A system and method for cloud-based collaborative intelligent processing applications based on heterogeneous Internet of Things (IoT)

Abstract

This invention provides a system and method for cloud-based collaborative intelligent processing applications based on heterogeneous Internet of Things (IoT). The system is characterized by being divided into a cloud side and an edge side. The cloud side primarily manages heterogeneous IoT devices and IoT processing model version libraries. The edge side, through dynamically configured and distributed IoT processing models from the cloud side, enables on-demand configuration of functions such as data access, protocol parsing, algorithm processing, and command control for heterogeneous IoT devices or systems. This invention, through standardized device model algorithm management, enhances the intelligent application capabilities of various types of edge devices in large-scale equipment integration systems, meets the management and security requirements of collaborative applications of different types of intelligent devices, and effectively improves the standardization, security, and stability of heterogeneous IoT device collaboration.

Description

Technical Field

[0001] This invention relates to a system and method for cloud-based collaborative applications based on heterogeneous Internet of Things (IoT), belonging to the field of communication technology. Specifically, it is a system and method for cloud-based collaborative intelligent processing applications based on heterogeneous Internet of Things (IoT). Background Technology

[0002] While the Internet of Things (IoT) in the power industry has a certain foundation, compared with the construction goals of the new power system digital technology support system—namely, accurate reflection, timely status, full-domain computing, and collaborative linkage—the existing smart IoT system still has shortcomings in overall construction and application, especially in the field of digital power distribution. These shortcomings are mainly reflected in two aspects: low efficiency of multi-protocol data acquisition, processing, and fusion of heterogeneous equipment on the low-voltage side, and a lack of intelligent algorithm application capabilities to cope with various edge IoT devices. This makes it difficult to support the access, processing, and computing functions of numerous large-scale IoT sensing and control devices at the park / enterprise level, which greatly limits the rapid standardization, promotion, and application of IoT devices.

[0003] Currently, the Internet of Things (IoT) devices in the power industry involve a wide variety of core intelligent equipment systems. These different types of intelligent equipment systems come from various equipment suppliers across the country, and their communication protocols, business processing, data analysis, control decisions, and management requirements vary. This results in high on-site customization costs and significant difficulties in data access and integration. There is an urgent need for a standardized management method and device to enable cloud-based collaborative intelligent processing applications for numerous types of IoT devices. Summary of the Invention

[0004] The purpose of this invention is to propose a system and method for cloud-based collaborative intelligent processing applications based on heterogeneous Internet of Things (IoT) to address the current challenges in the power industry's IoT sector. This is because the IoT devices involve a wide variety of core intelligent equipment systems from diverse suppliers, each with varying communication protocols, business processing, data analysis, control decisions, and management requirements. This leads to high on-site customization costs and significant difficulties in data access and integration. Applying this invention to the power IoT enables dynamic management of heterogeneous terminal devices, communication protocol processing, and edge processing algorithms. It improves device communication capabilities and efficiency, reduces the development and management costs of heterogeneous IoT systems on both the cloud and edge sides, and enhances the efficiency of heterogeneous IoT data fusion and deployment.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A system and method for cloud-based collaborative intelligent processing applications based on heterogeneous IoT are disclosed. The system can be divided into a cloud side and an edge side. The cloud side mainly implements the management of heterogeneous IoT devices and the version library management of IoT processing models (i.e., units or modules used to process southbound device system data), and is responsible for managing different heterogeneous IoT devices and the configuration of model algorithms for their operation. The edge side mainly uses the IoT processing models dynamically configured and distributed by the cloud side to achieve on-demand configuration of functions such as data access, protocol parsing, algorithm processing, and command control for heterogeneous IoT devices or systems. The IoT processing model refers to an independent processing unit or module that, according to predetermined protocols and rules, can receive structured, semi-structured, and unstructured IoT data input, complete data processing and edge computing such as IoT protocol processing, protocol conversion, algorithm processing, and control tasks, and finally output the calculation results data.

[0007] The cloud-side functions specifically include the following:

[0008] Based on business application requirements, the list of devices to be run is first specified; corresponding model algorithm packages and input data templates are specified according to different device types; each IoT processing model configuration is managed through task management. Before the IoT processing model configuration task is initialized, the terminal device list needs to be loaded according to the task configuration, and then the corresponding model version package and model data configuration file under the type of smart terminal are configured under the task, and sent to the corresponding end-side system through instructions; after the IoT processing model runs on the end-side, the cloud-side system receives the model task execution results of the specified end-side system through the communication protocol, and summarizes the results to the cloud-side application; it is mainly divided into: end-side device management, model version management, model association configuration, model data template configuration, and model dynamic configuration task management.

[0009] The end-side device management specifically includes: establishing end-side heterogeneous IoT device management, supporting the association between end-side devices and end-side device channels, IoT device models, IoT processing models, and IoT device numbers, supporting dynamic access to end-side device configurations, and being able to specify IoT processing models for different devices;

[0010] The model version management specifically includes: establishing an IoT processing model version management mechanism, supporting the addition, deletion, modification, and query functions of relevant model versions for the association relationship between the IoT device model and the device model data template corresponding to each device model version; supporting the management of different model versions, which can meet the needs of frequent changes in the functions of underlying IoT devices and the dynamic updates of edge-side algorithm applications, and decouple the application of a single model version from other business functions of the system during runtime.

[0011] The specific steps of the model-associated devices include: associating a list of application models with a specified end-side device, configuring a list of IoT processing models according to the actual IoT business needs of the end-side device, and then associating all IoT devices managed by the end-side device with the input and output of the application model respectively.

[0012] The model data template configuration specifically includes: supporting the setting of corresponding input data template management for different model versions, uniquely determining the input data preprocessing rules and format corresponding to a model version; generally, IoT devices access the entire amount of data, but the end-side model operation and processing does not require the entire amount of data according to actual application needs. This invention dynamically sets the model data preprocessing rule configuration, selecting only the data to be processed by the IoT processing model. After data extraction, conversion, unit conversion, normalization, cleaning and filtering, etc., the raw message data to be processed by the southbound device system received on the device bus is processed by the pre-configured parameters in the data template, such as device bus type, bus address, start bit, end bit, length, base, coefficient, offset, formula, precision, maximum and minimum range, to obtain input data that meets the processing requirements of the IoT processing model.

[0013] The dynamic model configuration task management specifically includes: creating model configuration tasks, setting the application device list, setting device model versions and data templates, and executing the function of issuing model running tasks; this is achieved through the following steps:

[0014] S1. Establish IoT processing model, configure running tasks, and manage the IoT processing model that will run on the edge.

[0015] S2. From the complete list of edge devices, specify the list of edge devices that need to run the edge model according to the device type and business application requirements (such as intelligent analysis, decision-making, control, etc.);

[0016] S3. Configure the device model version and data template for the end devices in the task list;

[0017] S4. Send the model task to the end-side device in the task list and start running the model;

[0018] S5. Receive feedback on model processing results, track the execution status of model tasks, and handle abnormal situations during model execution.

[0019] The terminal-side functions specifically include the following:

[0020] The terminal-side functions, under the control of the cloud-side system, dynamically select and load the corresponding model library version and input data template based on the type of IoT device system connected, including but not limited to distributed photovoltaic, charging piles, energy storage stations, large energy-consuming equipment, energy metering, air conditioning, smart lighting, environmental monitoring, and security monitoring equipment systems, according to the device type and business application requirements (such as intelligent analysis, decision-making, and control of device system data). They differentiate between different channels, selecting communication protocols, preprocessing rules configured for input model data templates, and intelligent algorithm version packages for processing, thereby achieving intelligent functional collaboration with various IoT device systems connected to the terminal side. Even when changes occur in the connected IoT devices, communication protocols, device IoT data, and control parameter configurations, the terminal-side basic functions remain unchanged. The system is designed for stability, minimizing the need for synchronous modifications due to changes in southbound IoT devices and processing algorithms. It enables multi-channel management access to IoT device subsystems and concurrent control of model processing. Based on the actual access status of the southbound IoT device subsystems, it dynamically creates channels, loads models, preprocesses model input data, processes model algorithms, outputs model analysis data, and outputs control commands. When an anomaly occurs in the model processing of a single type of device on the edge side through its corresponding channel, it does not affect the normal model processing of devices on other channels connected to the edge system. Through the dynamic configuration and decoupling design of the edge-side concurrent IoT device channel application functions, the efficiency and flexibility of heterogeneous edge-side IoT devices in intelligent data processing are improved, enhancing the overall intelligent computing and collaborative capabilities.

[0021] The edge-side functions include: heterogeneous IoT multi-channel control, dynamic model configuration update and model execution functions; for edge-side intelligent fusion terminals, each device channel is handled by a separate running task. During channel initialization, IoT device protocols and algorithms can be dynamically specified for each channel, supporting model algorithm and protocol configuration updates, and ensuring that different channels adopt different protocols and different algorithms for concurrent processing and execution.

[0022] The heterogeneous IoT multi-channel control specifically includes the following:

[0023] It supports dynamic loading of channel management / channel execution models by the edge smart gateway. Based on the channel model loaded by the IoT device, the channel communication protocol, data preprocessing, and intelligent algorithm of the specified model are applied to the channel. The channel protocol supports, but is not limited to, IEC101 / 104, 698, 376.1 and other energy consumption / environment monitoring protocols. The intelligent algorithm application configuration of the model can be executed and the algorithm execution results can be sent back to the cloud through the protocol extension fields. When the model is updated, the application layer function can be turned off as needed, the channel parameters and protocol can be re-initialized, and the channel device can be re-registered and reconfigured.

[0024] The dynamic update of the model configuration specifically includes the following:

[0025] The platform issues a model update command, and the client downloads the corresponding model version package and model data template from the platform. Specifically, the model can be provided through dynamic link libraries or independent processes, and the data template can be provided through XML templates, etc. The client configuration management task loads and updates the data to the specified device channel task. By pre-setting the model data template, the rules and formats for processing input data are specified for the model, so as to set corresponding input data processing rules for different model versions.

[0026] The model execution function specifically includes the following:

[0027] After completing the model configuration for the corresponding device type of this channel, the data is first preprocessed according to the data configuration template. This involves data extraction, cleaning, transformation, and processing. The raw message data to be processed from the southbound device system received on the device bus is processed using parameters such as device bus type, bus address, start bit, end bit, length, radix, coefficient, offset, formula, precision, and maximum / minimum range pre-configured in the data template. This pre-processing yields the input dataset required by the IoT processing model algorithm. Then, the model algorithm package is loaded using methods such as dynamic link libraries / inter-process calls. The function interface of the specified model algorithm is dynamically called for calculation, and the calculation results are obtained. Finally, the model algorithm execution results are uploaded to the cloud side through the IoT communication protocol extension fields.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0029] The heterogeneous IoT cloud-based collaborative intelligent processing solution provided by this invention can be widely applied to various IoT intelligent terminals, concentrators, collectors, station gateways, and downstream southbound equipment systems, such as distributed photovoltaics, charging piles, energy storage stations, large energy-consuming equipment, energy metering, air conditioning, intelligent lighting, environmental monitoring, and security monitoring equipment systems in intelligent application scenarios. Through cloud collaboration, the system dynamically configures the operating model algorithm for the end side, increasing the artificial intelligence application and intelligent collaboration capabilities of the end-side system.

[0030] This system and method, through standardized device model algorithms, enhance the intelligent application capabilities of various types of end-side devices in large-scale integrated systems (such as power systems, industrial control, and communication equipment), meeting the management and security requirements of collaborative applications of different types of intelligent devices, and effectively improving the standardization, security, and stability of heterogeneous IoT device collaboration. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the application structure of the heterogeneous IoT cloud collaborative application provided by the present invention;

[0032] Figure 2A schematic diagram of the system deployment structure for the heterogeneous IoT cloud collaborative application provided by the present invention;

[0033] Figure 3 This is a basic flowchart illustrating the application of the model algorithm in an embodiment of the present invention. Detailed Implementation

[0034] To clarify the technical problems, technical solutions, implementation processes, and performance demonstrations, the present invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative. The present invention is not intended to limit the scope of the invention. Various exemplary embodiments, features, and aspects of this disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the drawings denote elements with the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0035] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0036] Furthermore, to better illustrate this disclosure, numerous specific details are set forth in the following detailed description. Those skilled in the art will understand that this disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art have not been described in detail in order to highlight the main points of this disclosure.

[0037] Example 1

[0038] A system and method for cloud-based collaborative intelligent processing applications based on heterogeneous IoT are disclosed. The system can be divided into a cloud side and an edge side. The cloud side mainly implements heterogeneous IoT device management and IoT processing model version library management functions, and is responsible for managing different heterogeneous IoT devices and the configuration of model algorithms running on heterogeneous IoT devices. The edge side mainly uses the IoT processing model dynamically configured and distributed by the cloud side to realize on-demand configuration of functions such as data access, protocol parsing, algorithm processing, and command control for heterogeneous IoT devices or systems. The IoT processing model refers to an independent processing unit or module that can receive structured, semi-structured, and unstructured IoT data input according to predetermined protocols and rules, complete data processing and edge computing such as IoT protocol processing, protocol conversion, algorithm processing, and control tasks, and finally output the calculation result data.

[0039] The cloud-side functions specifically include the following:

[0040] Based on business application requirements, the list of devices to be run is first specified; corresponding model algorithm packages and input data templates are specified according to different device types; each IoT processing model configuration is managed through task management. Before the IoT processing model configuration task is initialized, the terminal device list needs to be loaded according to the task configuration, and then the corresponding model version package and model data configuration file under the type of smart terminal are configured under the task, and sent to the corresponding end-side system through instructions; after the IoT processing model runs on the end-side, the cloud-side system receives the model task execution results of the specified end-side system through the communication protocol, and summarizes the results to the cloud-side application; it is mainly divided into: end-side device management, model version management, model association configuration, model data template configuration, and model dynamic configuration task management.

[0041] The end-side device management specifically includes: establishing end-side heterogeneous IoT device management, supporting the association between end-side devices and end-side device channels, IoT device models, IoT processing models, and IoT device numbers, supporting dynamic access to end-side device configurations, and being able to specify IoT processing models for different devices;

[0042] The model version management specifically includes: establishing an IoT processing model version management mechanism, supporting the addition, deletion, modification, and query functions of relevant model versions for the association relationship between the IoT device model and the device model data template corresponding to each device model version; supporting the management of different model versions, which can meet the needs of frequent changes in the functions of underlying IoT devices and the dynamic updates of edge-side algorithm applications, and decouple the application of a single model version from other business functions of the system during runtime.

[0043] The specific steps of the model-associated devices include: associating a list of application models with a specified end-side device, configuring a list of IoT processing models according to the actual IoT business needs of the end-side device, and then associating all IoT devices managed by the end-side device with the input and output of the application model respectively.

[0044] The model data template configuration specifically includes: supporting the setting of corresponding input data template management for different model versions, uniquely determining the input data preprocessing rules and format corresponding to a model version; generally, IoT devices access the entire amount of data, but the end-side model operation and processing does not require the entire amount of data according to actual application needs. This invention dynamically sets the model data preprocessing rule configuration, selecting only the data to be processed by the IoT processing model. After data extraction, conversion, unit conversion, normalization, cleaning and filtering, etc., the raw message data to be processed by the southbound device system received on the device bus is processed by the pre-configured parameters in the data template, such as device bus type, bus address, start bit, end bit, length, base, coefficient, offset, formula, precision, maximum and minimum range, to obtain input data that meets the processing requirements of the IoT processing model.

[0045] The dynamic model configuration task management specifically includes: creating model configuration tasks, setting the application device list, setting device model versions and data templates, and executing the function of issuing model running tasks; this is achieved through the following steps:

[0046] S1. Establish IoT processing model, configure running tasks, and manage the IoT processing model that will run on the edge.

[0047] S2. From the complete list of edge devices, specify the list of edge devices that need to run the edge model according to the device type and business application requirements (such as intelligent analysis, decision-making, control, etc.);

[0048] S3. Configure the device model version and data template for the end devices in the task list;

[0049] S4. Send the model task to the end-side device in the task list and start running the model;

[0050] S5. Receive feedback on model processing results, track the execution status of model tasks, and handle abnormal situations during model execution.

[0051] The terminal-side functions specifically include the following:

[0052] The terminal-side functions, under the control of the cloud-side system, dynamically select and load the corresponding model library version and input data template based on the type of IoT device system accessed, including but not limited to distributed photovoltaic, charging piles, energy storage stations, large energy-consuming equipment, energy metering, air conditioning, smart lighting, environmental monitoring, and security monitoring equipment systems, according to the device type and business application requirements (such as intelligent analysis, decision-making, and control of device system data). It differentiates between different channels, selecting communication protocols, preprocessing rules configured for input model data templates, and intelligent algorithm version packages for processing. This enables intelligent functional collaboration with various IoT device systems connected to the terminal side. The dynamic selection is primarily based on two factors: first, the type of device accessed from the south (different models are selected for different devices); second, the application from the north (device control or device data analysis, for example, different models are used for controlling air conditioning, elevators, and lighting, and for analyzing security alarms and data). The models for lane departure in autonomous driving also differ; the system maintains the stability of the basic application version on the end side when the connected IoT devices, communication protocols, device IoT data, and control parameter configurations change, minimizing the need for synchronous modifications to the system due to changes in southbound IoT devices and processing algorithms; it enables multi-channel management access to IoT device subsystems and concurrent control of model processing, and dynamically creates channels, loads models, preprocesses model input data, processes model algorithms, outputs model analysis, and outputs control commands based on the actual access situation of the southbound IoT device subsystems. When an anomaly occurs in the model processing of a single type of device on the end side through its corresponding channel, it does not affect the normal model processing of devices on other channels connected to the end side system; through the dynamic configuration and decoupling design of the concurrent IoT device channel application functions on the end side, the efficiency and flexibility of heterogeneous IoT devices in intelligent data processing can be improved, enhancing the overall intelligent computing collaboration capability.

[0053] The edge-side functions include: heterogeneous IoT multi-channel control, dynamic model configuration update and model execution functions; for edge-side intelligent fusion terminals, each device channel is handled by a separate running task. During channel initialization, IoT device protocols and algorithms can be dynamically specified for each channel, supporting model algorithm and protocol configuration updates, and ensuring that different channels adopt different protocols and different algorithms for concurrent processing and execution.

[0054] The heterogeneous IoT multi-channel control specifically includes the following:

[0055] It supports dynamic loading of channel management / channel execution models by the edge smart gateway. Based on the channel model loaded by the IoT device, the channel communication protocol, data preprocessing, and intelligent algorithm of the specified model are applied to the channel. The channel protocol supports, but is not limited to, IEC101 / 104, 698, 376.1 and other energy consumption / environment monitoring protocols. The intelligent algorithm application configuration of the model can be executed and the algorithm execution results can be sent back to the cloud through the protocol extension fields. When the model is updated, the application layer function can be turned off as needed, the channel parameters and protocol can be re-initialized, and the channel device can be re-registered and reconfigured.

[0056] The dynamic update of the model configuration specifically includes the following:

[0057] The platform issues a model update command, and the client downloads the corresponding model version package and model data template from the platform. Specifically, the model can be provided through dynamic link libraries or independent processes, and the data template can be provided through XML templates, etc. The client configuration management task loads and updates the data to the specified device channel task. By pre-setting the model data template, the rules and formats for processing input data are specified for the model, so as to set corresponding input data processing rules for different model versions.

[0058] The model execution function specifically includes the following:

[0059] After completing the model configuration for the corresponding device type of this channel, the data is first preprocessed according to the data configuration template. Based on the device requirements and application needs, data extraction, cleaning, transformation, and processing are performed. The raw message data to be processed from the southbound device system received on the device bus is processed using parameters such as device bus type, bus address, start bit, end bit, length, radix, coefficient, offset, formula, precision, and maximum and minimum range pre-configured in the data template. This pre-processing yields the input dataset required by the IoT processing model algorithm. Then, the model algorithm package is loaded through methods such as dynamic link libraries / inter-process calls. The function interface of the specified model algorithm is dynamically called to perform calculations and obtain the results. Finally, the model algorithm execution results are uploaded to the cloud side through the IoT communication protocol extension fields.

[0060] like Figure 1 The diagram illustrates the system application structure in this embodiment of the invention, which can be specifically divided into cloud-side and edge-side applications. The cloud-side application primarily manages heterogeneous IoT devices and the IoT processing model version library. The edge-side application, mainly located on the smart gateway, utilizes the IoT processing model dynamically configured and distributed from the cloud side to achieve functions such as data access, protocol parsing, algorithm processing, and outputting control commands or analysis results for heterogeneous IoT devices or systems.

[0061] The edge-side functionality supports dynamically selecting and loading the appropriate model library version and input data template based on different device types. It also supports different channels for selecting communication protocols, input preprocessing rules, and intelligent algorithm processing, enabling intelligent functional collaboration with various device systems connected to the edge. It maintains the stability of the edge-side basic application version even when connected IoT devices, communication protocols, device IoT data, and control parameter configurations change, minimizing the need for system synchronization modifications due to changes in southbound IoT devices and processing algorithms. It enables multi-channel management of access device subsystems and concurrent model control processing. Based on the actual access status of southbound device subsystems, it dynamically implements functions such as channel establishment, model loading, model input data preprocessing, model algorithm processing, model analysis output, and control command output. When one type of device is processed through its corresponding channel model, it does not affect the normal model processing of devices on other channels connected to the edge system. Through the edge-side concurrent multi-channel functional decoupling design, it improves the efficiency and flexibility of intelligent data processing for heterogeneous IoT devices on the edge, enhancing the overall intelligent computing collaboration capability.

[0062] like Figure 2 The diagram illustrates the system deployment structure in this embodiment of the invention. The cloud-side platform includes IoT device management (201) and IoT processing model version management (202). IoT device management provides IoT device configuration management, IoT device channel management, and maintains the correspondence between IoT devices, model versions, and model configuration data. The IoT processing model version is an independent processing module for various IoT devices or systems, handling communication protocol processing, input data preprocessing, algorithm computation, edge computing, and control output. It supports different model version management, ensuring decoupling from other business functions within the system when IoT device data changes or algorithm applications are updated.

[0063] The edge system uses the intelligent terminal 203 to achieve dynamic model configuration and intelligent processing collaboration for numerous types of southbound device subsystems 204. It supports dynamically selecting and loading the corresponding model library version and input data template according to different types of connected devices, and supports distinguishing between different channels to select communication protocols, preprocessing rules, and manual algorithm processing, thereby realizing intelligent functional collaboration with various device systems connected to the edge.

[0064] like Figure 3 The diagram shows the basic flowchart of the model algorithm application provided in this embodiment of the invention. This invention illustrates a cloud-based intelligent collaboration method, which includes:

[0065] The cloud-based intelligent collaborative platform system channel configuration 301 provided in this embodiment of the invention firstly manages multiple independent IoT device systems under the end-side devices in a unified manner on the cloud-side platform. Through IoT device type and device model, IoT device management configures communication channels for the end-side devices and initializes relevant resources and configurations for accessing southbound IoT device data.

[0066] The model configuration 302 provided in this embodiment of the invention supports different model version management. It is necessary to specify the specific associated IoT processing model or model list for each channel of the specified end device to meet the needs of southbound IoT device function, protocol changes, end-side algorithm application configuration and update, and decouple it from other business functions of the system.

[0067] The model data configuration 303 provided in this embodiment of the invention unifies the data processing rules and formats for model input, and supports setting corresponding data input rule management for different model versions. Generally, IoT devices access the system with full data, but the edge model operation and processing does not require full data. This invention dynamically sets the model data preprocessing rule configuration, selects the southbound device system data required by the model (specifically depending on which systems are accessed from the south, including but not limited to photovoltaic system data, charging pile system data, energy storage system data, air conditioning system data, lighting system data, etc.), and processes it through data conversion, unit conversion, normalization, cleaning and filtering, etc. After processing the data rules, the raw message data to be processed from the southbound device system received on the device bus is processed by the pre-configured parameters in the data template, such as device bus type, bus address, start bit, end bit, length, base, coefficient, offset, formula, precision, maximum and minimum range, to obtain input data that meets the processing requirements of the IoT processing model.

[0068] The model task issuance 304 provided in this embodiment of the invention includes the following implementation steps:

[0069] S1. Establish IoT processing model, configure running tasks, and manage the IoT processing model that will run on the edge.

[0070] S2. From the complete list of end-side devices, specify the list of end-side devices that need to run the end-side model;

[0071] S3. Set the device model version and data template for the end devices in the task list;

[0072] S4. Send the model task to the end-side device in the task list and start running the model;

[0073] S5. Receive feedback on model processing results, track the execution status of model tasks, and handle abnormal situations during model execution.

[0074] The channel dynamic loading 305 provided in this embodiment dynamically loads the channel management / execution model through an end-side smart gateway, and specifies the channel communication protocol, data preprocessing, and intelligent algorithm application method according to the channel model loaded by the IoT device. In this embodiment, the channel protocol supports, but is not limited to, IEC101 / 104, 698, 376.1, and other energy consumption / environment monitoring protocols. The protocol extension fields are used to execute the model's intelligent algorithm application configuration and transmit the algorithm execution results back to the cloud. When the model is updated, the application layer function can be disabled as needed, and the channel parameters and protocol can be re-initialized, allowing the channel device to re-register and reconfigure its parameters.

[0075] The model data input 306 provided in this embodiment of the invention allows the terminal to download the corresponding version of the model algorithm package and model data input template from the platform. The terminal configuration management task then loads and updates the data into the designated device channel task. Through pre-set data processing rules and formats, such as XML configuration, it supports setting corresponding data input rules for different model versions. The terminal dynamically sets model data preprocessing rule configurations, selects the data required by the IoT processing model, and processes it through data conversion, unit conversion, normalization, cleaning, and filtering. The raw message data to be processed from the southbound device system received on the device bus is processed using pre-configured parameters in the data template, such as device bus type, bus address, start bit, end bit, length, radix, coefficient, offset, formula, precision, and maximum / minimum range, to obtain input data that meets the processing requirements of the IoT processing model.

[0076] The model algorithm processing 307 provided in this embodiment of the invention loads the model algorithm package through methods including but not limited to dynamic link libraries / inter-process calls, dynamically calls the function interface of the specified model algorithm to perform calculations, and obtains the calculation results.

[0077] The model output 308 provided in this embodiment of the invention uploads the model algorithm execution result to the cloud via an extended field in the Internet of Things communication protocol.

[0078] The model output 309 provided in this embodiment of the invention receives the execution results of the model task from the specified end-side system on the cloud side and summarizes the results to the cloud-side application for further processing. During the execution process on the end-side, the cloud side periodically monitors the execution status of the model task and handles any abnormal situations during model execution.

[0079] The termination judgment 310 provided in this embodiment of the invention allows the end-side channel to determine whether to terminate. If all processing model tasks are completed, or if internal or external abnormalities cause termination, or if the cloud side actively terminates the process, resulting in the inability to continue model processing, then channel unloading 311 is executed to return the end-side system resources occupied by IoT device access. If it is determined that the processing model tasks still need to be processed, then the model processing tasks continue to be executed. After channel unloading, the cloud side can choose to load a new IoT processing model.

[0080] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A cloud-based collaborative intelligent processing application system based on heterogeneous Internet of Things (IoT), characterized in that, The system is divided into cloud-side and edge-side. The cloud-side implements heterogeneous IoT device management and IoT processing model version library management functions, and is responsible for managing different heterogeneous IoT devices and the configuration of model algorithms for running on heterogeneous IoT devices. The edge-side, through the IoT processing model dynamically configured and distributed by the cloud-side, can realize on-demand configuration of data access, protocol parsing, algorithm processing, and command control functions for heterogeneous IoT devices or systems. The IoT processing model refers to an independent processing unit or module that, according to established protocols and rules, can receive structured, semi-structured, and unstructured IoT data input, complete IoT protocol processing, protocol conversion, algorithm processing, control task data processing, and edge computing, and finally realize the output of calculation result data. The cloud-side functions specifically include the following: Based on business application requirements, the first step is to specify the list of running edge devices; then, based on different device types, specify the corresponding model algorithm packages and input data templates; before the IoT processing model configuration task is initialized, each IoT processing model configuration is managed through task management. First, the terminal device list is loaded according to the task configuration, and then the corresponding model version package and model data configuration file in the configuration task are configured according to different terminal types, and sent to the corresponding edge system via instructions; after the IoT processing model runs on the edge, the cloud system receives the execution results of the specified edge system model task through the communication protocol, and summarizes the results to the cloud application; It is divided into: edge device management, model version management, model association configuration, model data template configuration, and model dynamic configuration task management; The end-side device management specifically includes: establishing end-side heterogeneous IoT device management, supporting the association between end-side devices and end-side device channels, IoT device models, IoT processing models, and IoT device numbers, supporting dynamic access to end-side device configurations, and being able to specify IoT processing models for different devices; The model association configuration specifically includes: associating a list of application models with a specified end-side device, configuring a list of IoT processing models according to the actual IoT business needs of the end-side device, and then associating all IoT devices managed by the end-side device with the input and output of the application model respectively.

2. The cloud-based collaborative intelligent processing application system based on heterogeneous Internet of Things as described in claim 1, characterized in that, The model version management specifically includes: establishing an IoT processing model version management mechanism, supporting the addition, deletion, modification, and query functions of relevant model versions for the association relationship between the IoT device model and the device model data template corresponding to each device model version; supporting the management of different model versions, which can meet the needs of frequent changes in the functions of underlying IoT devices and the dynamic updates of edge-side algorithm applications, and decoupling the runtime of a single model version application from the system business functions.

3. The cloud-based collaborative intelligent processing application system based on heterogeneous Internet of Things as described in claim 1, characterized in that, The model data template configuration specifically includes: supporting the setting of corresponding input data template management for different model versions, uniquely determining the input data preprocessing rules and format corresponding to a model version; by dynamically setting the model data preprocessing rule configuration, only the data to be processed by the IoT processing model is selected, and after data extraction, conversion, unit conversion, normalization, cleaning and filtering data processing rules, the raw message data to be processed by the southbound device system received on the device bus is processed through the device bus type, bus address, start bit, end bit, length, base, coefficient, offset, formula, precision, maximum and minimum range parameters pre-configured in the data template, and the input data meets the processing requirements of the IoT processing model.

4. The cloud-based collaborative intelligent processing application system based on heterogeneous Internet of Things as described in claim 1, characterized in that, The dynamic model configuration task management specifically includes: creating model configuration tasks, setting the application device list, setting device model versions and data templates, and executing the function of issuing model running tasks; this is achieved through the following steps: S1. Establish IoT processing model, configure running tasks, and manage the IoT processing model that will run on the edge. S2. From the complete list of terminal devices, specify the list of terminal devices corresponding to the terminal operation model according to the device type and business application requirements. S3. Configure the device model version and data template for the end devices in the task list; S4. Send the model task to the end-side device in the task list and start running the model; S5. Receive feedback on model processing results, track the execution status of model tasks, and handle abnormal situations during model execution.

5. The cloud-based collaborative intelligent processing application system based on heterogeneous Internet of Things as described in claim 1, characterized in that, The terminal-side functions specifically include the following: The terminal-side functions, under the control of the cloud-side system, dynamically select and load the corresponding model library version and input data template based on the IoT device system type and business application requirements. They differentiate between different channels, selecting communication protocols, preprocessing rules configured for input model data templates, and intelligent algorithm version packages for processing. This enables intelligent functional collaboration with various IoT device systems connected to the terminal. Even when the connected IoT devices, communication protocols, device IoT data, and control parameter configurations change, the terminal-side basic application version remains stable, reducing the need for synchronous modifications to the system due to changes in southbound IoT devices and processing algorithms. This allows for multi-channel management and access of IoT devices. The system integrates a connected device subsystem and a concurrent model processing control mechanism. Based on the actual access status of the southbound IoT device subsystem, it dynamically creates channels, loads models, preprocesses model input data, processes model algorithms, analyzes and outputs model data, and outputs control commands. When an anomaly occurs in the model processing of a single type of device on the edge side through its corresponding channel, it does not affect the normal model processing of devices on the channels connected to the edge system. Through the dynamic configuration and decoupling design of the edge-side concurrent IoT device channel application functions, the efficiency and flexibility of heterogeneous IoT devices in intelligent data processing can be improved, enhancing the overall intelligent computing and collaborative capabilities. The edge-side functions include: heterogeneous IoT multi-channel control, dynamic model configuration update and model execution functions; for edge-side intelligent fusion terminals, each device channel is handled by a separate running task. During channel initialization, IoT device protocols and algorithms are dynamically specified for each channel, supporting model algorithm and protocol configuration updates, and ensuring that different channels adopt different protocols and different algorithms for concurrent processing and execution.

6. The cloud-based collaborative intelligent processing application system based on heterogeneous Internet of Things as described in claim 5, characterized in that, The heterogeneous IoT multi-channel control specifically includes the following: It supports dynamic loading of channel management / channel execution models by the edge-side smart gateway. Based on the channel model loaded by the IoT device, the channel communication protocol, data preprocessing, and intelligent algorithm of the specified model are applied to the channel. The channel protocol supports IEC101 / 104, 698, 376.1 and energy consumption / environmental monitoring protocols. The intelligent algorithm application configuration of the model is executed and the algorithm execution results are sent back to the cloud through the protocol extension fields. When the model is updated, the application layer function is turned off, the channel parameters and protocol are re-initialized, and the channel device is re-registered and the parameters are reconfigured.

7. The cloud-based collaborative intelligent processing application system based on heterogeneous Internet of Things as described in claim 5, characterized in that, The dynamic update of the model configuration specifically includes the following: The platform issues a model update command, and the client downloads the corresponding model version package and model data template from the platform. Specifically, the model is provided through a dynamic link library or an independent process, and the data template is provided through an XML template. The client configuration management task loads and updates the data to the specified device channel task. By pre-setting the model data template, the rules and formats for processing input data are specified for the model, so as to set corresponding input data processing rules for different model versions.

8. The cloud-based collaborative intelligent processing application system based on heterogeneous Internet of Things as described in claim 5, characterized in that, The model execution function specifically includes the following: After completing the model configuration for the corresponding device type of this channel, the data is first preprocessed according to the data configuration template. This involves data extraction, cleaning, transformation, and processing. The raw message data to be processed from the southbound device system received on the device bus is processed using the device bus type, bus address, start bit, end bit, length, radix, coefficient, offset, formula, precision, and maximum / minimum range parameters pre-configured in the data template. This pre-processing yields the input dataset for the IoT processing model algorithm. Then, the model algorithm package is loaded via dynamic link library / inter-process call, and the function interface of the specified model algorithm is dynamically called for calculation, and the calculation results are obtained. Finally, the model algorithm execution results are uploaded to the cloud side through the IoT communication protocol extension fields.

Images