A multi-scenario data acquisition application method and system

Abstract

The present invention proposes a multi-scenario data collection application method and system. The multi-scenario data collection application method includes: the Internet of Things platform determines the scenario requirements corresponding to the current data collection, and starts the terminal device cluster corresponding to the scenario requirements to perform data collection according to the scenario requirements; the main edge node identifies the scene requirement change state in real time, uses the sub-edge node cluster to control the data collection of the terminal device cluster according to the scene requirement change state, and feeds back the processed scene data to the Internet of Things platform through the main edge node; and dynamically adjusts the operation strategy of the sub-edge node cluster corresponding to each scenario requirement according to the load status of the sub-edge node. The system includes modules corresponding to the steps of the method.

Description

Technical Field

[0001] The present invention proposes a multi-scenario data acquisition application method and system, belonging to the technical field of data acquisition control. Background Art

[0002] With the rapid development of information technology, the Internet of Things (IoT) has been widely applied across various industries, from smart homes to smart cities, from industrial control to environmental monitoring. The IoT is gradually changing how people live and work. The core of the IoT lies in data collection, transmission, processing, and application, with data collection being the foundation of the entire IoT system. However, with the increasing diversity and complexity of application scenarios, traditional single-scenario data collection methods are no longer able to meet practical needs.

[0003] Data collection in multi-scenario applications faces numerous challenges. First, different scenarios have varying data requirements and collection methods. For example, environmental monitoring may require high-frequency, high-precision data collection, while smart homes may prioritize low-power, long-life data collection. Therefore, flexibly adjusting data collection strategies based on scenario requirements has become a pressing issue. Second, with the surge in the number of IoT devices, data volume and complexity have also increased, placing higher demands on the real-time and accuracy of data collection. Traditional centralized data collection methods are prone to data congestion and latency, impacting data application effectiveness. Therefore, it is necessary to explore more efficient and distributed data collection methods. Furthermore, the rise of edge computing technology offers new insights into IoT data collection. By decentralizing computing power to the edge of the network, edge computing can significantly reduce data transmission latency and improve data processing efficiency. However, in multi-scenario applications, effectively utilizing edge computing resources to achieve intelligent data collection and control for clusters of terminal devices remains a technical challenge. While some data collection methods have been proposed for specific scenarios, these methods often lack versatility and flexibility, making them difficult to meet the needs of multi-scenario applications. At the same time, these methods also have shortcomings in data processing and resource management, making it difficult to fully leverage the advantages of the Internet of Things and edge computing technologies.

[0004] Therefore, it is necessary to propose a multi-scenario data collection application method, which can flexibly adjust the data collection strategy according to the scenario requirements, utilize edge computing resources to achieve efficient and real-time data collection and processing, and dynamically adjust the operation strategy according to the load status of the sub-edge node to improve the efficiency and accuracy of data collection. Summary of the Invention

[0005] The present invention provides a multi-scenario data acquisition application method and system to solve the technical problems existing in the above-mentioned prior art. The technical solutions adopted are as follows:

[0006] A multi-scenario data collection application method, the multi-scenario data collection application method comprising:

[0007] The IoT platform determines the scenario requirements corresponding to the current data collection, and starts the terminal device cluster corresponding to the scenario requirements to collect data according to the scenario requirements;

[0008] The master edge node identifies the scene demand change state in real time, uses the sub-edge node cluster to control the data collection of the terminal device cluster according to the scene demand change state, and feeds back the processed scene data to the Internet of Things platform through the master edge node;

[0009] The operation strategy of the sub-edge node cluster corresponding to each scenario requirement is dynamically adjusted according to the load status of the sub-edge node.

[0010] Furthermore, the IoT platform determines the scenario requirements corresponding to the current data collection, and starts the terminal device cluster corresponding to the scenario requirements to perform data collection according to the scenario requirements, including:

[0011] The IoT platform monitors the scenario demand instructions sent by the user end in real time;

[0012] The IoT platform obtains the scene type currently requiring data collection based on the scene requirement instruction sent by the user; wherein the scene types include industrial equipment monitoring scene, industrial environment monitoring scene, energy consumption monitoring scene, and warehousing and logistics monitoring scene;

[0013] The IoT platform starts the terminal device cluster corresponding to the scenario type according to the scenario type that currently requires data collection.

[0014] Furthermore, the master edge node can identify the change of scene requirements in real time, including:

[0015] The terminal device cluster collects data information corresponding to the scene type in real time according to its corresponding scene type, and sends the data information to the main edge node;

[0016] The master edge node performs data analysis on the data information to determine whether there is any abnormal scene operation event under each scene type;

[0017] When it is determined that there is a scene operation abnormality event under the scene type, it is determined that there is a scene demand state change under the current scene type, and a data acquisition control adjustment instruction is sent to the sub-edge node cluster.

[0018] Furthermore, according to the scenario requirement change state, the sub-edge node cluster is used to control the data collection of the terminal device cluster, including:

[0019] When each sub-edge node in the sub-edge node cluster receives the data acquisition control adjustment instruction, the sub-edge node extracts the data information collected by each corresponding terminal device;

[0020] Obtaining data difference information between the data information corresponding to the time when the abnormal event occurs and the data information before the time when the abnormal event occurs based on the data information collected by each terminal device;

[0021] Compare the data difference information corresponding to each terminal device with a preset deviation reference value;

[0022] The terminal device whose data difference information exceeds the preset deviation reference value is used as the main control device;

[0023] The terminal devices whose data difference information does not exceed the preset deviation reference value are used as auxiliary control devices;

[0024] Obtaining a data deviation coefficient using the data difference information of the main control device and the data difference information of the auxiliary control device;

[0025] Comparing the data deviation coefficient with a preset data deviation coefficient threshold;

[0026] When the data deviation coefficient exceeds a preset data deviation coefficient threshold, the data collection frequency of the terminal device is regulated at a first level;

[0027] When the data deviation coefficient does not exceed the preset data deviation coefficient threshold, the data collection frequency of the terminal device is regulated at the secondary level.

[0028] Furthermore, obtaining a data deviation coefficient by using the data difference information of the primary control device and the data difference information of the secondary control device includes:

[0029] Extracting data difference information of the main control device and the weight coefficient of the main control device;

[0030] Extracting data difference information of auxiliary control equipment and weight coefficients of auxiliary control equipment;

[0031] Obtaining a first deviation factor using the data difference information of the main control device and a weight coefficient of the main control device;

[0032] Obtaining a second deviation factor using the data difference information of the auxiliary control device and a weight coefficient of the auxiliary control device;

[0033] A data deviation coefficient is obtained using the first deviation factor and the second deviation factor.

[0034] Furthermore, the primary regulation includes:

[0035] When the data deviation coefficient exceeds a preset data deviation coefficient threshold, the sub-edge node cluster sends data information corresponding to the terminal device at the time when the abnormal event occurs to the master edge node;

[0036] The master edge node generates a scene intention vector through a built-in multimodal intention graph parsing model;

[0037] The master edge node generates a dynamic acquisition strategy matrix according to the scenario intention vector; wherein the dynamic acquisition strategy matrix includes device acquisition control parameters and sub-edge node cluster operation regulation parameters; wherein the device acquisition control parameters include data acquisition mode and terminal device acquisition control parameters; the sub-edge node cluster operation regulation parameters include task allocation weights and resource constraints of the sub-edge nodes; wherein the task allocation weights of the sub-edge nodes are set according to data difference information of the terminal devices corresponding to the sub-edge nodes; and the resource constraints of the sub-edge nodes are set according to the task allocation weights of the sub-edge nodes;

[0038] The master edge node sends the device acquisition control parameter to the sub-edge node cluster;

[0039] The sub-edge node cluster performs data collection and control on its corresponding terminal device according to the device collection control parameters;

[0040] The master edge node controls the operating state of the sub-edge node cluster according to the sub-edge node cluster operating control parameters.

[0041] Furthermore, the secondary regulation includes:

[0042] When the data deviation coefficient does not exceed the preset data deviation coefficient threshold, the sub-edge node determines whether the terminal device it is responsible for controlling generates data difference information at the time when the abnormal event occurs;

[0043] When the sub-edge node identifies that there is a terminal device that generates data difference information at the time when the abnormal event occurs among the terminal devices it is responsible for controlling, the terminal device that generates data difference information at the time when the abnormal event occurs is used as the target terminal device;

[0044] Get the data collection frequency and CPU usage of the target terminal device during data collection;

[0045] Get the CPU usage of the sub-edge node corresponding to the target terminal device during the terminal device acquisition control process;

[0046] Ratio processing is performed on the CPU occupancy rate of the target terminal device and the CPU occupancy rate of the sub-edge node corresponding to the target terminal device to obtain a CPU occupancy rate ratio parameter;

[0047] The data collection frequency of the target terminal device is adjusted by utilizing the CPU occupancy ratio parameter in combination with the weight coefficient corresponding to the target terminal device.

[0048] Furthermore, the operation strategy of the sub-edge node cluster corresponding to each scenario requirement is dynamically adjusted according to the load status of the sub-edge node, including:

[0049] Monitor the load status data of each sub-edge node in real time, wherein the load status data includes CPU occupancy, memory occupancy, and network communication delay ratio between the sub-edge node and the master edge node;

[0050] Normalizing the load status data of the edge node to obtain the normalized load status data;

[0051] Generate a load state vector corresponding to each unit time according to the normalized load state data corresponding to each unit time of the sub-edge node; wherein, the load state vector U=[G c , G n , G y ], and G c , G n and G y Respectively represent the CPU usage, memory usage, and network communication delay ratio between the sub-edge node and the main edge node corresponding to each unit time;

[0052] Obtaining the Euclidean norm corresponding to the load state vector according to the load state vector corresponding to each unit time;

[0053] The Euclidean norm of the load state vector corresponding to each unit time is used to determine whether the task allocation weight of the sub-edge node needs to be adjusted.

[0054] Furthermore, whether the task allocation weight of the sub-edge node needs to be adjusted is determined based on the Euclidean norm of the load state vector corresponding to each unit time, including:

[0055] Obtaining a difference in the Euclidean norms of the load state vectors corresponding to every two adjacent unit times by using the Euclidean norm of the load state vector corresponding to the unit time;

[0056] comparing the Euclidean norm of the load state vector corresponding to the unit time with a preset norm threshold;

[0057] comparing the Euclidean norm difference with a preset difference threshold;

[0058] When the continuous duration of the Euclidean norm of the load state vector exceeding the preset norm threshold exceeds the preset time length threshold, or the number of load state vector groups corresponding to each two adjacent unit times for which the Euclidean norm difference exceeds the preset difference threshold exceeds the preset group number threshold, the task allocation weight of the sub-edge node is adjusted using the Euclidean norm of the load state vector corresponding to the unit time.

[0059] A multi-scenario data acquisition application system is used to execute the multi-scenario data acquisition application method described above, characterized in that the multi-scenario data acquisition application system includes:

[0060] The data collection startup module is used by the IoT platform to determine the scenario requirements corresponding to the current data collection, and to start the terminal device cluster corresponding to the scenario requirements to perform data collection according to the scenario requirements;

[0061] A data acquisition and control module is used to identify the scene demand change state in real time through the main edge node, use the sub-edge node cluster to control the data acquisition of the terminal device cluster according to the scene demand change state, and feed back the processed scene data to the Internet of Things platform through the main edge node cluster;

[0062] The operation strategy dynamic adjustment module is used to dynamically adjust the operation strategy of the sub-edge node cluster corresponding to each scenario requirement according to the load status of the sub-edge node.

[0063] Beneficial effects of the present invention:

[0064] The present invention proposes a multi-scenario data acquisition application method and system that determines scenario requirements through an Internet of Things platform and activates a corresponding terminal device cluster for data acquisition. The system can flexibly adjust the data acquisition strategy according to the actual application scenario, thereby improving the accuracy and practicality of data acquisition. By utilizing a cluster of master edge nodes and sub-edge nodes for data acquisition and control, the system can achieve efficient and real-time data acquisition and processing. The master edge node is responsible for real-time identification of changes in scenario requirements and control adjustments, while the sub-edge node is responsible for specific data acquisition task execution and resource scheduling optimization. By dynamically adjusting the operating strategy according to the load status of the sub-edge node, the system can maintain an efficient and stable operating state. At the same time, as the application scenarios expand and the number of terminal devices increases, the system can expand data acquisition and processing capabilities by adding sub-edge nodes, thereby meeting a wider range of application needs. By achieving local processing and storage of data through edge computing technology, the system can significantly reduce data transmission delays and costs. This is particularly important for application scenarios that require high-frequency, high-precision data acquisition. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 A flow chart of the method of the present invention;

[0066] Figure 2 This is a system block diagram of the system of the present invention. DETAILED DESCRIPTION

[0067] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0068] The embodiment of the present invention proposes a multi-scenario data collection application method, such as Figure 1 As shown, the multi-scenario data collection application method includes:

[0069] S1. The IoT platform determines the scenario requirements corresponding to the current data collection, and starts the terminal device cluster corresponding to the scenario requirements to collect data according to the scenario requirements;

[0070] S2. Using the master edge node to identify the scene demand change state in real time, using the sub-edge node cluster to control data acquisition of the terminal device cluster according to the scene demand change state, and feeding back the processed scene data to the Internet of Things platform through the master edge node;

[0071] S3. Dynamically adjust the operation strategy of the sub-edge node cluster corresponding to each scenario requirement according to the load status of the sub-edge node.

[0072] The working principle of the above technical solution is as follows: The IoT platform, serving as the hub of the entire data collection system, first determines the current data collection requirements based on the actual application scenario. Based on the determined requirements, the IoT platform activates the corresponding cluster of terminal devices to collect data. These terminal devices may include sensors, cameras, smart meters, etc., and they collect data according to preset parameters and instructions. During the data collection process, the master edge node is responsible for identifying changes in the scenario requirements in real time. This is achieved by analyzing data uploaded by the terminal device cluster or combining other external information. When the master edge node detects a change in scenario requirements, it uses the sub-edge node cluster to control data collection for the terminal device cluster based on the new requirements. This includes adjusting the frequency and accuracy of data collection or initiating new data collection tasks. The master edge node also performs preliminary processing on the collected data, such as data cleaning and format conversion, and then feeds the processed scenario data back to the IoT platform. To improve the efficiency and stability of the data collection system, the system dynamically adjusts its operating strategy based on the load status of the sub-edge nodes. If the load on a sub-edge node is too high, the system may transfer some of its data collection tasks to other sub-edge nodes with lower loads to achieve load balancing. At the same time, the system will also schedule and optimize resources based on the performance status of the sub-edge nodes to ensure the smooth progress of data collection tasks.

[0073] The above technical solution achieves the following: By using the IoT platform to identify scenario requirements and activate corresponding terminal device clusters for data collection, the system can flexibly adjust data collection strategies based on actual application scenarios, thereby improving data collection accuracy and practicality. Utilizing a cluster of master edge nodes and sub-edge nodes for data collection and control, the system achieves efficient and real-time data collection and processing. The master edge node is responsible for real-time identification of changing scenario requirements and making control adjustments, while the sub-edge nodes are responsible for executing specific data collection tasks and optimizing resource scheduling. Dynamically adjusting operating strategies based on the load status of the sub-edge nodes ensures efficient and stable operation. Furthermore, as application scenarios expand and the number of terminal devices increases, the system can expand data collection and processing capabilities by adding sub-edge nodes to meet a wider range of application needs. By enabling localized data processing and storage through edge computing technology, the system can significantly reduce data transmission latency and costs. This is particularly important for application scenarios requiring high-frequency, high-precision data collection.

[0074] In one embodiment of the present invention, an IoT platform determines a scenario requirement corresponding to current data collection, and activates a terminal device cluster corresponding to the scenario requirement to perform data collection according to the scenario requirement, including:

[0075] S101, the IoT platform monitors the scenario demand instructions sent by the user end in real time;

[0076] S102. The IoT platform obtains the scene type currently requiring data collection based on the scene requirement instruction sent by the user; wherein the scene types include industrial equipment monitoring scene, industrial environment monitoring scene, energy consumption monitoring scene, and warehousing and logistics monitoring scene;

[0077] S103. The Internet of Things platform starts a terminal device cluster corresponding to the scene type according to the scene type that currently requires data collection.

[0078] Among them, the terminal device cluster is used to collect the data information required for its corresponding scenario type, and each terminal device cluster corresponds to a sub-edge node cluster. At the same time, the terminal device cluster contains multiple terminal devices for data collection; and a sub-edge node contained in the sub-edge node cluster controls one or more terminal devices, for example, a sub-edge node manages and controls the data collection operation of multiple sensors.

[0079] The working principle of the above technical solution is as follows: The IoT platform has real-time monitoring capabilities and can continuously receive and interpret scenario-specific instructions sent by the user. These instructions typically include the specific scenario type for which the user wishes to collect data and related parameter settings. Based on the scenario-specific instructions sent by the user, the IoT platform can accurately determine the scenario type currently requiring data collection. These scenario types include, but are not limited to, industrial equipment monitoring, industrial environment monitoring, energy consumption monitoring, and warehouse logistics monitoring. Based on the determined scenario type, the IoT platform activates the corresponding terminal device cluster for data collection. Each terminal device cluster is optimized for a specific scenario type, enabling efficient collection of the required data information. A terminal device cluster contains multiple data collection terminal devices, such as sensors and cameras. These terminal devices, under the control of a sub-edge node cluster, collect data at a preset collection frequency and accuracy. Each terminal device cluster corresponds to a sub-edge node cluster. The sub-edge node cluster is responsible for managing and controlling its subordinate terminal devices to ensure smooth data collection. Specifically, a sub-edge node can manage and control the data collection operations of multiple terminal devices, enabling real-time data transmission and processing.

[0080] The above technical solution achieves the following: By real-time monitoring of scenario-specific requirements sent by the user end, the IoT platform can rapidly respond and activate the corresponding terminal device cluster for data collection. This flexible data collection approach can meet data collection requirements in different scenarios, improving data collection flexibility. Furthermore, because each terminal device cluster is optimized for a specific scenario type, it can efficiently collect the required data information, improving data collection accuracy. Data collected by the terminal devices is transmitted in real time to the sub-edge node cluster for processing. The sub-edge node cluster possesses powerful data processing capabilities, enabling preliminary analysis and screening of the collected data, thereby reducing the processing burden on the IoT platform. Furthermore, real-time data transmission and processing provides users with timely data feedback and decision support, improving overall system performance and user experience. By managing and controlling data collection from terminal devices through the sub-edge node cluster, this technical solution enables local processing and storage of data. This significantly reduces data transmission latency and costs, and improves system responsiveness and operational efficiency. This technical solution adopts a modular design approach, allowing each component of the system to be independently upgraded and maintained. This enhances the system's scalability and maintainability, providing users with a more convenient and efficient service experience.

[0081] In one embodiment of the present invention, real-time identification of scene demand change states by a master edge node includes:

[0082] S201a. The terminal device cluster collects data information corresponding to the scene type in real time according to the corresponding scene type, and sends the data information to the master edge node;

[0083] S202a: The master edge node analyzes the data information to determine whether there is any abnormal scenario operation event under each scenario type;

[0084] S203a: When it is determined that there is a scene operation abnormality event under the scene type, it is determined that there is a scene demand state change under the current scene type, and a data acquisition control adjustment instruction is sent to the sub-edge node cluster.

[0085] At the same time, the abnormal events corresponding to each scenario type include but are not limited to:

[0086] Abnormal events corresponding to industrial equipment monitoring scenarios: equipment operation mechanical abnormalities (for example, abnormal vibration, abnormal temperature, abnormal operating index parameters, etc.), operational abnormalities (for example, sudden drop in efficiency, sudden shutdown, etc.), and wear and aging (for example, component deformation and cracking, seal failure, etc.);

[0087] Abnormal events corresponding to industrial environment monitoring scenarios: abnormal environmental indicator parameters (for example, temperature, humidity, dust concentration, etc.) and leakage of hazardous substances;

[0088] Abnormal events corresponding to energy consumption monitoring scenarios: abnormal energy consumption, equipment no-load energy consumption, peak and off-peak electricity price mismatch, illegal electricity theft, and pipeline leakage;

[0089] Abnormal events corresponding to warehousing and logistics monitoring scenarios: abnormal inventory management, logistics equipment failure and safety risks (for example, fire hazards, intrusion by non-staff, personnel entering high-risk areas such as robot work areas), etc.

[0090] The working principle of this technical solution is as follows: a cluster of terminal devices collects relevant data in real time based on their respective scenarios (such as industrial equipment monitoring, industrial environment monitoring, energy consumption monitoring, and warehouse logistics monitoring). This data includes, but is not limited to, equipment operating status, environmental parameters, energy consumption, and inventory status. This collected data is then sent to the primary edge node for further analysis and processing.

[0091] The master edge node receives data from the terminal device cluster and performs detailed data analysis. During this analysis, the master edge node determines whether any abnormal events exist within each scenario type. These abnormal events are determined based on pre-set rules and thresholds, such as mechanical anomalies in equipment operation, abnormal environmental parameters, abnormal energy consumption, and abnormal inventory management. If the master edge node determines that an abnormal event exists within a scenario type, it deems that the scenario demand state for that scenario type has changed. When the master edge node determines that a change in scenario demand state has occurred, it sends data collection control adjustment instructions to the corresponding sub-edge node cluster based on the specific type and severity of the abnormal event. These instructions may include increasing data collection frequency, changing data collection accuracy, or initiating new data collection tasks to better adapt to the new scenario demand state. Upon receiving the instructions, the sub-edge node cluster adjusts the data collection strategy of its subordinate terminal devices to ensure accurate and timely data collection.

[0092] The above technical solution achieves the following: By collecting and analyzing data from a cluster of terminal devices in real time, the master edge node can quickly identify changes in scenario requirements. This real-time identification capability enables the system to promptly respond to changes in scenario requirements, adjust data collection strategies, and ensure data accuracy and timeliness. The master edge node possesses powerful data analysis capabilities and can accurately identify abnormal events in each scenario type. Once an abnormal event is identified, the system immediately sends data collection control adjustment instructions to the sub-edge node cluster to address the impact of the abnormal event. This abnormality handling capability helps to promptly identify and resolve potential problems, improving system stability and reliability. The system can dynamically adjust data collection strategies based on changing scenario requirements. This flexibility and adaptability enables the system to better adapt to data collection needs in different scenarios, improving data collection efficiency and accuracy. By performing data analysis and abnormality determination at the master edge node, the transmission of large amounts of raw data to the IoT platform is avoided. This reduces data transmission latency and costs, while also alleviating the processing burden on the IoT platform. Real-time collected and analyzed data provides timely and accurate data support for decision-making. This helps managers better understand the system's operating status, promptly identify and resolve potential problems, and improve the scientific and effective nature of decision-making.

[0093] In summary, this technical solution achieves real-time identification and processing of changing scenario requirements through data collection from a cluster of terminal devices, data analysis and anomaly determination by the master edge node, and the issuance of data collection control adjustment instructions. This technical solution improves the system's anomaly handling capabilities, data collection flexibility, and adaptability, reduces data transmission latency and costs, and provides timely and accurate data support for decision-making.

[0094] In one embodiment of the present invention, data collection and control of a terminal device cluster using a sub-edge node cluster according to the scenario requirement change state includes:

[0095] S201b: When each sub-edge node in the sub-edge node cluster receives a data acquisition control adjustment instruction, the sub-edge node extracts data information collected by each corresponding terminal device;

[0096] S202b. Obtaining data difference information between the data information corresponding to the time when the abnormal event occurred and the data information before the time when the abnormal event occurred based on the data information collected by each terminal device; and, when the data information collected by the terminal device is a physical parameter, the data difference information refers to the deviation amplitude ratio of the physical parameter, for example, equipment operating parameters or environmental indicator parameters such as ambient temperature data, temperature data, energy consumption data, voltage data, and current data; when the data information collected by the terminal device is deterministic data, the data difference information value is 1 and 0, 1 indicating a determination result of "yes" and 0 indicating a determination result of "no", for example, whether a non-staff member has intruded or whether a person has intruded into a high-risk area;

[0097] S203b, comparing the data difference information corresponding to each terminal device with a preset deviation reference value;

[0098] S204b, the terminal device whose data difference information exceeds the preset deviation reference value is used as the main control device;

[0099] S205b, using the terminal device whose data difference information does not exceed the preset deviation reference value as an auxiliary control device;

[0100] S206b, obtaining a data deviation coefficient using the data difference information of the primary control device and the data difference information of the auxiliary control device;

[0101] S207b, comparing the data deviation coefficient with a preset data deviation coefficient threshold;

[0102] S208b, when the data deviation coefficient exceeds a preset data deviation coefficient threshold, performing a primary regulation on the data collection frequency of the terminal device;

[0103] S209b: When the data deviation coefficient does not exceed the preset data deviation coefficient threshold, the data collection frequency of the terminal device is regulated at a secondary level.

[0104] The working principle of the above technical solution is as follows: When each sub-edge node in the sub-edge node cluster receives a data collection control adjustment instruction, it extracts the data information collected by its corresponding terminal device. Based on the data information collected by each terminal device, the sub-edge node calculates the difference between the data information at the time of the abnormal event and the data before the abnormal event. The data difference information can be the deviation ratio of a physical parameter (such as ambient temperature, temperature, energy consumption, voltage, current, etc.), or it can be a "yes" or "no" judgment data (such as whether a non-staff member has intruded or whether a person has entered a high-risk area). The sub-edge node compares the data difference information of each terminal device with a preset deviation reference value. Terminal devices whose data difference information exceeds the preset deviation reference value are classified as primary control devices. Terminal devices whose data difference information does not exceed the preset deviation reference value are classified as secondary control devices. Using the data difference information of the primary and secondary control devices, a data deviation coefficient is calculated. The calculated data deviation coefficient is compared with a preset data deviation coefficient threshold. If the data deviation coefficient exceeds the preset threshold, the terminal device's data collection frequency is adjusted to a primary level (possibly increasing the collection frequency to obtain more data). If the data deviation coefficient does not exceed the preset threshold, the data collection frequency of the terminal device is regulated at the secondary level (the collection frequency may be reduced to save resources).

[0105] The effect of the above technical solution is: by classifying terminal devices into primary control devices and auxiliary control devices, and regulating them according to data difference information, key data can be obtained more accurately and the targeted nature of data collection can be improved. According to the comparison result of the data deviation coefficient and the preset threshold, the data collection frequency of the terminal device is dynamically adjusted, which not only ensures the real-time and accuracy of the data, but also avoids unnecessary waste of resources. By calculating data difference information and data deviation coefficients in real time, the system can quickly identify abnormal events and potential risks, and respond in a timely manner. Through refined data collection and control, this technical solution helps to timely discover and solve potential problems, thereby improving the stability and reliability of the system. By performing data processing and decision-making at the sub-edge node level, the amount of data transmitted to the IoT platform is reduced, reducing the burden of data transmission and processing.

[0106] In one embodiment of the present invention, obtaining a data deviation coefficient using the data difference information of the primary control device and the data difference information of the secondary control device includes:

[0107] Step 1: extract data difference information of the main control device and the weight coefficient of the main control device;

[0108] Step 2: extracting data difference information of the auxiliary control equipment and the weight coefficient of the auxiliary control equipment;

[0109] Obtaining a first deviation factor using the data difference information of the main control device and a weight coefficient of the main control device;

[0110] The first deviation factor is obtained by the following formula:

[0111]

[0112] Among them, P 01 represents the first deviation factor; n represents the number of main control devices; S zi Indicates the value corresponding to the data difference information of the i-th main control device; S y Indicates the preset deviation reference value; w zi Represents the weight coefficient of the i-th master control device; In the above mathematical formula, the data difference information value S of each master control device is first calculated. zi Deviation from the preset reference value S y Perform ratio calculation to obtain the difference ratio of each main control device relative to the preset reference value, reflecting the degree to which the data of a single main control device deviates from the reference value. Then multiply it by the weight coefficient w of the main control device. zi The weight coefficient reflects the importance of different main control devices in the overall evaluation, highlighting the impact of important devices. Finally, the sum of all main control devices is calculated and averaged (divided by the number of main control devices n). Taking into account the situation of all main control devices, a factor that can represent the overall data deviation of the main control devices is obtained. Corresponding physical meaning: P 01 It is a quantitative indicator that reflects the deviation of the overall data of the main control equipment from the preset deviation reference value, which is obtained by comprehensively considering the number of main control equipment, the degree of difference in data of each equipment and their weight. It is used to measure the overall deviation level of the main control equipment data.

[0113] Step 3: Obtain a second deviation factor using the data difference information of the auxiliary control device and the weight coefficient of the auxiliary control device;

[0114]

[0115] Among them, P 02 represents the second deviation factor; m represents the number of auxiliary control devices; S fi Indicates the value corresponding to the data difference information of the i-th auxiliary control device; S y Indicates the preset deviation reference value; w fi Represents the weight coefficient of the i-th auxiliary control device; Specifically, similar to the first deviation factor, the data difference information value S of each auxiliary control device is first fi Deviation from the preset reference value S y By comparison, we can get the ratio of the deviation of the data of a single auxiliary control device from the reference value. Multiply it by its weight coefficient w fi, reflecting the impact of the varying importance of different auxiliary control devices on the results. The sum of all auxiliary control devices and the average (divided by the number of auxiliary control devices, m) is used to obtain the overall data deviation index for the auxiliary control devices. Corresponding physical meaning: P02 is a quantitative indicator that measures the degree of deviation of the overall auxiliary control device data from the preset deviation reference value, taking into account the number of auxiliary control devices, the data differences between each device, and their weights. It is used to characterize the overall deviation of the auxiliary control device data.

[0116] Step 4: Obtain a data deviation coefficient using the first deviation factor and the second deviation factor.

[0117] The data deviation coefficient is obtained by the following formula:

[0118]

[0119] Among them, P represents the data deviation coefficient; P 01 represents the first deviation factor; P 02 Represents the second deviation factor.

[0120] The working principle of the above technical solution is: extracting data difference information and corresponding weight coefficients from the main control device and extracting data difference information and corresponding weight coefficients from the auxiliary control device.

[0121] First deviation factor calculation:

[0122] The first deviation factor (P01) is calculated using the data difference information and weight coefficients of the master control devices through a specific formula. This formula takes into account the number of master control devices (n), the data difference information of each master control device (Szi), the preset deviation reference value (Sy), and the weight coefficient (wzi) of each master control device.

[0123] Second deviation factor calculation:

[0124] Similarly, the second deviation factor (P02) is calculated using the data difference information and weight coefficients of the auxiliary control devices through a specific formula. This formula also takes into account the number of auxiliary control devices (m), the data difference information of each auxiliary control device (Szi; note that Szi here is a common symbol for the main control and auxiliary control devices, but represents different data), the preset deviation reference value (Sy), and the weight coefficient (wfi) of each auxiliary control device.

[0125] Data deviation coefficient calculation:

[0126] Finally, the data deviation coefficient (P) is calculated using another formula using the first and second deviation factors. This formula combines the effects of the first and second deviation factors to reflect the data deviation of the entire system.

[0127] The effect of the above technical solution is: by calculating the data deviation coefficient, the data deviation of the entire system can be accurately quantified, providing an accurate basis for subsequent regulation. When calculating the deviation factor, the weight coefficient of the equipment is taken into account, which reflects the importance and influence of different equipment in the system, making the calculation results more reasonable and accurate. The control strategy based on the data deviation coefficient can adjust the data collection frequency or take other measures in a targeted manner to improve the response speed and control efficiency of the system. By real-time monitoring and calculating the data deviation coefficient, potential data deviation problems can be discovered and resolved in a timely manner, which helps to enhance the stability and reliability of the system. This technical solution is not only suitable for specific application scenarios, but can also be expanded and modified according to actual needs to adapt to different systems and equipment.

[0128] Furthermore, by calculating the first deviation factor P01 and the second deviation factor P02 in separate steps and then combining them into the data deviation coefficient P, the complex direct calculation of the combined deviations of multiple devices is avoided, reducing computational complexity and improving efficiency. The weight coefficients wzi and wfi can be pre-calculated and stored, reducing the burden of real-time calculations and further improving efficiency. The preset deviation reference value Sy can be reused for deviation calculations on all devices, avoiding duplication and improving efficiency. The technical solution supports any number of primary and secondary control devices (n and m can be dynamically adjusted), facilitating system expansion and device addition and removal. Users can customize the weight coefficients wzi and wfi based on actual needs, flexibly adjusting the contribution of different devices to deviation assessment. The preset deviation reference value Sy can be adjusted based on specific application scenarios to accommodate diverse deviation assessment requirements. By monitoring the values ​​of the first deviation factor P01 and the second deviation factor P02, abnormal data fluctuations on primary and secondary control devices can be promptly detected, providing early warning of potential problems. The data deviation coefficient P comprehensively reflects the overall deviation of the primary and secondary control devices, facilitating the rapid location of abnormal devices or data sources. The technical solution adopts a modular design, allowing each step (such as weight coefficient extraction, deviation factor calculation, and data deviation coefficient merging) to be independently optimized and expanded, facilitating system maintenance and upgrades. The multi-device joint evaluation design enables the system to adapt to complex control architectures and support potential future device expansion and functional enhancements.

[0129] In one embodiment of the present invention, the primary regulation includes:

[0130] When the data deviation coefficient exceeds a preset data deviation coefficient threshold, the sub-edge node cluster sends data information corresponding to the terminal device at the time when the abnormal event occurs to the master edge node;

[0131] The master edge node generates a scene intention vector through a built-in multimodal intention graph parsing model;

[0132] The master edge node generates a dynamic acquisition strategy matrix based on the scenario intention vector; wherein the dynamic acquisition strategy matrix includes device acquisition control parameters and sub-edge node cluster operation regulation parameters; wherein the device acquisition control parameters include but are not limited to data acquisition modes (for example, event-triggered acquisition mode, periodic polling acquisition mode, and continuous streaming acquisition mode) and terminal device acquisition control parameters (for example, acquisition frequency, resolution); the sub-edge node cluster operation regulation parameters include task allocation weights and resource constraints of sub-edge nodes; wherein the task allocation weights of the sub-edge nodes are set according to data difference information of the terminal devices corresponding to the sub-edge nodes; the resource constraints of the sub-edge nodes are set according to the task allocation weights of the sub-edge nodes;

[0133] Specifically, the task allocation weight of the sub-edge node is obtained by the following formula:

[0134]

[0135] Where R represents the task allocation weight of the child edge node; P max and P min Indicates the maximum and minimum values ​​of the data difference information of the terminal device corresponding to the sub-edge node; P y Indicates the numerical average value of the data difference information of the terminal devices corresponding to the sub-edge node; P x Indicates the numerical standard deviation of the data difference information of the terminal device corresponding to the sub-edge node; specifically, middle, This section calculates the ratio of the range (the difference between the maximum and minimum values) of the data difference information value to the average value, reflecting the degree of dispersion of the terminal device data difference information value relative to the average level, and measuring the degree of deviation of the data difference from the average state in the numerical span. Multiply the previously obtained ratio by , is to map the ratio value to [0, ] interval, in preparation for the subsequent use of the sine function, so that subsequent calculations can be performed within the appropriate function definition domain, and the characteristics of the sine function are used to further quantify the data differences. It means that based on the previous sine function calculation results, the task allocation weight is further adjusted in combination with the degree of data dispersion. The larger the standard deviation, the more dispersed the data is. (1+P x) value, the final task allocation weight R will be affected and adjusted accordingly, reflecting the impact of data discreteness on task allocation. The task allocation weight R comprehensively considers factors such as the range, mean, and dispersion (standard deviation) of the data difference information values ​​of the sub-edge node's corresponding terminal devices. Through a series of calculations, these factors are quantified into a single value, which is used to determine the task allocation weight of the sub-edge node. The greater the data variance and the higher the dispersion, the corresponding adjustment of the task allocation weight will be.

[0136] The resource constraints of the sub-edge node are obtained by the following formula:

[0137]

[0138] Where K represents the limit corresponding to the resource constraint of the sub-edge node, and when the limit corresponding to the resource constraint reaches the theoretically agreed maximum limit, it is set according to the maximum limit; K0 represents the initial limit corresponding to the resource constraint of the sub-edge node; R represents the task allocation weight of the sub-edge node; c represents the number of terminal devices corresponding to the sub-edge node; c m Indicates the preset number reference value. Specifically, This part of the calculation is the ratio of the preset reference number to the actual number of terminal devices, reflecting the situation of the actual number of terminal devices relative to the preset reference number, and measuring the relative scale of the number of terminal devices managed by the sub-edge node. This part calculates the exponential function with e as the base, and takes the negative of the previous proportional value as the exponent. Due to the characteristics of the exponential function, when c is relative to c m When c is large (i.e., the actual number of terminal devices is large), this value approaches 0; when c is small relative to cm, this value approaches 1. It is used to adjust subsequent calculations based on the relative scale of the number of terminal devices, reflecting the impact of the number of devices on resource constraints. The task allocation weight R of a child edge node is multiplied by an exponential value reflecting the relative scale of the number of terminal devices, comprehensively considering the impact of both the task allocation weight and the number of terminal devices on resource constraints. The greater the task allocation weight and the appropriate relative scale of the number of terminal devices (adjusted according to the exponential function), the more pronounced the adjustment effect of this product on resource constraints. The above formula model multiplies the adjustment amount by the initial resource constraint K0 to obtain the final resource constraint limit K. By comprehensively considering factors such as the task allocation weight and the number of terminal devices, dynamic adjustments are made to the initial resource constraints based on actual conditions to determine the resource constraints of the child edge node. When the theoretical maximum limit is reached, the maximum limit is set to ensure that resource constraints remain within a reasonable range.

[0139] The master edge node sends the device acquisition control parameter to the sub-edge node cluster;

[0140] The sub-edge node cluster performs data collection and control on its corresponding terminal device according to the device collection control parameters;

[0141] The master edge node controls the operating state of the sub-edge node cluster according to the sub-edge node cluster operating control parameters.

[0142] The working principle of the above technical solution is as follows: when the data deviation coefficient exceeds a preset threshold, the first-level control mechanism is triggered. The sub-edge node cluster sends the data information corresponding to the terminal device at the time of the abnormal event to the master edge node. The master edge node uses its built-in multimodal intent graph parsing model to parse the received data information and generate a scenario intent vector. The scenario intent vector reflects the data characteristics and changing trends of the current scenario, providing a basis for subsequent policy generation. Based on the scenario intent vector, the master edge node generates a dynamic collection policy matrix. The dynamic collection policy matrix includes device collection control parameters and sub-edge node cluster operation control parameters. Device collection control parameters include data collection mode, frequency, and resolution, which are used to adjust the data collection behavior of the terminal device. The sub-edge node cluster operation control parameters include task allocation weights and resource constraints, which are used to optimize task allocation and resource utilization of the sub-edge nodes. The task allocation weights of the sub-edge nodes are calculated based on the data difference information of the corresponding terminal device, taking into account the maximum, minimum, average, and standard deviation of the data difference information. The resource constraints of the sub-edge nodes are calculated based on the task allocation weights, the number of terminal devices, and a preset reference value to ensure the rational and effective use of resources. The master edge node sends device collection control parameters to the sub-edge node cluster. Based on the received parameters, the sub-edge node cluster regulates data collection for its corresponding terminal devices. Simultaneously, the master edge node controls the operating status of the sub-edge node cluster based on the sub-edge node cluster's operational control parameters.

[0143] The effects of the above technical solution are: by real-time monitoring of the data deviation coefficient and triggering the corresponding control mechanism, rapid response and dynamic adjustment to abnormal events are achieved. According to the scenario intention vector, a dynamic collection strategy matrix is ​​generated to optimize and customize the data collection strategy, thereby improving the efficiency and accuracy of data collection. By calculating the task allocation weights and resource constraints of the sub-edge nodes, reasonable allocation and effective utilization of resources are achieved, avoiding resource waste and bottleneck problems. By dynamically adjusting the data collection strategy and the operating status of the sub-edge node cluster, the stability and reliability of the system are enhanced, and the overall performance of the system is improved. This technical solution is not only suitable for specific application scenarios, but can also be flexibly adjusted and expanded according to actual needs, and has strong applicability and scalability.

[0144] When the data deviation coefficient exceeds a preset threshold, the sub-edge node cluster immediately sends the terminal device data at the time of the anomaly to the master edge node, enabling rapid reporting and response to the anomaly. The master edge node generates a scenario intent vector using a built-in multimodal intent graph parsing model, quickly understanding the core characteristics of the anomaly scenario and providing a precise basis for subsequent policy generation. A dynamic collection policy matrix is ​​generated based on the scenario intent vector, directly guiding the control behavior of the sub-edge node cluster, reducing manual intervention and improving response speed. The dynamic collection policy matrix supports multiple collection modes, including event triggering, periodic polling, and continuous streaming, and can be flexibly switched based on scenario requirements to improve data collection accuracy and efficiency. Device collection control parameters (such as collection frequency and resolution) can be dynamically adjusted based on scenario requirements, avoiding resource waste and improving collection efficiency. Collection policies are dynamically generated based on real-time data variance information, enabling adaptive adjustment of collection behavior to adapt to complex and changing operating environments. Task allocation weights for sub-edge nodes are dynamically calculated based on terminal device data variance information (maximum, minimum, average, and standard deviation), ensuring fair and reasonable task allocation and preventing resource overload or idleness. Resource constraints are dynamically adjusted based on task allocation weights and the number of terminal devices, ensuring efficient operation of sub-edge nodes even with limited resources. By coordinating the optimization of task allocation weights and resource constraints, resource utilization of the sub-edge node cluster is maximized, improving overall operational efficiency. Resource constraints are set to theoretical maximum limits and automatically adjust to the maximum limit when the limit is reached, preventing system crashes caused by resource overload. Resource constraints are dynamically adjusted based on initial limits, ensuring basic system operation even under resource constraints and improving system stability and reliability. The multimodal intent graph parsing model and dynamic collection policy generation algorithm, based on the dynamic collection policy matrix and resource constraints, adopt a lightweight design to reduce computational complexity and ensure real-time performance. The calculation of task allocation weights and resource constraints is performed in steps, avoiding performance bottlenecks caused by one-time calculations and improving computational efficiency. The dynamic collection policy matrix and resource constraints can be updated in real time, ensuring the system always operates in an optimal state. The technical solution adopts a modular design, allowing each step (such as data reporting, policy generation, and resource constraint adjustment) to be independently optimized and extended, facilitating system maintenance and upgrades. Preset thresholds, initial limits, reference values, and other parameters can be adjusted based on actual needs to adapt to different application scenarios. The multi-edge node collaboration mechanism enables the system to adapt to complex distributed architectures and support possible future node expansion and function enhancements.

[0145] In summary, this technical solution realizes dynamic regulation and optimization of terminal device data collection and sub-edge node cluster operation status by real-time monitoring of data deviation coefficient, generation of scenario intention vector, generation of dynamic collection strategy matrix, calculation of task allocation weights and resource constraints, and distribution and execution of strategy parameters, thereby improving the overall performance and stability of the system.

[0146] In one embodiment of the present invention, the secondary regulation includes:

[0147] When the data deviation coefficient does not exceed the preset data deviation coefficient threshold, the sub-edge node determines whether the terminal device it is responsible for controlling generates data difference information at the time when the abnormal event occurs;

[0148] When the sub-edge node identifies that there is a terminal device that generates data difference information at the time when the abnormal event occurs among the terminal devices it is responsible for controlling, the terminal device that generates data difference information at the time when the abnormal event occurs is used as the target terminal device;

[0149] Get the data collection frequency and CPU usage of the target terminal device during data collection;

[0150] Get the CPU usage of the sub-edge node corresponding to the target terminal device during the terminal device acquisition control process;

[0151] Ratio processing is performed on the CPU occupancy rate of the target terminal device and the CPU occupancy rate of the sub-edge node corresponding to the target terminal device to obtain a CPU occupancy rate ratio parameter;

[0152] The data collection frequency of the target terminal device is adjusted by utilizing the CPU occupancy ratio parameter in combination with the weight coefficient corresponding to the target terminal device.

[0153] The adjusted data collection frequency of the target terminal device is obtained by the following formula:

[0154]

[0155] Where F represents the data collection frequency of the target terminal device after adjustment; F0 represents the data collection frequency of the target terminal device before adjustment; B c Indicates the CPU usage ratio parameter; P s Indicates the deviation factor value corresponding to the target terminal device. Specifically, by dividing the data collection frequency F0 before adjustment by the adjustment coefficient Multiplying them together yields the adjusted data collection frequency F for the target terminal device. Taking into account the relative relationship between device resource usage and data differences, the original data collection frequency of the device is dynamically adjusted so that the collection frequency can be optimized based on the actual device operation and data conditions.

[0156] The working principle of the above technical solution is as follows: When the data deviation coefficient does not exceed the preset data deviation coefficient threshold, the secondary control process begins. The sub-edge node determines whether the terminal device it controls generated data discrepancy information at the time the abnormal event occurred. The terminal device with data discrepancy information is identified and designated as the target terminal device. The data collection frequency and CPU usage of the target terminal device during the data collection process are retrieved. Simultaneously, the CPU usage of the sub-edge node corresponding to the target terminal device during the terminal device collection and control process is retrieved. The CPU usage of the target terminal device is compared with the CPU usage of the sub-edge node corresponding to the target terminal device to obtain a CPU usage ratio parameter (Bc). The CPU usage ratio parameter (Bc) is combined with the target terminal device's corresponding weight coefficient (Ps, which may reflect the importance of the terminal device, but its source and calculation method are not clearly stated in the original document) to calculate the adjusted data collection frequency (F) of the target terminal device using a specific formula. The sub-edge node adjusts the data collection frequency of the target terminal device based on the calculated adjusted data collection frequency (F).

[0157] The effect of the above technical solution is: by identifying the terminal device that generates data difference information at the time of the abnormal event and adjusting its data collection frequency, it is possible to achieve fine-grained control of the data collection behavior of the terminal device. By considering the CPU occupancy ratio parameter and combining it with the weight coefficient of the terminal device to adjust the data collection frequency, it helps to optimize the use of system resources and avoid resource waste. By adjusting the data collection frequency, the problem of excessive system load caused by too frequent data collection can be reduced, thereby improving the stability and reliability of the system. This technical solution can be flexibly adjusted according to the actual conditions of different terminal devices (such as CPU occupancy, weight coefficient, etc.), and has strong adaptability and scalability. By optimizing the data collection frequency, the efficiency of data collection and processing can be improved while ensuring data quality, thereby improving the performance of the entire system.

[0158] On the other hand, when the data deviation coefficient does not exceed the preset threshold, the sub-edge node proactively determines the data discrepancy information of the terminal device at the time of the abnormal event and accurately identifies the target terminal device that requires adjustment, avoiding interference with normal devices and improving resource utilization efficiency. By accessing the CPU usage of the target terminal device and the corresponding sub-edge node and calculating the CPU usage ratio parameter, the target terminal device's data collection frequency is dynamically adjusted to ensure resource allocation matches the device's actual load and reduce resource waste. The adjusted data collection frequency is dynamically calculated based on the CPU usage ratio parameter and the target terminal device's deviation factor. This reduces the collection frequency when the device load is high and increases it when the load is low, achieving intelligent control of the collection frequency. By adjusting the collection frequency, resource overload caused by excessive collection is avoided while ensuring the timely collection of critical data, improving overall system efficiency. By monitoring and adjusting the CPU usage ratio parameter, performance degradation or system crashes caused by resource overload on the target terminal device and sub-edge node are prevented, thereby improving system stability. When the target terminal device generates data discrepancy information at the time of the abnormal event, the system can quickly adjust the collection frequency, reducing the impact of the abnormality on the system and promoting rapid system recovery. The adjusted data collection frequency calculation formula utilizes a lightweight design, requiring only simple multiplication and addition operations, reducing computational complexity and ensuring real-time performance. Sub-edge nodes can quickly identify abnormal devices and adjust the collection frequency, avoiding performance issues caused by computational delays and improving system responsiveness. The technical solution utilizes a modular design, allowing steps such as abnormal device identification, resource utilization retrieval, and frequency adjustment to be independently optimized and expanded, facilitating system maintenance and upgrades. Parameters such as the deviation factor value can be adjusted based on actual needs to accommodate different application scenarios and device characteristics. The collaborative mechanism between sub-edge nodes and terminal devices enables the system to adapt to complex distributed architectures and support potential future node expansion and functionality enhancements.

[0159] In one embodiment of the present invention, the operation strategy of the sub-edge node cluster corresponding to each scenario requirement is dynamically adjusted according to the load status of the sub-edge node, including:

[0160] S301. Monitor the load status data of each sub-edge node in real time, wherein the load status data includes CPU occupancy, memory occupancy, and network communication delay ratio between the sub-edge node and the master edge node;

[0161] S302: normalize the load status data of the edge node to obtain the normalized load status data;

[0162] S303, generating a load state vector corresponding to each unit time according to the normalized load state data corresponding to each unit time of the sub-edge node; wherein, the load state vector U=[Gc , G n , G y ], and G c , G n and G y Respectively represent the CPU usage, memory usage, and network communication delay ratio between the sub-edge node and the main edge node corresponding to each unit time;

[0163] S304. Obtain the Euclidean norm corresponding to the load state vector according to the load state vector corresponding to each unit time;

[0164] S305 : Determine whether it is necessary to adjust the task allocation weight of the sub-edge node according to the Euclidean norm of the load state vector corresponding to each unit time.

[0165] Specifically, whether the task allocation weight of the sub-edge node needs to be adjusted is determined based on the Euclidean norm of the load state vector corresponding to each unit time, including:

[0166] S3051. Obtain a difference in the Euclidean norms of the load state vectors corresponding to every two adjacent unit times using the Euclidean norm of the load state vector corresponding to the unit time;

[0167] S3052: Compare the Euclidean norm of the load state vector corresponding to the unit time with a preset norm threshold;

[0168] S3053, comparing the Euclidean norm difference with a preset difference threshold;

[0169] S3054. When the continuous duration of the Euclidean norm of the load state vector exceeding the preset norm threshold exceeds the preset time length threshold, or the number of load state vector groups corresponding to each two adjacent unit times whose Euclidean norm difference exceeds the preset difference threshold exceeds the preset group number threshold, the task allocation weight of the sub-edge node is adjusted using the Euclidean norm of the load state vector corresponding to the unit time.

[0170] The adjusted task allocation weight of the child edge node is obtained by the following formula:

[0171]

[0172] Among them, R t represents the adjusted task allocation weight of the sub-edge node; R represents the task allocation weight of the sub-edge node; U p represents the Euclidean norm average value of the load state vector of the child edge node; ΔU prepresents the average value of the Euclidean norm difference of the child edge nodes; ε represents the smoothing factor, and the value range of the smoothing factor is 0.05-0.1. The functions of the smoothing factor include preventing the denominator from being zero, adjusting the weight sensitivity and balancing resource utilization. Specifically, The numerator, ΔUp, is the average magnitude of the change in the Euclidean norm of the load state vector. The denominator, Up+ε, represents the average load size. The smoothing factor, ε, is added to prevent the denominator from reaching zero (if Up is 0) and to adjust the magnitude of the calculated result. This fraction calculates the ratio of the load change magnitude to the average load size (accounting for the smoothing factor), and is used to measure the relative impact of load change on task allocation weight adjustments. Add the result of the previous fraction to 1 to form an adjustment coefficient. 1 represents the baseline for keeping the original task allocation weight unchanged. It is the incremental part of adjusting the task allocation weight according to the load change. The adjustment coefficient is used to determine the change ratio of the final task allocation weight relative to the initial weight. Overall, by dividing the original task allocation weight R of the child edge node by the adjustment coefficient Multiply them together to get the adjusted task allocation weight R of the child edge node t Taking into account the average size of the sub-edge node load and the average magnitude of the load change, the original task allocation weights are dynamically adjusted according to the load situation, so that task allocation can better adapt to the load status changes of the sub-edge nodes.

[0173] The working principle of the above technical solution is to monitor the load status data of each sub-edge node in real time, including CPU utilization, memory utilization, and the network communication delay ratio between the sub-edge node and the main edge node. The collected load status data is normalized to eliminate the influence of different dimensions and obtain normalized load status data. Based on the normalized load status data corresponding to each unit time, a load status vector U = [Gc, Gn, Gy] is generated, where Gc, Gn, and Gy represent CPU utilization, memory utilization, and network communication delay ratio, respectively. Based on the load status vector corresponding to each unit time, its Euclidean norm is calculated to quantify the degree of load status change.

[0174] By comparing the Euclidean norm of the load state vector with a preset norm threshold, and comparing the Euclidean norm difference between each two adjacent unit times with a preset difference threshold, it is determined whether the task allocation weight of the sub-edge node needs to be adjusted. When the Euclidean norm of the load state vector exceeds the preset norm threshold for a continuous period exceeding the preset time length threshold, or the number of groups of load state vectors corresponding to each two consecutive unit times for which the Euclidean norm difference exceeds the preset difference threshold exceeds the preset group number threshold, the adjustment mechanism is triggered. The Euclidean norm of the load state vector corresponding to the unit time is used in combination with a preset formula to adjust the task allocation weight of the sub-edge node. The adjustment formula takes into account the average Euclidean norm of the load state vector, the average value of the Euclidean norm difference, and the smoothing factor to ensure the smoothness and rationality of the adjustment.

[0175] The effect of the above technical solution is: by real-time monitoring of the load status data of the sub-edge nodes and dynamically adjusting the task allocation weight according to the changes in the load status, the real-time optimization of the sub-edge node cluster operation strategy is achieved. By adjusting the task allocation weight, the resources of the sub-edge nodes can be more effectively utilized, resource overload or idleness can be avoided, load balancing can be achieved, and the overall performance of the system can be improved. The dynamic adjustment strategy helps to reduce the risk of system crash or performance degradation due to insufficient or overloaded resources, and improve the stability and reliability of the system. This technical solution can be flexibly adjusted according to the needs of different scenarios and the actual load status of the sub-edge nodes, and has strong flexibility and scalability. By introducing mathematical tools such as the Euclidean norm for quantitative analysis and judgment, intelligent management of the load status of the sub-edge nodes is achieved, and the accuracy and efficiency of management are improved.

[0176] Furthermore, by calculating the Euclidean norm (L2 norm) and its difference of the load state vector per unit time, the system can detect load changes on child edge nodes in real time. When the Euclidean norm of the load state vector exceeds a preset threshold, or when its difference exceeds the threshold, the system automatically triggers an adjustment in task allocation weights to ensure dynamic matching of resource allocation with node load. The load state vector can include multi-dimensional metrics such as CPU utilization, memory usage, and network bandwidth. By calculating the Euclidean norm, these multi-dimensional features are integrated into a single comprehensive metric, avoiding the limitations of a single metric and improving the accuracy of adjustments. By determining the duration that the Euclidean norm of the load state vector exceeds the threshold, the system can identify persistent load anomalies rather than transient fluctuations, thereby avoiding frequent adjustments due to misjudgments and improving system stability. Adjustments are triggered only when the number of consecutive sets of Euclidean norm differences exceeds the threshold, further filtering out random fluctuations and ensuring that adjustments only occur when load changes are significant, minimizing the impact of unnecessary adjustments on the system. A smoothing factor (ε) is introduced into the adjustment formula, with a value range of 0.05-0.1. This prevents the denominator from being zero and adjusts the sensitivity of weight adjustment. The smoothing factor makes weight adjustment smoother, avoids drastic weight changes caused by load fluctuations, and improves resource utilization stability. The adjusted task allocation weight (Rt) not only considers the average value (Up) of the load state vector but also incorporates the average value of the Euclidean norm difference (ΔUp). This allows weight adjustment to respond to both the absolute level and changing trend of the load, further optimizing resource allocation. The calculation of the Euclidean norm and the comparison of the difference are lightweight operations with low computational complexity and can be completed quickly, ensuring efficient system operation even in scenarios with high real-time requirements. The system can respond quickly when the load state vector or difference value continuously exceeds a threshold, adjusting task allocation weights in a timely manner to avoid performance degradation caused by latency. Parameters such as the norm threshold, difference threshold, time threshold, and group number threshold can be adjusted according to actual needs to adapt to different application scenarios and system scales. Steps such as load state vector calculation, Euclidean norm comparison, and weight adjustment can be independently optimized and extended, facilitating system maintenance and upgrades. The collaborative mechanism between sub-edge nodes and the master edge node enables the system to adapt to complex distributed architectures and support potential future node expansion and functionality enhancements.

[0177] The embodiment of the present invention proposes a multi-scenario data acquisition application system, such as Figure 2 As shown, the method for executing the multi-scenario data acquisition application method is characterized in that the multi-scenario data acquisition application system includes:

[0178] The data collection startup module is used by the IoT platform to determine the scenario requirements corresponding to the current data collection, and to start the terminal device cluster corresponding to the scenario requirements to perform data collection according to the scenario requirements;

[0179] A data acquisition and control module is used to identify the scene demand change state in real time through the main edge node, use the sub-edge node cluster to control the data acquisition of the terminal device cluster according to the scene demand change state, and feed back the processed scene data to the Internet of Things platform through the main edge node cluster;

[0180] The operation strategy dynamic adjustment module is used to dynamically adjust the operation strategy of the sub-edge node cluster corresponding to each scenario requirement according to the load status of the sub-edge node.

[0181] The working principle of the above technical solution is as follows: The IoT platform, serving as the hub of the entire data collection system, first determines the current data collection requirements based on the actual application scenario. Based on the determined requirements, the IoT platform activates the corresponding cluster of terminal devices to collect data. These terminal devices may include sensors, cameras, smart meters, etc., and they collect data according to preset parameters and instructions. During the data collection process, the master edge node is responsible for identifying changes in the scenario requirements in real time. This is achieved by analyzing data uploaded by the terminal device cluster or combining other external information. When the master edge node detects a change in scenario requirements, it uses the sub-edge node cluster to control data collection for the terminal device cluster based on the new requirements. This includes adjusting the frequency and accuracy of data collection or initiating new data collection tasks. The master edge node also performs preliminary processing on the collected data, such as data cleaning and format conversion, and then feeds the processed scenario data back to the IoT platform. To improve the efficiency and stability of the data collection system, the system dynamically adjusts its operating strategy based on the load status of the sub-edge nodes. If the load on a sub-edge node is too high, the system may transfer some of its data collection tasks to other sub-edge nodes with lower loads to achieve load balancing. At the same time, the system will also schedule and optimize resources based on the performance status of the sub-edge nodes to ensure the smooth progress of data collection tasks.

[0182] The above technical solution achieves the following: By using the IoT platform to identify scenario requirements and activate corresponding terminal device clusters for data collection, the system can flexibly adjust data collection strategies based on actual application scenarios, thereby improving data collection accuracy and practicality. Utilizing a cluster of master edge nodes and sub-edge nodes for data collection and control, the system achieves efficient and real-time data collection and processing. The master edge node is responsible for real-time identification of changing scenario requirements and making control adjustments, while the sub-edge nodes are responsible for executing specific data collection tasks and optimizing resource scheduling. Dynamically adjusting operating strategies based on the load status of the sub-edge nodes ensures efficient and stable operation. Furthermore, as application scenarios expand and the number of terminal devices increases, the system can expand data collection and processing capabilities by adding sub-edge nodes to meet a wider range of application needs. By enabling localized data processing and storage through edge computing technology, the system can significantly reduce data transmission latency and costs. This is particularly important for application scenarios requiring high-frequency, high-precision data collection.

[0183] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A multi-scenario data collection application method, characterized in that: The multi-scenario data collection application method includes: The IoT platform determines the scenario requirements corresponding to the current data collection, and starts the terminal device cluster corresponding to the scenario requirements to collect data according to the scenario requirements; The master edge node identifies the scene demand change state in real time, uses the sub-edge node cluster to control the data collection of the terminal device cluster according to the scene demand change state, and feeds back the processed scene data to the Internet of Things platform through the master edge node; Dynamically adjust the operation strategy of the sub-edge node cluster corresponding to each scenario requirement based on the load status of the sub-edge node; The data collection and control of the terminal device cluster by using the sub-edge node cluster according to the scene requirement change state includes: When each sub-edge node in the sub-edge node cluster receives the data acquisition control adjustment instruction, the sub-edge node extracts the data information collected by each corresponding terminal device; Obtaining data difference information between the data information corresponding to the time when the abnormal event occurs and the data information before the time when the abnormal event occurs based on the data information collected by each terminal device; Compare the data difference information corresponding to each terminal device with a preset deviation reference value; The terminal device whose data difference information exceeds the preset deviation reference value is used as the main control device; The terminal devices whose data difference information does not exceed the preset deviation reference value are used as auxiliary control devices; Obtaining a data deviation coefficient using the data difference information of the main control device and the data difference information of the auxiliary control device; Comparing the data deviation coefficient with a preset data deviation coefficient threshold; When the data deviation coefficient exceeds a preset data deviation coefficient threshold, the data collection frequency of the terminal device is regulated at a first level; When the data deviation coefficient does not exceed the preset data deviation coefficient threshold, the data collection frequency of the terminal device is regulated at the secondary level.

2. The multi-scenario data acquisition application method according to claim 1, characterized in that: The IoT platform determines the scenario requirements corresponding to the current data collection, and starts the terminal device cluster corresponding to the scenario requirements to collect data according to the scenario requirements, including: The IoT platform monitors the scenario demand instructions sent by the user end in real time; The IoT platform obtains the scene type currently requiring data collection based on the scene requirement instruction sent by the user; wherein the scene types include industrial equipment monitoring scene, industrial environment monitoring scene, energy consumption monitoring scene, and warehousing and logistics monitoring scene; The IoT platform starts the terminal device cluster corresponding to the scenario type according to the scenario type that currently requires data collection.

3. The multi-scenario data acquisition application method according to claim 1, characterized in that: The master edge node identifies the changing state of scenario requirements in real time, including: The terminal device cluster collects data information corresponding to the scene type in real time according to its corresponding scene type, and sends the data information to the main edge node; The master edge node performs data analysis on the data information to determine whether there is any abnormal scene operation event under each scene type; When it is determined that there is a scene operation abnormality event under the scene type, it is determined that there is a scene demand state change under the current scene type, and a data acquisition control adjustment instruction is sent to the sub-edge node cluster.

4. The multi-scenario data acquisition application method according to claim 1, characterized in that: Obtaining a data deviation coefficient using the data difference information of the primary control device and the data difference information of the auxiliary control device includes: Extracting data difference information of the main control device and the weight coefficient of the main control device; Extracting data difference information of auxiliary control equipment and weight coefficients of auxiliary control equipment; Obtaining a first deviation factor using the data difference information of the main control device and a weight coefficient of the main control device; Obtaining a second deviation factor using the data difference information of the auxiliary control device and a weight coefficient of the auxiliary control device; A data deviation coefficient is obtained using the first deviation factor and the second deviation factor.

5. The multi-scenario data acquisition application method according to claim 1, characterized in that: The first-level regulation includes: When the data deviation coefficient exceeds a preset data deviation coefficient threshold, the sub-edge node cluster sends data information corresponding to the terminal device at the time when the abnormal event occurs to the master edge node; The master edge node generates a scene intention vector through a built-in multimodal intention graph parsing model; The master edge node generates a dynamic acquisition strategy matrix according to the scenario intention vector; wherein the dynamic acquisition strategy matrix includes device acquisition control parameters and sub-edge node cluster operation regulation parameters; wherein the device acquisition control parameters include data acquisition mode and terminal device acquisition control parameters; the sub-edge node cluster operation regulation parameters include task allocation weights and resource constraints of the sub-edge nodes; wherein the task allocation weights of the sub-edge nodes are set according to data difference information of the terminal devices corresponding to the sub-edge nodes; and the resource constraints of the sub-edge nodes are set according to the task allocation weights of the sub-edge nodes; The master edge node sends the device acquisition control parameter to the sub-edge node cluster; The sub-edge node cluster performs data collection and control on its corresponding terminal device according to the device collection control parameters; The master edge node controls the operating state of the sub-edge node cluster according to the sub-edge node cluster operating control parameters.

6. The multi-scenario data acquisition application method according to claim 1, characterized in that: The secondary regulation includes: When the data deviation coefficient does not exceed the preset data deviation coefficient threshold, the sub-edge node determines whether the terminal device it is responsible for controlling generates data difference information at the time when the abnormal event occurs; When the sub-edge node identifies that there is a terminal device that generates data difference information at the time when the abnormal event occurs among the terminal devices it is responsible for controlling, the terminal device that generates data difference information at the time when the abnormal event occurs is used as the target terminal device; Get the data collection frequency and CPU usage of the target terminal device during data collection; Get the CPU usage of the sub-edge node corresponding to the target terminal device during the terminal device acquisition control process; Ratio processing is performed on the CPU occupancy rate of the target terminal device and the CPU occupancy rate of the sub-edge node corresponding to the target terminal device to obtain a CPU occupancy rate ratio parameter; The data collection frequency of the target terminal device is adjusted by utilizing the CPU occupancy ratio parameter in combination with the weight coefficient corresponding to the target terminal device.

7. The multi-scenario data acquisition application method according to claim 1, characterized in that: Dynamically adjust the operation strategy of the sub-edge node cluster corresponding to each scenario requirement based on the load status of the sub-edge node, including: Monitor the load status data of each sub-edge node in real time, wherein the load status data includes CPU occupancy, memory occupancy, and network communication delay ratio between the sub-edge node and the master edge node; Normalizing the load status data of the edge node to obtain the normalized load status data; Generate a load state vector corresponding to each unit time according to the normalized load state data corresponding to each unit time of the sub-edge node; wherein, the load state vector U=[G c , G n , G y ], and G c , G n and G y Respectively represent the CPU usage, memory usage, and network communication delay ratio between the sub-edge node and the main edge node corresponding to each unit time; Obtaining the Euclidean norm corresponding to the load state vector according to the load state vector corresponding to each unit time; The Euclidean norm of the load state vector corresponding to each unit time is used to determine whether the task allocation weight of the sub-edge node needs to be adjusted.

8. The multi-scenario data acquisition application method according to claim 7, characterized in that: The Euclidean norm of the load state vector corresponding to each unit time is used to determine whether the task allocation weight of the sub-edge node needs to be adjusted, including: Obtaining a difference in the Euclidean norms of the load state vectors corresponding to every two adjacent unit times by using the Euclidean norm of the load state vector corresponding to the unit time; comparing the Euclidean norm of the load state vector corresponding to the unit time with a preset norm threshold; comparing the Euclidean norm difference with a preset difference threshold; When the continuous duration of the Euclidean norm of the load state vector exceeding the preset norm threshold exceeds the preset time length threshold, or the number of load state vector groups corresponding to each two adjacent unit times for which the Euclidean norm difference exceeds the preset difference threshold exceeds the preset group number threshold, the task allocation weight of the sub-edge node is adjusted using the Euclidean norm of the load state vector corresponding to the unit time.

9. A multi-scenario data acquisition application system, used to execute any one of the multi-scenario data acquisition application methods according to claims 1-8, characterized in that: The multi-scenario data acquisition application system includes: The data collection startup module is used by the IoT platform to determine the scenario requirements corresponding to the current data collection, and to start the terminal device cluster corresponding to the scenario requirements to perform data collection according to the scenario requirements; A data acquisition and control module is used to identify the scene demand change state in real time through the main edge node, use the sub-edge node cluster to control the data acquisition of the terminal device cluster according to the scene demand change state, and feed back the processed scene data to the Internet of Things platform through the main edge node cluster; The operation strategy dynamic adjustment module is used to dynamically adjust the operation strategy of the sub-edge node cluster corresponding to each scenario requirement according to the load status of the sub-edge node.

Images