a visual business process orchestration engine according to the microservice architecture to obtain an orchestration result;

[0041] An intelligent decision-making module, connected to the process orchestration module, performs intelligent judgment on the decision points in the process according to the orchestration results and preset rules to obtain a judgment result;

[0042] The event-driven module is connected to the process orchestration module and the intelligent decision-making module respectively, and is used to monitor business events and adjust the process according to the judgment results.

[0043] Furthermore, the intelligent decision-making module includes:

[0044] Rule management unit, used to define, store and update decision rules to obtain management results;

[0045] A decision engine unit, connected to the rule management unit, for judging the decision points in the business process according to the management results to obtain a decision result;

[0046] The feedback adjustment unit is connected to the decision engine unit and is used to evaluate the decision result to obtain an evaluation result, and adjust the decision rules and processes according to the evaluation result.

[0047] Furthermore, the firmware upgrade layer includes:

[0048] A state perception module is used to capture network state data in real time, including physical layer indicators, transport layer indicators, and application layer multi-dimensional indicators;

[0049] an intelligent analysis module, connected to the state perception module, for performing predictions using a prediction algorithm based on the physical layer indicators, the transport layer indicators, and the application layer multi-dimensional indicators to generate an adaptive decision matrix;

[0050] an intelligent execution module, connected to the intelligent analysis module, and selecting an upgrade strategy according to the decision matrix to execute the firmware upgrade to obtain an execution result;

[0051] The firmware management module is connected to the intelligent execution module and is responsible for generating and applying the difference package when the intelligent execution module selects the differential upgrade strategy.

[0052] Furthermore, the intelligent analysis module includes:

[0053] Data preprocessing unit, which cleans, normalizes and extracts features from the collected data to obtain processing results;

[0054] a state prediction unit, connected to the data preprocessing unit, and using a long short-term memory network to predict the network state according to the processing result to obtain a prediction result;

[0055] a comprehensive evaluation unit, connected to the state prediction unit, for performing a multi-dimensional evaluation and calculating a comprehensive score of the network using a comprehensive scoring formula based on the prediction result and the network state data to obtain a comprehensive evaluation result;

[0056] A decision matrix generating unit is connected to the comprehensive evaluation unit and is used to generate the adaptive decision matrix according to the comprehensive evaluation result.

[0057] Furthermore, the comprehensive evaluation unit includes:

[0058] An indicator classification subunit, which classifies the network status data and the prediction results to extract indicators related to the comprehensive scoring formula to obtain classification results;

[0059] A weight configuration subunit, connected to the indicator classification subunit, configures the weight of the classification result according to different network environments and application scenarios to obtain a configuration result;

[0060] A score calculation subunit, connected to the weight configuration subunit, calculates a comprehensive score based on the comprehensive score formula and the configuration result;

[0061] The level determination subunit is connected to the score calculation subunit and is used to determine the network level according to the comprehensive score and a preset network interval.

[0062] Compared with existing technologies, the present invention offers significant advantages in that, through device modeling and interface abstraction at the device abstraction layer, it shields the heterogeneity of different devices, freeing upper-layer applications from concerns about protocol differences between underlying devices and significantly reducing integration and maintenance costs. Resource pooling enables unified scheduling and optimization of device resources, improving resource utilization. This is particularly true in large-scale IoT deployments, where resources can be dynamically allocated and reclaimed to avoid waste. The converged communication layer supports multi-protocol adaptation, ensuring seamless communication between devices. Intelligent routing and QoS guarantees optimize data transmission paths and quality, improving system responsiveness and stability. Routing guarantees and queue mapping enhance communication fault tolerance, ensuring data loss through backup paths or queue buffering even in the event of network fluctuations or device failures. The business logic layer utilizes a microservices architecture, enabling independent deployment and scalability of the process orchestration module and the intelligent decision-making module, facilitating rapid iteration and customized development. An event-driven mechanism enhances the system's real-time performance. The intelligent decision-making module integrates AI algorithms to support advanced features such as predictive maintenance and energy optimization, further enhancing the system's intelligence. The firmware upgrade layer uses state awareness and policy-driven decision-making to enable on-demand, batched firmware updates, avoiding network congestion or service interruptions caused by large-scale simultaneous device upgrades. Differential execution technology only transmits the firmware differences, significantly reducing bandwidth usage and upgrade time, making it particularly suitable for low-power devices. The firmware management module supports version control and rollback mechanisms, ensuring rapid recovery from upgrade failures and enhancing system robustness.

[0063] In particular, the device connection unit automatically records the device connection status and basic information, reducing manual intervention and improving device access efficiency. The initialization configuration unit performs initial configuration based on standardized models and device description information, so that the device reaches the predetermined working state at startup, ensuring the synchronization and stability of multi-device collaborative work. Personalized configuration is performed according to the characteristics of different devices, enhancing the system's adaptability to devices of different types and manufacturers. The operation monitoring unit monitors the operating status of the equipment in real time, and can quickly detect signs of equipment performance degradation or failure, reducing the impact of equipment failure on the system. The monitoring frequency and items are dynamically adjusted based on historical operating data, and monitoring resources are reasonably allocated to improve monitoring efficiency and reduce system load. The status assessment unit provides a quantitative basis for equipment maintenance through scientific assessment, enabling maintenance decisions based on the actual status of the equipment, rather than fixed maintenance cycles. The fault repair unit quickly performs repair operations according to preset strategies, reducing equipment fault repair time and improving equipment availability.

[0064] In particular, the path evaluation unit monitors the quality indicators of each node and link in the network in real time, can quickly obtain the real-time status of the network, and provide an accurate basis for routing decisions. Using scientific evaluation algorithms to analyze the collected information, the quality of each transmission path is comprehensively and accurately evaluated, laying the foundation for selecting the optimal path. Based on the evaluation results provided by the path evaluation unit, the routing decision unit uses an effective routing algorithm to select the optimal transmission path, significantly improving data transmission efficiency and reliability. When network conditions change, such as link failures or congestion, the dynamic adjustment unit can quickly adjust the transmission path according to the latest routing decision to ensure the continuity and stability of data transmission. Dynamic adjustment of the transmission path can promptly avoid network problems, reduce transmission interruptions or delays, and improve the stability and reliability of the system in complex network environments.

[0065] In particular, the rule management unit provides centralized management of decision-making rules, ensuring their accuracy and timeliness. The decision engine unit evaluates data in real time at key decision points in business processes based on the latest rules, quickly delivering accurate decisions. This significantly improves the efficiency and quality of business processing and reduces human error. The feedback adjustment unit comprehensively evaluates decision results, collects feedback data from multiple channels, and, after in-depth analysis, pinpoints issues within rules and processes. Adjustments and optimizations are made in a timely manner, creating a closed-loop management system that continuously improves the system's decision-making capabilities and accurately responds to changing business needs.

[0066] In particular, the indicator classification subunit clarifies which indicators are relevant to the comprehensive score through classification, helping to focus on in-depth analysis and processing of key indicators, thereby improving the efficiency and accuracy of the assessment. The weight configuration subunit configures the weights of the classification results according to different network environments and application scenarios, allowing the comprehensive assessment to flexibly adapt to various actual situations and improving the relevance and practicality of the assessment results. The score calculation subunit calculates the comprehensive score based on the comprehensive scoring formula and configuration results, presenting the network status in a quantitative manner, facilitating an intuitive understanding of the overall performance and health of the network. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 A schematic diagram of the structure of an intelligent edge-based collaborative management system for IoT devices provided by an embodiment of the present invention;

[0068] Figure 2 A schematic diagram of the structure of the device abstraction layer in the intelligent edge-based collaborative management system for IoT devices provided by an embodiment of the present invention;

[0069] Figure 3 A schematic diagram of the structure of the converged communication layer in the intelligent edge-based IoT device collaborative management system provided by an embodiment of the present invention;

[0070] Figure 4A schematic diagram of the structure of the business logic layer in the intelligent edge-based collaborative management system for IoT devices provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0071] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0072] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0073] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside", and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0074] Furthermore, it should be noted that, in the description of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0075] See also Figure 1 As shown, the present invention provides an IoT device collaborative management system based on intelligent edge, including:

[0076] Device abstraction layer 10, used to connect to various peripheral devices to implement abstract management and resource pooling of the peripheral devices;

[0077] The fusion communication layer 20 is connected to the device abstraction layer 10 to achieve adaptation and communication of different protocols;

[0078] The business logic layer 30 is connected to the integrated communication layer 20 to implement specific business logic and functions;

[0079] The firmware upgrade layer 40 is connected to the device abstraction layer 10 and the converged communication layer 20 respectively to implement the upgrade of the device firmware;

[0080] The device abstraction layer includes a device modeling module, an interface abstraction module, a resource pool management module, a device description module and a cycle management module;

[0081] The converged communication layer includes a plug-in management module, a protocol adaptation module, a routing guarantee module, an intelligent routing module, a QoS guarantee module and a queue mapping module;

[0082] The business logic layer includes a microservices partitioning module, an architecture building module, a process orchestration module, an intelligent decision-making module, and an event-driven module;

[0083] The firmware upgrade layer includes a state perception module, a policy decision module, a differential execution module and a firmware management module.

[0084] Specifically, the device abstraction layer's primary function is to connect to various peripheral devices and implement abstract management and resource pooling for these devices. This layer serves as the foundation for interaction between the entire system and peripheral devices, ensuring that the diversity and complexity of devices do not directly impact the business logic and communication processes of upper layers. The device modeling module uses a device model to predefine the various connectable peripheral devices in the system in a metaclass format. A standardized model is created for each device, including key information such as its functions, attributes, and behavior. Based on the results of unified device modeling, the interface abstraction module standardizes device driver interfaces. This module shields the underlying interface differences between different devices and provides a unified access interface to upper layers. This ensures that upper-layer modules do not need to be aware of specific device hardware details, thus enabling unified access to diverse devices. The resource pool management module utilizes a unified interface abstracted by standardized interfaces to centrally manage device resources through a converged resource pool. Its function is to achieve centralized resource management and optimized configuration, thereby improving resource utilization. The device description module uses a unified device description language (such as XML or JSON) to describe device information. This module enhances device manageability and configurability, making it easier to read and modify device information. The Lifecycle Management Module leverages the device management core. Based on the standardized models, unified interfaces, resource management, and description languages ​​established in all previous steps, it implements full device lifecycle management. This module covers device connection, initialization, operation, status monitoring, fault handling, and disconnection, ensuring device stability and reliability throughout its lifecycle. The Converged Communication Layer is responsible for adapting and communicating different protocols, ensuring efficient and reliable data transmission between devices. It bridges the gap between the device abstraction layer and the business logic layer, handling all communication-related matters. The Plugin Management Module loads and manages independent plug-ins for various communication protocols. This module supports hot-swapping and runtime updates, ensuring system flexibility and maintainability. The Protocol Adaptation Module interfaces with the Protocol Plugin Loading Management Module and adapts, directs, and parses data from different protocols based on loaded protocol plug-ins. This module ensures unified processing of data across different protocols within the system. The Routing Assurance Module provides routing assurance during communication, ensuring reliable data transmission paths. Its function is to maintain the stability of communication links and the continuity of data transmission. The Intelligent Routing Module uses converted standardized messages to dynamically select the optimal transmission path based on node status and link quality. This module intelligently optimizes data transmission paths and improves communication efficiency. The QoS assurance module classifies and schedules data based on service type and priority. This module ensures the reliable transmission of critical business data and prioritizes important data. The queue mapping module maps different service flow characteristics to different queues based on the assurance results. This module prioritizes high-priority control instructions and messages, optimizing the order of data processing. The business logic layer implements specific business logic and functions.This layer performs business processing and decision-making based on data transmitted by the converged communication layer. The microservice partitioning module divides business rules into several microservices based on business needs. Each microservice corresponds to a specific domain boundary, achieving modularization and independence of business logic. The architecture building module connects to the microservice partitioning module and builds a microservice architecture based on domain boundaries. This module provides architectural support for the implementation of business logic. The process orchestration module connects to the architecture building module and introduces a visual business process orchestration engine based on the microservice architecture. This module enables visual design and management of business processes, improving their flexibility and maintainability. The intelligent decision-making module connects to the process orchestration module and makes intelligent judgments at decision points in the process based on orchestration results and preset rules. The event-driven module connects to the process orchestration module and the intelligent decision-making module, respectively, to monitor business events. Based on the judgment results, it adjusts the process, achieving dynamic adaptation and optimization of the business process. The firmware upgrade layer implements device firmware upgrades, ensuring timely updates to device performance and functionality. The status awareness module captures network status indicators in real time, including various indicators at the physical, transport, and application layers. This module provides real-time network status information for firmware upgrades. The policy decision module uses an evaluation algorithm based on the perceived network status to output an adaptive decision matrix. This module determines the appropriate upgrade strategy based on the network level, ensuring the efficiency and adaptability of the firmware upgrade process. In differential upgrade scenarios, the server uses tools to compare the new and old firmware versions and generate a differential package. After receiving the differential package, the device uses the patch application engine and tools to restore the complete firmware, and the verification module ensures data consistency. This module enables efficient and secure firmware upgrades. The firmware management module manages the entire firmware upgrade process, including uploading and checking the upgrade package, and updating the device.

[0085] Specifically, the device abstraction layer shields the heterogeneity of different devices through device modeling and interface abstraction, freeing upper-layer applications from concerns about protocol differences between underlying devices and significantly reducing integration and maintenance costs. Resource pooling enables unified scheduling and optimization of device resources, improving resource utilization. This is particularly true in large-scale IoT deployments, where resources can be dynamically allocated and reclaimed to avoid waste. The converged communication layer supports multi-protocol adaptation, ensuring seamless communication between different devices. Intelligent routing and QoS guarantees optimize data transmission paths and quality, improving system responsiveness and stability. Routing guarantees and queue mapping enhance communication fault tolerance, ensuring data loss through alternate paths or queue buffering even in the event of network fluctuations or device failures. The business logic layer utilizes a microservices architecture, enabling independent deployment and scalability of the process orchestration and intelligent decision-making modules, facilitating rapid iteration and customized development. An event-driven mechanism enhances the system's real-time performance. The intelligent decision-making module integrates AI algorithms to support advanced features such as predictive maintenance and energy optimization, further enhancing the system's intelligence. The firmware upgrade layer uses state awareness and policy-driven decision-making to enable on-demand, batched firmware updates, avoiding network congestion or service interruptions caused by large-scale simultaneous device upgrades. Differential execution technology only transmits the firmware differences, significantly reducing bandwidth usage and upgrade time, making it particularly suitable for low-power devices. The firmware management module supports version control and rollback mechanisms, ensuring rapid recovery from upgrade failures and enhancing system robustness.

[0086] Specifically, the device abstraction layer 10 includes:

[0087] The device modeling module 11 uses a device model to predefine various peripheral devices that can be connected to the system in a metaclass manner and creates a standardized model for each device;

[0088] An interface abstraction module 12 is connected to the device modeling module 11 and performs standardization processing on the device driver interface according to the standardization model to obtain a standardized interface;

[0089] The resource pool management module 13 is connected to the interface abstraction module 12 and is used to centrally manage the resources of the device through the fusion resource pool according to the standardized interface to obtain a management result;

[0090] The device description module 14 is used to describe the device information using a unified device description language to obtain a description result;

[0091] The cycle management module 15 is connected to the device modeling module 11, the interface abstraction module 12, the resource pool management module 13, and the device description module 14 respectively, and performs full life cycle management of the device according to the standardized model, the standardized interface, the management results, and the description results.

[0092] Specifically, the device modeling module identifies the types of peripheral devices that may be connected to the system, such as sensors, actuators, and controllers. A metaclass model is defined for each device type, including basic device attributes (such as name, ID, and type), functions (such as reading and writing data), and behaviors (such as response time and operating frequency). Based on the metaclass model, a standardized model is created for each device, ensuring that all device models have a consistent structure and interface. The interface abstraction module obtains the standardized model for each device from the device modeling module. It analyzes the underlying interface differences between different devices, including communication protocols, data formats, and electrical interfaces. Based on the standardized model, a unified device access interface is designed, defining common interface functions and data structures. Specific interface conversion logic is implemented for each device, converting the underlying interface into a standardized interface. The resource pool management module creates a converged resource pool for centralized device resource management. It obtains the standardized interfaces of the devices from the interface abstraction module. Device resources are registered in the resource pool, including device connection information, resource type, and resource status. Resources are allocated from the resource pool to devices based on their needs. When a device is disconnected or resources are no longer needed, the resources are reclaimed from the device and returned to the resource pool. Define a unified interface for device driver plug-ins to ensure dynamic plug-in loading and management. Maintain a plug-in repository to store all available device driver plug-ins. Load the appropriate driver plug-in from the plug-in repository based on device type and requirements. When a device is disconnected or a plug-in is no longer needed, uninstall the corresponding driver plug-in to free up system resources. Support dynamic plug-in updates to ensure the system can adapt to new devices and new functions. The device description module uses a unified device description language, such as XML or JSON. Define a device description template, including basic device information, function descriptions, and interface descriptions. Generate a description file for each device, detailing its features and configuration. Provide a parsing tool to read and parse device description files to obtain device-related information. The lifecycle management module obtains the standardized device model from the device modeling module, the standardized interface from the interface abstraction module, the management results from the resource pool management module, and the description results from the device description module. When a device is connected, it establishes the connection based on the standardized model and interface and records the device connection status. During device initialization, it configures device parameters based on the description results to ensure proper operation. During device operation, it continuously monitors device status, including performance indicators and resource utilization. When a device fails, corresponding fault handling operations are performed based on the standardized model and description results, such as restarting the device, reconfiguring parameters, etc. When the device is disconnected, ensure that all related resources are properly recovered and the device connection status is updated.

[0093] Specifically, the device modeling module creates standardized models for each peripheral device, enabling unified management and abstract representation of diverse devices. This allows the system to handle a wide range of devices in a consistent manner, simplifying device management. The interface abstraction module standardizes interfaces, effectively shielding differences in underlying interfaces across devices. This decouples software from hardware, improving system portability and maintainability. The resource pool management module centrally manages device resources by integrating resource pools, achieving efficient resource utilization and optimal allocation, avoiding resource waste and overallocation. The device description module uses a unified device description language to describe device information, making it easier to read and modify device information and enhancing device manageability and configurability. Based on the information provided by the previous modules, the lifecycle management module implements full device lifecycle management, covering all aspects of the device lifecycle, including connection, initialization, operation, status monitoring, fault handling, and disconnection. This lifecycle management enables timely detection and resolution of device faults, reducing downtime, extending device lifespan, and improving device reliability and overall system performance.

[0094] Specifically, the cycle management module includes:

[0095] The device connection unit is used to record the connection status and related basic information of the device into the standardized model to obtain the connection result;

[0096] an initialization configuration unit connected to the device connection unit, performing initial configuration on the device according to the standardized model and device description information to obtain a configuration result;

[0097] An operation monitoring unit, for monitoring the operation status of the equipment in real time, and dynamically adjusting the monitoring frequency and monitoring items according to the historical operation data recorded in the management result to obtain the monitoring result;

[0098] a status evaluation unit, connected to the operation monitoring unit, for evaluating the health status and performance level of the equipment according to the monitoring results to obtain an evaluation result;

[0099] The fault repair unit is connected to the status evaluation unit and is used to perform corresponding repair or recovery operations according to the evaluation result and the fault handling strategy in the management result, and update the device status information in the management result.

[0100] Specifically, based on user actions or the system's pre-set scanning mechanism, a connection instruction is sent to the target peripheral device, for example, via a communication protocol such as Bluetooth, Wi-Fi, or Ethernet. Based on the device's feedback, a link is established using the corresponding communication protocol. If the device responds correctly, the link establishment is confirmed successful. Basic device information is obtained, such as the device type (sensor, actuator, etc.), unique identifier (MAC address, serial number, etc.), manufacturer information, and firmware version. The device's connection status (connected, connection failed, etc.) and collected basic information are recorded in the device's standardized model, forming a connection result for subsequent management. The initialization configuration unit ensures that the device operates according to the established settings at startup. Based on the device's standardized model and description information, appropriate initial parameters are set for the device. The device's standardized model and device description file, containing initial configuration parameters such as the default communication baud rate, data format, and operating mode, are obtained from the device abstraction layer. Based on this information, initialization instructions are generated for the device. For example, for a temperature sensor, the sampling frequency is set to once per second and the communication protocol is Modbus. The initialization instruction is sent to the device, and the device waits for feedback on the execution results. If the device confirms that the new configuration has been applied, initialization is successful. If it fails, diagnosis and adjustments are performed based on the error message. The operation monitoring unit continuously tracks the device's operating status and dynamically adjusts monitoring strategies to ensure timely detection of potential problems. Through the device driver or communication protocol, it regularly reads operational data from the device, such as sensor measurements, actuator operating status, device temperature, voltage, and other operating parameters. The range and fluctuation patterns of key indicators during operation are analyzed. Based on historical data analysis, the monitoring frequency and monitored items are adjusted. For example, if a device experiences high load during a specific period, the monitoring frequency can be increased during that period. The monitoring frequency can be reduced for parameters with high stability, while it can be increased for parameters with large fluctuations. The condition assessment unit quantitatively assesses the device's health and performance level, providing a basis for subsequent maintenance decisions. The data transmitted by the operation monitoring unit is deeply analyzed, using algorithms and models to evaluate device performance indicators such as sensor accuracy and actuator response time. The assessment results are compared with preset performance indicator thresholds to determine whether the device is operating normally. For example, if the vibration amplitude of a motor exceeds a set threshold, a mechanical failure may be identified. Combined with historical performance data, the system predicts the likely time of future equipment failures, providing a reference for preventive maintenance. The equipment's health and performance assessment results are recorded in detail in the management results for easy query and analysis. Based on the assessment results and pre-defined troubleshooting strategies, the fault repair unit takes effective measures to repair or restore normal equipment operation. The results from the status assessment unit are thoroughly analyzed, combined with the troubleshooting strategies in the management results, to determine a specific repair plan.For faults that can be automatically recovered, such as communication interruptions, a retry mechanism can be set up. For faults that require manual intervention, such as hardware damage, a repair work order is generated. Follow the repair plan and perform actions, such as restarting the device, reconfiguring parameters, updating firmware, or invoking remote repair services. After the repair is complete, monitor the device status to verify the success of the repair. If the device returns to normal, the device status in the management results will be updated to normal. If the repair fails, further analyze the cause and take remedial measures.

[0101] Specifically, the device connection unit automatically records the device connection status and basic information, reducing manual intervention and improving device access efficiency. The initialization configuration unit performs initial configuration based on standardized models and device description information, so that the device reaches the predetermined working state at startup, ensuring the synchronization and stability of multi-device collaborative work. Personalized configuration is performed according to the characteristics of different devices, enhancing the system's adaptability to devices of different types and manufacturers. The operation monitoring unit monitors the operating status of the equipment in real time, and can quickly detect signs of equipment performance degradation or failure, reducing the impact of equipment failure on the system. The monitoring frequency and items are dynamically adjusted based on historical operating data, and monitoring resources are reasonably allocated to improve monitoring efficiency and reduce system load. The status assessment unit provides a quantitative basis for equipment maintenance through scientific assessment, enabling maintenance decisions based on the actual status of the equipment, rather than fixed maintenance cycles. The fault repair unit quickly performs repair operations according to preset strategies, reducing equipment fault repair time and improving equipment availability.

[0102] Specifically, if Figure 3 As shown, the converged communication layer 20 includes:

[0103] A plug-in management module 21 is used to load and manage independent plug-ins of various communication protocols;

[0104] The protocol adaptation module 22 is connected to the plug-in management module 21 and adapts, diverts and parses data of different protocols according to the loaded protocol plug-ins to obtain adaptation results;

[0105] The message conversion module 23 is connected to the protocol adaptation module 22, converts the adaptation result into a standardized message format, and defines a standardized data packet structure to encapsulate and convert the data to obtain a conversion result;

[0106] An intelligent routing module 24, connected to the message conversion module 23, is used to use the conversion result to dynamically select an optimal transmission path according to node status and link quality to obtain a routing result;

[0107] QoS guarantee module 25, used to classify and schedule data according to the type and priority of business data to obtain guarantee results;

[0108] The queue mapping module 26 is connected to the intelligent routing module 24 and the QoS guarantee module 25 respectively, and maps different service flow characteristics to different queues according to the guarantee result and the routing result to obtain mapping results.

[0109] Specifically, the plug-in management module identifies the communication protocols the system needs to support, such as MQTT, HTTP, and TCP / IP, and loads the corresponding plug-ins from the plug-in library. Each plug-in runs independently and fully functionally. When a plug-in is no longer needed or needs to be updated, it is safely stopped and uninstalled to free up system resources. Hot-swappable functionality allows plug-ins to be updated while the system is running, without restarting the system, minimizing business impact. The module continuously monitors the running status of plug-ins, promptly detecting and addressing anomalies to ensure stable plug-in operation. The protocol adapter module receives data from different protocols and accurately identifies its type based on protocol characteristics. It converts the data to the target protocol format, including adjusting the data structure and encoding method to meet the requirements of the system or the target protocol. Based on the protocol type and data flow, the data is sent to the corresponding processing module or plug-in to ensure correct data processing. The received data is parsed to extract key information for subsequent processing. The message conversion module receives the adapted data from the protocol adapter module as input for message conversion. It converts data from different protocols into an internally standardized message format, including standardized header, payload, and metadata structures. Data is encapsulated according to a standardized packet structure to ensure data integrity and consistency. When necessary, data format conversion is performed, such as converting XML data to JSON, to meet the requirements of different system components. The intelligent routing module uses the standardized messages output by the message conversion module as the basis for intelligent routing. It monitors the status of each node and link quality in the network in real time, including metrics such as bandwidth, latency, and packet loss rate. Based on current network conditions and pre-set routing algorithms, it selects the optimal path for each message. When network conditions change, it recalculates routes based on the latest information to ensure efficient and reliable data transmission. The QoS assurance module classifies service data based on its type and priority, such as voice data, video data, and text data. It implements differentiated service scheduling for different data types, allocating more resources to high-priority data. It rationally allocates system resources, such as bandwidth and computing resources, to ensure the transmission quality of critical service data. Data flow is controlled based on network conditions and resource usage to avoid network congestion. The queue mapping module uses the QoS assurance module's assurance results and the intelligent routing module's routing results as the basis for queue mapping. Based on service flow characteristics and priorities, data is mapped to different queues, such as high-priority data to a higher-priority queue. Queues can be created and deleted, and queue priorities and resource allocation can be adjusted to ensure efficient queue management. Data transmission is scheduled based on queue priority, with data from high-priority queues being sent first, ensuring the timeliness of critical service data.

[0110] Specifically, the plug-in management module supports the flexible loading and management of multiple communication protocol plug-ins, enabling the system to adapt to the communication needs of different devices and application scenarios. The protocol adaptation module adapts, directs, and parses data from different protocols, addressing the challenge of diverse communication protocols across IoT devices and enabling interoperability between heterogeneous systems. Efficient protocol parsing and adaptation algorithms ensure fast and accurate data conversion between different protocols, reducing system latency. The message conversion module converts data from different protocols into a standardized message format, facilitating subsequent data processing and analysis, and improving system integration and scalability. The intelligent routing module dynamically selects the optimal transmission path based on real-time node status and link quality, improving data transmission efficiency and reliability and reducing the risk of network congestion. Intelligent routing decisions fully utilize network resources, enabling rapid data transmission and enhancing overall system performance and responsiveness. The QoS assurance module classifies services and schedules traffic based on data type and priority, ensuring reliable transmission of critical business data and improving system stability and availability. The queue mapping module maps different service flow characteristics to different queues based on QoS assurance and routing results, enabling refined data flow management and scheduling.

[0111] Specifically, the intelligent routing module includes:

[0112] A path evaluation unit, configured to collect and analyze in real time the status information of each node in the network and the quality indicators of the links to obtain collection results, and to evaluate the collection results to obtain evaluation results;

[0113] a routing decision unit, connected to the path evaluation unit, configured to use a routing algorithm to select an optimal transmission path according to the evaluation result to obtain a decision result;

[0114] The dynamic adjustment unit is connected to the routing decision unit and is used to perform dynamic adjustment according to the decision result when the network conditions change.

[0115] Specifically, the path evaluation unit continuously monitors the operational status of each node in the network, including metrics such as node load, CPU utilization, and memory usage. It detects quality indicators for each link, such as bandwidth, latency, packet loss rate, and jitter, by sending probe packets or leveraging information provided by network protocols. The collected node status and link quality information is statistically analyzed and organized, calculating statistics such as the average, maximum, and minimum values ​​of the relevant indicators. An evaluation algorithm is then used to comprehensively evaluate each path and calculate its weight or score. The evaluation algorithm can be designed based on specific needs. For example, a weighted average method can be used to assign weights based on the importance of each indicator to calculate a comprehensive path score. Alternatively, a fuzzy evaluation method can be used to account for the ambiguity and uncertainty of the indicators. Based on the results of the evaluation algorithm, the quality of each path is determined and an evaluation result is generated, which provides a basis for routing decisions. The routing decision unit selects an appropriate routing algorithm, such as the shortest path first algorithm, based on network characteristics and service requirements. Initial parameters required by the routing algorithm are set, such as the initial routing table and initial path weights. The evaluation result generated by the path evaluation unit is received, including information such as the weights and scores of each path. Based on the evaluation results, the routing algorithm calculates the optimal transmission path. For example, in the SPF algorithm, a shortest path tree is constructed based on path weights, and the shortest path from the source node to the destination node is selected as the optimal transmission path. After determining the optimal transmission path, a decision result is generated, including information such as the selected path and next-hop node, to provide guidance for data transmission. The dynamic adjustment unit closely monitors changes in network conditions, such as link failures, node failures, and network congestion. It receives feedback from other modules or network devices, such as congestion notifications and link status updates. When a network condition change is detected or feedback is received, it determines whether the trigger conditions for dynamic adjustment are met. For example, a link failure may render the current optimal path unavailable, or network congestion may exceed a set threshold. Based on the decision result of the routing decision unit and the current network conditions, the optimal transmission path is recalculated and the routing decision is updated. For example, if the original optimal path fails, a suboptimal path is selected as the new optimal transmission path, and data transmission direction is adjusted accordingly. Based on the updated routing decision, the data transmission path is adjusted to ensure data transmission along the new optimal path.

[0116] Specifically, the path evaluation unit monitors the quality indicators of each node and link in the network in real time, can quickly obtain the real-time status of the network, and provide an accurate basis for routing decisions. Using scientific evaluation algorithms to analyze the collected information, it comprehensively and accurately evaluates the quality of each transmission path, laying the foundation for selecting the optimal path. Based on the evaluation results provided by the path evaluation unit, the routing decision unit uses an effective routing algorithm to select the optimal transmission path, significantly improving data transmission efficiency and reliability. When network conditions change, such as link failures or congestion, the dynamic adjustment unit can quickly adjust the transmission path according to the latest routing decision to ensure the continuity and stability of data transmission. Dynamic adjustment of the transmission path can promptly avoid network problems, reduce transmission interruptions or delays, and improve the stability and reliability of the system in complex network environments.

[0117] Specifically, if Figure 4 As shown, the business logic layer 30 includes:

[0118] The microservice partitioning module 31 divides the business rules into several microservices according to business requirements, and each microservice corresponds to each domain boundary;

[0119] An architecture building module 32, connected to the microservice partitioning module 31, is used to build a microservice architecture according to the domain boundaries;

[0120] The process orchestration module 33 is connected to the architecture building module 32 and introduces a visual business process orchestration engine according to the microservice architecture to obtain an orchestration result;

[0121] An intelligent decision-making module 34 is connected to the process orchestration module 33 and performs intelligent judgment on the decision points in the process according to the orchestration results and preset rules to obtain a judgment result;

[0122] The event-driven module 35 is connected to the process orchestration module 33 and the intelligent decision-making module 34 respectively, and is used to monitor business events and adjust the process according to the judgment results.

[0123] Specifically, the microservice partitioning module first conducts an in-depth analysis of specific IoT application scenarios and business requirements, such as "security monitoring" and "environmental control" in smart homes, and "predictive equipment maintenance" and "production process optimization" in the Industrial IoT. Using a domain-driven design approach, different business domains and core business rules are identified. For example, in a smart building, "energy management" and "visitor management" are two distinct domains. Based on the identified domain boundaries and business rules, system functions are partitioned into a set of highly cohesive and loosely coupled microservices. Each microservice focuses on a specific business capability, such as "sensor data acquisition service," "device control service," "alarm analysis service," and "user configuration service." A clear, standardized API (such as a RESTful API) is defined for each microservice, clearly defining its inputs, outputs, and functionality. A service registration mechanism is designed or implemented to enable microservices to be discovered and invoked by other components in the system. Finally, a microservice list is generated, clearly defining each microservice's name, domain, core functionality, and external interfaces. This serves as input for the architecture building module. Based on the microservice list, the architecture building module designs a system architecture blueprint for the entire business logic layer. This includes: defining dependencies and data flows between microservices; selecting a communication method between microservices, such as synchronous calls (REST / gRPC) or asynchronous messaging (MQTT, Kafka); designing a data storage solution for microservices (such as a relational database, NoSQL database, or time-series database), taking into account data consistency and isolation; packaging one or more related microservices into a single deployment unit (such as a Docker container); clarifying the microservices' dependencies on underlying infrastructure (such as message queues, caches, and databases); defining interfaces between the business logic layer and the converged communication layer and firmware upgrade layer; and designing authentication, authorization, and communication encryption mechanisms between microservices. The process orchestration module integrates a visual business process orchestration engine (such as a BPMN 2.0 engine). This engine provides a graphical interface or scripting language for defining business processes. Based on the microservice architecture and business requirements provided by the architecture building module, the orchestration engine is used to define specific business processes. A process consists of a series of steps (tasks), each of which typically calls one or more microservices. Developers or business analysts use a visual interface to drag and drop service nodes, define process branches, set conditions, and configure data transmission to build a business process diagram. Embed logic such as microservice invocation, data conversion, and simple judgment into process steps. The orchestration engine saves the designed process as an executable process definition (orchestration result), which includes the process step sequence, dependencies, parameter passing rules, and more. When a business event triggers, the process orchestration module instantiates the orchestration result, creates a specific process execution instance, and drives the execution of microservices according to the defined sequence and logic. The intelligent decision-making module integrates a rules engine (such as Drools) or a machine learning model service.Predefine a series of decision rules or train machine learning models based on business needs. Rules are typically based on business logic (e.g., "If the temperature exceeds a threshold and the humidity is below a threshold, it is considered a fire risk"), while models may learn complex patterns based on historical data (e.g., predicting the probability of equipment failure). Within the process defined by the process orchestration module, clearly identify decision points requiring intelligent judgment. When the process reaches a decision point, the intelligent decision-making module obtains the necessary data as input from sources such as the process context, equipment data, and historical data. The input data is fed into the rule engine or machine learning model for rule matching or model inference. The rule engine returns the conclusion of the matched rule, or the model returns the prediction result / classification label. This judgment result is a clear instruction or suggestion (e.g., "Perform action A," "Jump to step X," "Mark as high risk"). The judgment result is passed to the event-driven module for subsequent process adjustments. The event-driven module continuously monitors business events from multiple sources, including: device status changes, data reports, and alarms transmitted through the converged communication layer; process start, completion, and exception events generated within the process orchestration module; and instructions and notifications from other systems or users. This module is typically implemented using message queue or event bus technology to ensure reliable event delivery and decoupling. For each monitored event, the module searches for predefined processing logic based on the event type and content. This processing logic may be directly associated with a predefined process or combined with the judgment results of the intelligent decision-making module. Process Adjustment / Triggering: If an event matches the start conditions of a process, the process orchestration module is triggered to instantiate and start that process. If an event occurs during process execution and the intelligent decision-making module's judgment indicates that adjustment is required, the module sends instructions to the process orchestration module to modify the execution path of the current process instance (such as skipping a step, modifying parameters, terminating the process, or starting another sub-process).

[0124] Specifically, the microservice partitioning module splits complex business rules into independent microservices, each focusing on a specific business function, improving code maintainability and readability. The process orchestration module enables rapid adjustment and optimization of business processes to adapt to market changes and fluctuating business needs, enhancing enterprise competitiveness. The intelligent decision-making module reduces the need for manual intervention and achieves a higher level of automation by intelligently assessing decision points within the process. The event-driven module monitors business events in real time and adjusts processes accordingly, enabling the system to rapidly respond to business changes and improving real-time processing capabilities.

[0125] Specifically, the intelligent decision-making module includes:

[0126] Rule management unit, used to define, store and update decision rules to obtain management results;

[0127] A decision engine unit, connected to the rule management unit, for judging the decision points in the business process according to the management results to obtain a decision result;

[0128] The feedback adjustment unit is connected to the decision engine unit and is used to evaluate the decision result to obtain an evaluation result, and adjust the decision rules and processes according to the evaluation result.

[0129] Specifically, the rule management unit conducts in-depth communication with business experts and relevant teams to fully understand business logic and decision-making requirements. By analyzing business processes, key decision points and decision rules are identified. Business rules are converted into computer-readable representations, such as decision trees, decision tables, or rule languages. For example, for customer credit assessment, the rule definition might be: "If a customer's credit score is greater than 800 and there are no overdue payments, then the customer's credit rating is excellent." Use a relational or NoSQL database to store rules and ensure their organization. Design a reasonable database table structure to store rule attributes, conditions, actions, and other information. Implement a version control mechanism to record the creation, modification, and deletion history of rules. Use version control tools such as Git to manage rule files for easy tracking and rollback. Provide an intuitive user interface or API to allow authorized users to update rules promptly based on business changes. For example, develop a web-based rule management interface that allows business personnel to add, modify, or delete rules. Establish a strict review process to ensure the accuracy and compliance of rule updates. After rule updates are submitted, they are reviewed by relevant business experts and technical personnel to ensure their correctness and manageable impact on business processes. When the decision engine is started, it retrieves the latest rule set from the rule management unit through API calls or database queries. For example, during initialization, the decision engine sends an HTTP request to the rule management unit to retrieve all rules relevant to the current business. The rules are converted into executable logic for the decision engine. Rules based on rule languages ​​are parsed using the corresponding parser; rules in the form of decision trees or decision tables are converted into program logic. For example, the rule "If a customer's credit score is greater than 800 and there are no overdue payments, then the customer's credit rating is excellent" is parsed into a conditional statement in the program. Embedding the decision engine in the business process: The decision engine is integrated into key decision points in the business process. For example, in the credit approval process, the decision engine is embedded in the customer credit assessment phase. At the decision point, real-time data from the business process is collected through integration with business systems. For example, the customer's credit score and overdue payment history are obtained from the customer relationship management system. Based on real-time data and pre-set rules, logical judgment is performed to reach a decision result. For example, based on a customer's credit score and overdue payment history, the decision engine determines the customer's credit rating as excellent, good, or fair. This decision is fed back to the business process to guide subsequent operations. For example, the customer's credit rating is fed back to the credit approval system so that the approval personnel can approve the loan amount based on the credit rating. The feedback adjustment unit defines metrics for evaluating decision results, such as accuracy, timeliness, and business impact. For example, for a credit approval decision, the accuracy metric might be the degree of consistency between the credit rating assessment and the customer's actual default status.Collect feedback data on decision-making results through multiple channels, including business systems, user feedback, and system logs. For example, collect information on actual defaults after a customer takes out a loan, as well as customer satisfaction with the credit approval process. Use data analysis techniques (such as data mining and statistical analysis) to mine this feedback data and identify issues in rules and processes. For example, by analyzing the discrepancies between actual customer defaults and credit rating assessment results, it can be discovered that certain rules are inaccurate. Based on these analysis results, modify decision rules or optimize business processes. For example, adjust credit score thresholds or add new evaluation metrics.

[0130] Specifically, the rule management unit provides centralized management of decision-making rules, ensuring their accuracy and timeliness. The decision engine unit evaluates data in real time at key decision points in business processes based on the latest rules, quickly delivering accurate decisions. This significantly improves the efficiency and quality of business processing and reduces human error. The feedback adjustment unit comprehensively evaluates decision results, collects feedback data from multiple channels, and conducts in-depth analysis to pinpoint issues within rules and processes. Adjustments and optimizations are then made in a timely manner, creating a closed-loop management system that continuously improves the system's decision-making capabilities and accurately responds to changing business needs.

[0131] Specifically, the firmware upgrade layer includes:

[0132] A state perception module is used to capture network state data in real time, including physical layer indicators, transport layer indicators, and application layer multi-dimensional indicators;

[0133] an intelligent analysis module, connected to the state perception module, for performing predictions using a prediction algorithm based on the physical layer indicators, the transport layer indicators, and the application layer multi-dimensional indicators to generate an adaptive decision matrix;

[0134] an intelligent execution module, connected to the intelligent analysis module, and selecting an upgrade strategy according to the decision matrix to execute the firmware upgrade to obtain an execution result;

[0135] The firmware management module is connected to the intelligent execution module and is responsible for generating and applying the difference package when the intelligent execution module selects the differential upgrade strategy.

[0136] Specifically, network interfaces and hardware tools are used to capture physical layer metrics in real time, including port utilization, packet error rate, packet distortion rate, and CRC error rate. By monitoring these metrics, the network's stability and reliability at the physical transmission level are understood. By analyzing the transmission characteristics of network packets, transport layer metrics such as TCP retransmission rate, round-trip time (RTT), and throughput are obtained. These metrics reflect the network's efficiency and reliability during data transmission. At the application layer, service data is analyzed in detail, collecting multi-dimensional metrics such as service data transmission delay, packet loss rate, and application layer protocol response time. These metrics are directly related to user experience and service performance. The intelligent analysis module selects appropriate prediction algorithms, such as time series analysis and machine learning algorithms (such as LSTM), based on service needs and data characteristics. Time series analysis is suitable for processing time-ordered network status data and can capture trends and cyclical changes in the data. Machine learning algorithms can handle more complex nonlinear relationships, learn patterns from historical data, and make predictions. Prediction models are trained using historical network status data, and model parameters are adjusted to improve prediction accuracy. During the training process, new data is continuously input, allowing the model to gradually adapt to changes in network conditions and optimize prediction results. The intelligent execution module selects the most appropriate upgrade strategy based on the recommendations in the matrix, such as direct upgrade, rate-limited upgrade, or differential upgrade. Based on the selected upgrade strategy, the module executes the corresponding firmware upgrade operations. During execution, the module monitors the upgrade progress to ensure a smooth process. Firmware Management Module: When the intelligent execution module selects the differential upgrade strategy, the module uses specialized tools to compare the new and old firmware versions and generate a differential package. These tools accurately identify differences between firmware files, ensuring the accuracy and integrity of the differential package. On the device side, the firmware management module is responsible for applying the received differential package to the current firmware version to generate a new firmware version. Before applying the differential package, the module performs an integrity check to ensure it has not been tampered with or corrupted. This check typically uses a hash algorithm, such as MD5 or SHA-1, to verify the integrity of the differential package by calculating the hash value and comparing it with the hash value provided by the server.

[0137] Specifically, the state awareness module captures multi-dimensional metrics from the physical, transport, and application layers in real time, providing comprehensive and accurate network status data for subsequent intelligent analysis, ensuring that upgrade decisions are based on the latest network conditions. The intelligent analysis module uses a predictive algorithm to analyze this collected network status data and generate an adaptive decision matrix. This accurately predicts the optimal upgrade strategy based on current network conditions, providing a scientific basis for the intelligent execution module and ensuring the efficiency and adaptability of the firmware upgrade process. The intelligent execution module selects the most appropriate upgrade strategy based on the decision matrix. This flexibility ensures optimal firmware upgrades under varying network conditions, improving upgrade success and efficiency. When the intelligent execution module selects a differential upgrade strategy, the firmware management module is responsible for generating and applying differential packets. Differential upgrades significantly reduce the amount of data required, lowering network bandwidth requirements and accelerating upgrades. This is particularly suitable for large-scale device upgrades or bandwidth-constrained scenarios, saving time and resources and improving upgrade efficiency.

[0138] Specifically, the intelligent analysis module includes:

[0139] Data preprocessing unit, which cleans, normalizes and extracts features from the collected data to obtain processing results;

[0140] a state prediction unit, connected to the data preprocessing unit, and using a long short-term memory network to predict the network state according to the processing result to obtain a prediction result;

[0141] a comprehensive evaluation unit, connected to the state prediction unit, for performing a multi-dimensional evaluation and calculating a comprehensive score of the network using a comprehensive scoring formula based on the prediction result and the network state data to obtain a comprehensive evaluation result;

[0142] A decision matrix generating unit is connected to the comprehensive evaluation unit and is used to generate the adaptive decision matrix according to the comprehensive evaluation result.

[0143] Specifically, the data preprocessing unit uses a data deduplication algorithm to identify and delete duplicate data records. For example, data records can be hashed using a hash algorithm, with identical hash values ​​representing duplicate data, thereby removing duplicates. Missing values ​​can be filled using the mean, median, or mode. For example, for numerical data, the mean or median of the indicator is used; for categorical data, the mode is used. Outliers can be identified using statistical methods (such as the 3σ principle) or machine learning-based anomaly detection algorithms (such as the Isolation Forest algorithm) and corrected or deleted. The normalization method is selected based on the data characteristics and model requirements. For data with large numerical ranges, Min-Max scaling can be used; for data that needs to eliminate dimensionality effects, Z-Score normalization can be used. For example, Min-Max scaling can be used to scale the data to the [0, 1] range; Z-Score normalization can be used to transform the data into a distribution with a mean of 0 and a standard deviation of 1. Correlation analysis can be used to calculate the correlation coefficient between features and the target variable to select features with high correlation. Alternatively, dimensionality reduction techniques such as principal component analysis (PCA) can be used to extract the main components of the data as key features. The identified key features are combined into a feature vector. For example, for network status data, key features such as port utilization at the physical layer, TCP retransmission rate at the transport layer, and transmission delay at the application layer can be combined into a feature vector, which serves as input for the subsequent prediction model. The state prediction unit designs an LSTM model consisting of an input layer, two LSTM layers, and an output layer. The input layer receives the feature vector, the LSTM layer is used to capture long-term dependencies in time series data, and the output layer predicts future network status. Each LSTM layer has 128 neurons, and the activation function uses ReLU to introduce nonlinearity. The sequence length is set to 10, meaning that data from the past 10 time steps is used to predict the network status for the next time step. Historical network status data from the past month is collected, arranged in chronological order, and divided into a training set and a validation set with an 8:2 ratio. The LSTM model is trained using the training set, using the mean squared error (MSE) as the loss function and the Adam optimization algorithm for gradient descent. After each training epoch, the model performance is evaluated on the validation set, and the model parameters are adjusted based on the validation loss. After training for multiple epochs, until the validation loss no longer decreases significantly, the model is considered trained. The feature vector output by the data preprocessing unit is used as the input to the LSTM model. For example, the feature vector sequence for the past 10 time steps is input into the model. After calculation, the model outputs predicted network status values ​​for the next time step, including predicted values ​​for physical layer metrics (such as port utilization), transport layer metrics (such as TCP retransmission rate), and application layer metrics (such as transmission delay). The comprehensive evaluation unit determines the weight of each metric based on business needs and network status data analysis.For example, after expert evaluation and data analysis, the weights for physical layer indicators are determined to be 0.4, transport layer indicators to be 0.3, and application layer indicators to be 0.3. A comprehensive scoring formula is constructed based on the weighted summation: Comprehensive score = physical layer indicator score × 0.4 + transport layer indicator score × 0.3 + application layer indicator score × 0.3. The predicted future network state output by the state prediction unit and the current actual network state data are obtained. For example, the predicted physical layer port utilization, transport layer TCP retransmission rate, and application layer transmission delay are obtained, along with the current actual values. The actual and predicted values ​​of each indicator are substituted into the comprehensive scoring formula. For each indicator, a score is calculated based on its actual and predicted values. For example, for physical layer port utilization, the normal range is set to [0, 70]%. If the current actual value is 60% and the predicted value is 65%, the score is (70 - 60) / (70 - 0) × 100 = 14.29 points (assuming a full score of 100, calculated using a linear scale). Scores for other indicators are calculated similarly and then substituted into the comprehensive scoring formula to calculate the network's comprehensive score. The decision matrix generation unit collects comprehensive network score data over a period of time and analyzes its distribution. For example, if comprehensive scores are mostly concentrated between 70 and 90, it indicates a good network condition. Based on the score distribution, network status classification criteria are set. For example, a comprehensive score greater than 80 is considered excellent, 60-80 is considered medium, and less than 60 is considered poor. Three firmware upgrade strategies are defined: direct upgrade, speed-limited upgrade, and differential upgrade. Direct upgrade is suitable for fast upgrades when network conditions are excellent; speed-limited upgrade is suitable for controlling upgrade traffic when network conditions are medium; and differential upgrade is suitable for reducing data transmission volume when network conditions are poor. A two-dimensional table is constructed, with rows representing network status levels (excellent, medium, poor) and columns representing upgrade strategy types (direct upgrade, speed-limited upgrade, differential upgrade). The decision matrix options are populated based on business needs and experience. For example, when network conditions are excellent, direct upgrade is preferred; when they are medium, speed-limited upgrade is preferred; and when they are poor, differential upgrade is preferred. A comprehensive evaluation of the decision matrix is ​​conducted quarterly. Collect historical data on firmware upgrades using the decision matrix and analyze metrics such as upgrade success rate, upgrade time, and resource consumption. If the current upgrade strategy has a low success rate for a certain network status level, consider adjusting the upgrade strategy priority for that status level. For example, if the upgrade success rate for rate-limited upgrades is low when the network status is medium, try increasing the priority of differential upgrades. Re-evaluate the decision matrix when the network environment changes (such as network expansion or service type changes) or when service requirements adjust (such as increased upgrade time requirements). For example, if the service has increased upgrade time requirements, the rate limit for rate-limited upgrades can be appropriately lowered to speed up the upgrade.

[0144] Specifically, the data preprocessing unit uses feature extraction to convert raw data into features that better reflect the essence of the network status. This helps the model capture key information and improves the accuracy of predictions and evaluations. The state prediction unit uses an LSTM network to predict network status for a period of time in the future. This allows firmware upgrade decisions to be made based not on the current instantaneous state but on upcoming network conditions. This avoids waiting until network deterioration is predicted, thereby improving the upgrade success rate.

[0145] Specifically, the comprehensive evaluation unit includes:

[0146] An indicator classification subunit, which classifies the network status data and the prediction results to extract indicators related to the comprehensive scoring formula to obtain classification results;

[0147] A weight configuration subunit, connected to the indicator classification subunit, configures the weight of the classification result according to different network environments and application scenarios to obtain a configuration result;

[0148] A score calculation subunit, connected to the weight configuration subunit, calculates a comprehensive score based on the comprehensive score formula and the configuration result;

[0149] The level determination subunit is connected to the score calculation subunit and is used to determine the network level according to the comprehensive score and a preset network interval.

[0150] Specifically, the indicator classification subunit determines the classification dimensions, such as physical layer indicators, transport layer indicators, and application layer indicators, based on the requirements of the comprehensive scoring formula. The specific indicators within each dimension and their classification criteria are clearly defined. For example, physical layer indicators include port utilization and packet error rate; transport layer indicators include TCP retransmission rate and throughput; and application layer indicators include transmission delay and packet loss rate. Indicators relevant to the comprehensive scoring formula are extracted from network status data and prediction results. These extracted indicators are classified into corresponding dimensions according to predefined classification rules. The weight configuration subunit analyzes the impact of each indicator on the comprehensive network score based on different network environments and application scenarios. The weight configuration for each indicator is determined based on historical data and practical experience. For example, if transport layer indicators have a greater impact on network quality in a specific scenario, a relatively higher weight can be assigned to them. Based on the analysis results, initial weights are assigned to the classified indicators. Either subjective weighting methods (such as expert scoring) or objective weighting methods (such as principal component analysis) can be used. Weight configuration is dynamically adjusted based on changes in the network environment and adjustments to business needs. For example, as business grows, the importance of application-layer indicators may increase, necessitating a corresponding increase in their weight. The scoring calculation subunit determines the specific form of the comprehensive scoring formula based on business needs and indicator characteristics. A weighted summation approach is generally used, but other mathematical methods can also be incorporated depending on the specific situation. The classified indicators and their corresponding weights are substituted into the comprehensive scoring formula to calculate the network's comprehensive score. The formula is: Comprehensive score = ∑(indicator value × weight). The grade determination subunit defines the network grade classification criteria based on business needs and network performance requirements. For example, networks can be classified into three grades: excellent, medium, and poor. A corresponding comprehensive score range is set for each network grade. For example, the score range for excellent is [80, 100], for medium is [60, 80], and for poor is [0, 60]. The calculated comprehensive score is compared with the pre-set score range to determine the network grade. For example, a comprehensive score of 85 is considered excellent. Based on the judgment result, the network grade is output. The results can be displayed to users in a visual manner, such as using different colors to identify different network levels on the monitoring interface, so that users can quickly understand the network status.

[0151] Specifically, the indicator classification subunit clarifies which indicators are relevant to the comprehensive score through classification, helping to focus on in-depth analysis and processing of key indicators, thereby improving the efficiency and accuracy of the assessment. The weight configuration subunit configures the weights of the classification results according to different network environments and application scenarios, allowing the comprehensive assessment to flexibly adapt to various actual situations and improving the relevance and practicality of the assessment results. The score calculation subunit calculates the comprehensive score based on the comprehensive scoring formula and configuration results, presenting the network status in a quantitative manner, facilitating an intuitive understanding of the overall performance and health of the network.

[0152] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

[0153] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A collaborative management system for IoT devices based on intelligent edge, characterized in that: include: Device abstraction layer, used to connect to various peripheral devices to achieve abstract management and resource pooling of the peripheral devices; A fusion communication layer connected to the device abstraction layer to achieve adaptation and communication of different protocols; The business logic layer is connected to the integrated communication layer to implement specific business logic and functions; The firmware upgrade layer is connected to the device abstraction layer and the fusion communication layer respectively to implement the upgrade of device firmware; The device abstraction layer includes a device modeling module, an interface abstraction module, a resource pool management module, a device description module and a cycle management module; The converged communication layer includes a plug-in management module, a protocol adaptation module, a routing guarantee module, an intelligent routing module, a QoS guarantee module and a queue mapping module; The business logic layer includes a microservices partitioning module, an architecture building module, a process orchestration module, an intelligent decision-making module, and an event-driven module; The firmware upgrade layer includes a state perception module, a policy decision module, a differential execution module and a firmware management module.

2. The IoT device collaborative management system based on intelligent edge according to claim 1 is characterized in that: The device abstraction layer includes: The device modeling module uses the device model to predefine various peripheral devices that can be connected to the system in the form of metaclasses, and creates a standardized model for each device; an interface abstraction module, connected to the device modeling module, and performing standardization processing on the device driver interface according to the standardization model to obtain a standardized interface; A resource pool management module, connected to the interface abstraction module, for centrally managing the resources of the device through the fusion resource pool according to the standardized interface to obtain a management result; The device description module is used to describe the device information using a unified device description language to obtain a description result; A cycle management module is connected to the device modeling module, the interface abstraction module, the resource pool management module and the device description module respectively, and performs full life cycle management of the device according to the standardized model, the standardized interface, the management results and the description results.

3. The IoT device collaborative management system based on intelligent edge according to claim 2 is characterized in that: The cycle management module includes: The device connection unit is used to record the connection status and related basic information of the device into the standardized model to obtain the connection result; an initialization configuration unit connected to the device connection unit, performing initial configuration on the device according to the standardized model and device description information to obtain a configuration result; An operation monitoring unit, for monitoring the operation status of the equipment in real time, and dynamically adjusting the monitoring frequency and monitoring items according to the historical operation data recorded in the management result to obtain the monitoring result; a status evaluation unit, connected to the operation monitoring unit, for evaluating the health status and performance level of the equipment according to the monitoring results to obtain an evaluation result; The fault repair unit is connected to the status evaluation unit and is used to perform corresponding repair or recovery operations according to the evaluation result and the fault handling strategy in the management result, and update the device status information in the management result.

4. The IoT device collaborative management system based on intelligent edge according to claim 3 is characterized in that: The fusion communication layer includes: Plug-in management module, used to load and manage independent plug-ins of various communication protocols; The protocol adaptation module is connected to the protocol plug-in loading management module, and adapts, diverts and analyzes data of different protocols according to the loaded protocol plug-in to obtain the adaptation result; A message conversion module, connected to the protocol adaptation module, converts the adaptation result into a standardized message format, and defines a standardized data packet structure to encapsulate and convert the data to obtain a conversion result; an intelligent routing module, connected to the message conversion module, for using the conversion result to dynamically select an optimal transmission path according to node status and link quality to obtain a routing result; QoS guarantee module, which is used to classify and schedule data according to the type and priority of business data to obtain guarantee results; The queue mapping module is connected to the intelligent routing module and the QoS guarantee module respectively, and maps different service flow characteristics to different queues according to the guarantee result and the routing result to obtain mapping results.

5. The IoT device collaborative management system based on intelligent edge according to claim 4 is characterized in that: The intelligent routing module includes: A path evaluation unit, configured to collect and analyze in real time the status information of each node in the network and the quality indicators of the links to obtain collection results, and to evaluate the collection results to obtain evaluation results; a routing decision unit, connected to the path evaluation unit, configured to use a routing algorithm to select an optimal transmission path according to the evaluation result to obtain a decision result; The dynamic adjustment unit is connected to the routing decision unit and is used to perform dynamic adjustment according to the decision result when the network conditions change.

6. The IoT device collaborative management system based on intelligent edge according to claim 5, characterized in that: The business logic layer includes: The microservice division module divides business rules into several microservices according to business needs, with each microservice corresponding to each domain boundary; An architecture building module, connected to the microservice partitioning module, for building a microservice architecture according to the domain boundaries; A process orchestration module is connected to the architecture building module and introduces a visual business process orchestration engine according to the microservice architecture to obtain an orchestration result; An intelligent decision-making module, connected to the process orchestration module, performs intelligent judgment on the decision points in the process according to the orchestration results and preset rules to obtain a judgment result; The event-driven module is connected to the process orchestration module and the intelligent decision-making module respectively, and is used to monitor business events and adjust the process according to the judgment results.

7. The IoT device collaborative management system based on intelligent edge according to claim 6, characterized in that: The intelligent decision-making module includes: Rule management unit, used to define, store and update decision rules to obtain management results; A decision engine unit, connected to the rule management unit, for judging the decision points in the business process according to the management results to obtain a decision result; The feedback adjustment unit is connected to the decision engine unit and is used to evaluate the decision result to obtain an evaluation result, and adjust the decision rules and processes according to the evaluation result.

8. The IoT device collaborative management system based on intelligent edge according to claim 7, characterized in that: The firmware upgrade layer includes: A state perception module is used to capture network state data in real time, including physical layer indicators, transport layer indicators, and application layer multi-dimensional indicators; an intelligent analysis module, connected to the state perception module, for performing predictions using a prediction algorithm based on the physical layer indicators, the transport layer indicators, and the application layer multi-dimensional indicators to generate an adaptive decision matrix; an intelligent execution module, connected to the intelligent analysis module, and selecting an upgrade strategy according to the decision matrix to execute the firmware upgrade to obtain an execution result; The firmware management module is connected to the intelligent execution module and is responsible for generating and applying the difference package when the intelligent execution module selects the differential upgrade strategy.

9. The IoT device collaborative management system based on intelligent edge according to claim 8, characterized in that: The intelligent analysis module includes: Data preprocessing unit, which cleans, normalizes and extracts features from the collected data to obtain processing results; a state prediction unit, connected to the data preprocessing unit, and using a long short-term memory network to predict the network state according to the processing result to obtain a prediction result; a comprehensive evaluation unit, connected to the state prediction unit, for performing a multi-dimensional evaluation and calculating a comprehensive score of the network using a comprehensive scoring formula based on the prediction result and the network state data to obtain a comprehensive evaluation result; A decision matrix generating unit is connected to the comprehensive evaluation unit and is used to generate the adaptive decision matrix according to the comprehensive evaluation result.

10. The IoT device collaborative management system based on intelligent edge according to claim 9, characterized in that: The comprehensive assessment unit includes: An indicator classification subunit, which classifies the network status data and the prediction results to extract indicators related to the comprehensive scoring formula to obtain classification results; A weight configuration subunit, connected to the indicator classification subunit, configures the weight of the classification result according to different network environments and application scenarios to obtain a configuration result; A score calculation subunit, connected to the weight configuration subunit, calculates a comprehensive score based on the comprehensive score formula and the configuration result; The level determination subunit is connected to the score calculation subunit and is used to determine the network level according to the comprehensive score and a preset network interval.

Images