IoT equipment maintenance method and system for edge computing big data processing

Abstract

The present invention relates to the technical field of Internet of Things device maintenance, and discloses an Internet of Things device maintenance method and system for edge computing big data processing. The edge node is deployed through a data acquisition module to acquire multimodal data of a target Internet of Things device, thereby improving the reliability and stability of a maintenance system. The edge computing module processes the acquired multimodal data through an edge computing strategy to obtain edge data, and after processing and screening the multimodal data, only transmits valuable edge data to the cloud, thereby reducing a large amount of unnecessary data transmission and improving network resource utilization. The data transmission module transmits the edge data to the cloud. The maintenance management module determines whether a target Internet of Things device is abnormal based on the edge data to maintain the target Internet of Things device, and synchronously manages the edge computing strategy, records maintenance operations, thereby predicting the failure risk of the Internet of Things device in advance and reducing the downtime of the Internet of Things device.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things device maintenance, and in particular to an Internet of Things device maintenance method and system for edge computing big data processing. Background Art

[0002] Edge computing transfers data processing tasks from the cloud to devices at the edge of the network, processing them close to the data source, greatly reducing the delay in data transmission, which is very important for application scenarios with high real-time requirements; by performing preliminary data processing and screening at the edge node, only valuable data is sent to the cloud, reducing the load pressure on the cloud and improving processing efficiency. However, most of them have not solved how to obtain and process the multimodal data of IoT devices to obtain edge data, and after transmitting it to the cloud, judge and maintain IoT devices based on the edge data.

[0003] For example, a Chinese patent with publication number CN117395140A discloses an IoT-based equipment maintenance system and method, including an IoT platform, which is connected to the equipment in communication to obtain the equipment data of the equipment, and analyzes and processes the equipment data to generate a rule model; the equipment maintenance system includes an equipment maintenance module, an equipment repair module and an alarm notification module; the equipment maintenance module generates an equipment maintenance work order based on the rule model and the knowledge graph; the equipment repair module generates an equipment maintenance plan based on the equipment maintenance work order and the rule model based on the knowledge graph; the alarm notification module generates an alarm notification based on the rule model. The equipment is connected to the equipment maintenance system through the IoT platform, and the rule model assists the equipment maintenance system in generating a maintenance work order, which not only simplifies the data interaction method, but also allows for flexible configuration without manual operation, improves the efficiency and accuracy of equipment maintenance, and solves the problems of complex data interaction, insufficient flexibility and inability to fully utilize data in the existing IoT equipment maintenance method.

[0004] For example, the Chinese patent with the authorization announcement number CN118678255B discloses an operation and maintenance IoT maintenance method and system for intelligent optical transmission supporting equipment. The invention realizes operation and maintenance through the optical neural model, fully utilizing the processing speed of the optical neural model, thereby further improving the operation and maintenance efficiency of the IoT, thereby meeting the data transmission requirements of the intelligent optical transmission supporting equipment and ensuring reliability. The operation and maintenance of the IoT is realized through the time period operation parameters, ensuring that the fault can be accurately located and the network maintenance can be realized, thereby further meeting the data transmission requirements of the intelligent optical transmission supporting equipment and ensuring reliability.

[0005] The above patents have the problems raised by this background technology: the above-mentioned equipment maintenance system based on the Internet of Things generates equipment maintenance work orders according to the rule model based on the knowledge graph through the equipment maintenance module, generates equipment maintenance plans according to the equipment maintenance work orders and the rule model based on the knowledge graph through the equipment maintenance module, generates alarm notifications according to the rule model through the alarm notification module, and connects the equipment and the equipment maintenance system in communication with the Internet of Things platform, simplifies the data interaction method and improves the efficiency and accuracy of equipment maintenance; the above-mentioned operation and maintenance Internet of Things maintenance method realizes operation and maintenance through the optical neural model, makes full use of the processing speed of the optical neural model to improve the operation and maintenance efficiency of the Internet of Things, realizes the operation and maintenance of the Internet of Things through the time period operation parameters, and ensures that the fault can be accurately located and the network maintenance can be realized. The above two patents do not solve the problem of how to obtain and process the multimodal data of the Internet of Things device to obtain edge data, and judge and maintain the Internet of Things device according to the edge data after transmitting it to the cloud. To solve this problem, the present invention proposes an Internet of Things device maintenance method and system for edge computing big data processing. Summary of the invention

[0006] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0007] In view of the problems existing in the above-mentioned existing Internet of Things device maintenance method and system for edge computing big data processing, the present invention is proposed.

[0008] Therefore, the object of the present invention is to provide an Internet of Things device maintenance method and system for edge computing big data processing.

[0009] In order to solve the above technical problems, the present invention provides an IoT device maintenance system for edge computing big data processing, including: a data acquisition module, an edge computing module, a data transmission module and a maintenance management module;

[0010] The data acquisition module is used to deploy edge nodes to acquire multimodal data of target IoT devices. The deployment logic of the edge nodes includes determining the deployment location and number of edge nodes based on whether the target IoT devices are in a dense area and the operating status of the target IoT devices.

[0011] The edge computing module is used to process the acquired multimodal data through an edge computing strategy to obtain edge data. The edge computing strategy includes performing spatiotemporal data synchronization on the acquired multimodal data, dividing the spatiotemporal data synchronized multimodal data into modal blocks, judging whether to fuse the modal block pairs according to the attraction of the modal block pairs, and generating a modal group to obtain edge data;

[0012] The data transmission module is used to transmit edge data to the cloud;

[0013] The maintenance management module is used to determine whether the target IoT device has an abnormality based on the edge data to maintain the target IoT device, and synchronously manage the edge computing strategy and record the maintenance operations.

[0014] As a preferred solution of the IoT device maintenance system for edge computing big data processing described in the present invention, the strategy for obtaining multimodal data of the target IoT device includes:

[0015] By deploying edge nodes at the target IoT device, multimodal data of the target IoT device is obtained, and the frequency of obtaining multimodal data is dynamically adjusted according to the operating status of the target IoT device;

[0016] The logic of dynamically adjusting the acquisition frequency of multimodal data includes:

[0017] Obtain the operation evaluation results of the target IoT device , configure the evaluation threshold , compare the operation evaluation result of the target IoT device with the evaluation threshold to obtain the operation status of the target IoT device ,like ,but , that is, the operating state of the target IoT device is the state base value;

[0018] like ,but , that is, the operating state of the target IoT device is the state extreme value, and the acquisition frequency of multimodal data is dynamically adjusted according to the operating state of the target IoT device. , then the frequency of acquiring multimodal data is reduced;

[0019] like , then the frequency of acquiring multimodal data is increased.

[0020] As a preferred solution of the IoT device maintenance system for edge computing big data processing described in the present invention, the deployment logic of the edge node includes:

[0021] Get the density of target IoT devices , configure the density threshold , compare the density of the target IoT device with the density threshold to determine whether the target IoT device is in a dense area. , then the target IoT device is not in a densely populated area;

[0022] like , then the target IoT device is in a dense area;

[0023] Comprehensively check whether the target IoT device is in a densely populated area and the operating status of the target IoT device Determine the deployment location and number of edge nodes. If the target IoT devices are located in a densely populated area and , then deploy the edge node to the target IoT device and increase the number of deployed edge nodes;

[0024] If the target IoT device is in a densely populated area and , then deploy the edge nodes to the target IoT devices, keeping the number of deployed edge nodes unchanged;

[0025] If the target IoT device is not in a densely populated area and , then increase the number of deployed edge nodes;

[0026] If the target IoT device is not in a densely populated area and , the number of deployed edge nodes remains unchanged.

[0027] As a preferred solution of the IoT device maintenance system for edge computing big data processing described in the present invention, the edge computing strategy includes:

[0028] The acquired multimodal data is synchronized in time and space. The logic of the time and space data synchronization includes interpolation logic and dynamic synchronization logic. The interpolation logic includes:

[0029] Obtaining integrity of multimodal data , configure the complete threshold , the integrity of the multimodal data is compared with the complete threshold to determine whether to perform temporal and spatial interpolation on the multimodal data. If , then there is no need to perform temporal and spatial interpolation on multimodal data;

[0030] like , it is necessary to perform temporal and spatial interpolation on the multimodal data.

[0031] As a preferred solution of the IoT device maintenance system for edge computing big data processing described in the present invention, the dynamic synchronization logic includes:

[0032] The frequency of spatiotemporal data synchronization is determined according to whether the target IoT device is in a dense area. If the target IoT device is in a dense area, the frequency of spatiotemporal data synchronization is the frequency extreme value.

[0033] If the target IoT device is not in a densely populated area, the frequency of spatiotemporal data synchronization is the frequency base value.

[0034] As a preferred solution of the IoT device maintenance system for edge computing big data processing described in the present invention, the edge computing strategy also includes:

[0035] The multimodal data after spatiotemporal data synchronization is divided into modal blocks, the attraction of the modal block pair is calculated, the attraction threshold is configured, and the attraction of the modal block pair is compared with the attraction threshold to determine whether the modal block pair is fused. If the attraction of the modal block pair is less than the attraction threshold, the modal block pair is not fused.

[0036] If the attraction of the modal block pair is greater than or equal to the attraction threshold, the modal block pair is fused to generate a modal group;

[0037] After each fusion of modal block pairs, the list of modal groups is updated until the fusion of modal block pairs is completed, the final modal group is generated, and the edge data is obtained.

[0038] As a preferred solution of the IoT device maintenance system for edge computing big data processing of the present invention, the functional expression of the attraction of the modal block pair is as follows:

[0039] ;

[0040] In the formula, Represents a modal block pair and The attractiveness value, represents the time coefficient, represents a constant, represents the time decay factor, represents the time difference of the modal block pair, represents the spatial coefficient, represents the spatial attenuation factor, represents the spatial difference of the modal block pair, represents the correlation coefficient, Represents the data dependency of a modal block pair.

[0041] As a preferred solution of the IoT device maintenance system for edge computing big data processing described in the present invention, the logic for determining whether an abnormality occurs in the target IoT device includes:

[0042] Obtain the trend of edge data, configure the trend threshold, and compare the trend of edge data with the trend threshold to determine whether the target IoT device is abnormal. If the trend of edge data is less than or equal to the trend threshold, the target IoT device is not abnormal.

[0043] If the trend of the edge data is greater than the trend threshold, the target IoT device is abnormal and needs to be maintained.

[0044] The IoT device maintenance method for edge computing big data processing includes: S1, deploying edge nodes to obtain multimodal data of target IoT devices, and the deployment logic of the edge nodes includes comprehensively determining the deployment location and deployment quantity of the edge nodes based on whether the target IoT devices are in a dense area and the operating status of the target IoT devices;

[0045] S2. Processing the acquired multimodal data through an edge computing strategy to obtain edge data. The edge computing strategy includes performing spatiotemporal data synchronization on the acquired multimodal data, dividing the spatiotemporal data synchronized multimodal data into modal blocks, determining whether to fuse the modal block pairs according to the attraction of the modal block pairs, and generating a modal group to obtain edge data.

[0046] S3, transmit edge data to the cloud;

[0047] S4. Determine whether the target IoT device has any abnormality based on the edge data to maintain the target IoT device, and synchronously manage the edge computing strategy and record the maintenance operation.

[0048] A computer device includes a memory for storing instructions; and a processor for executing the instructions, so that the computer device executes an Internet of Things device maintenance method for edge computing big data processing.

[0049] A computer-readable storage medium stores a computer program, which, when executed, implements an Internet of Things device maintenance method for edge computing big data processing.

[0050] The beneficial effects of the present invention are as follows: the present invention deploys edge nodes through a data acquisition module to acquire multimodal data of a target IoT device, thereby greatly shortening the data transmission time, enabling the maintenance system to respond quickly to changes in the state of the IoT device, and improving the reliability and stability of the maintenance system; the edge computing module processes the acquired multimodal data through an edge computing strategy to obtain edge data, and after processing and screening the multimodal data, only transmits valuable edge data to the cloud, thereby reducing a large amount of unnecessary data transmission, reducing the pressure on network bandwidth, and improving network resource utilization; the data transmission module transmits the edge data to the cloud; the maintenance management module determines whether an abnormality occurs in the target IoT device based on the edge data to maintain the target IoT device, and synchronously manages the edge computing strategy, records maintenance operations, and can predict the failure risk of the IoT device in advance, and perform maintenance before the failure occurs, thereby reducing the downtime of the IoT device and improving the utilization rate of the IoT device. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0052] Figure 1 This is a system structure diagram of the Internet of Things device maintenance system for edge computing big data processing of the present invention;

[0053] Figure 2 A frequency logic diagram for dynamically adjusting and acquiring the IoT device maintenance system for edge computing big data processing of the present invention;

[0054] Figure 3 This is a flow chart of edge data generation for the IoT device maintenance system for edge computing big data processing of the present invention;

[0055] Figure 4 This is a method flow chart of the Internet of Things device maintenance method for edge computing big data processing of the present invention. DETAILED DESCRIPTION

[0056] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0057] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0058] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0059] Example 1

[0060] In this embodiment, a system structure diagram of an IoT device maintenance system for edge computing big data processing is provided, such as Figure 1 As shown, the IoT equipment maintenance system for edge computing big data processing includes a data acquisition module, an edge computing module, a data transmission module and a maintenance management module.

[0061] The following railway IoT devices represent the target IoT devices.

[0062] The data acquisition module is used to deploy edge nodes to obtain multimodal data of target IoT devices and dynamically adjust the frequency of multimodal data acquisition according to the operating status of the target IoT devices. The deployment logic of the edge nodes includes comprehensively determining the deployment location and number of edge nodes based on whether the target IoT devices are in a densely populated area and the operating status of the target IoT devices.

[0063] Strategies to obtain multimodal data from target IoT devices include:

[0064] The multimodal data of the target IoT device is obtained by deploying edge nodes at the target IoT device, and the frequency of obtaining multimodal data is dynamically adjusted according to the operating status of the target IoT device.

[0065] The logic of dynamically adjusting the acquisition frequency of multimodal data is as follows: Figure 2 As shown, specifically including:

[0066] Obtain the operation evaluation results of the target IoT device , configure the evaluation threshold , compare the operation evaluation result of the target IoT device with the evaluation threshold to obtain the operation status of the target IoT device ,like ,but , that is, the operating state of the target IoT device is the state base value;

[0067] like ,but , that is, the operating state of the target IoT device is the state extreme value, and the acquisition frequency of multimodal data is dynamically adjusted according to the operating state of the target IoT device. , then the frequency of acquiring multimodal data is reduced;

[0068] like , then the frequency of acquiring multimodal data is increased.

[0069] Exemplarily, the operation evaluation result of the railway Internet of Things device is obtained by weighted calculation of the CPU utilization rate, memory utilization rate, network bandwidth utilization rate, storage utilization rate and response time of the railway Internet of Things device. The value range of the weight coefficient of CPU utilization rate, memory utilization rate, network bandwidth utilization rate, storage utilization rate and response time is between 0 and 1, and the weight coefficients of CPU utilization rate, memory utilization rate, network bandwidth utilization rate, storage utilization rate and response time add up to 1, and the determination of the evaluation threshold depends on the actual situation; when the operation state of the railway Internet of Things device is the state base value, it means that the railway Internet of Things device is in a static or low-load state, and the frequency of obtaining multimodal data needs to be reduced to save computing resources and bandwidth; when the operation state of the railway Internet of Things device is the state extreme value, it means that the railway Internet of Things device is in an active state or a large change in the surrounding environment of the railway Internet of Things device is detected, and the frequency of obtaining multimodal data needs to be increased. By dynamically adjusting the frequency of obtaining multimodal data of the railway Internet of Things device, the efficiency and resource utilization of the maintenance system can be significantly improved.

[0070] The deployment logic of edge nodes includes:

[0071] Get the density of target IoT devices , configure the density threshold , compare the density of the target IoT device with the density threshold to determine whether the target IoT device is in a dense area. , then the target IoT device is not in a densely populated area;

[0072] like , the target IoT device is in a dense area.

[0073] Comprehensively check whether the target IoT device is in a densely populated area and the operating status of the target IoT device Determine the deployment location and number of edge nodes. If the target IoT devices are located in a densely populated area and , then deploy the edge node to the target IoT device and increase the number of deployed edge nodes;

[0074] If the target IoT device is in a densely populated area and , then deploy the edge nodes to the target IoT devices, keeping the number of deployed edge nodes unchanged;

[0075] If the target IoT device is not in a densely populated area and , then increase the number of deployed edge nodes;

[0076] If the target IoT device is not in a densely populated area and , the number of deployed edge nodes remains unchanged.

[0077] At the same time, at least one backup node is provided for each edge node to prevent service interruption due to edge node failure or network problems. The backup node can quickly take over processing tasks to ensure high availability of the maintenance system.

[0078] For example, in railway IoT devices, stations, signal equipment and track intersections (such as turnouts, tunnel exits and forks) are located in densely populated areas. Edge nodes should be deployed in these places for real-time monitoring and multimodal data processing. In particular, equipment such as turnouts and switch machines need to monitor their operating status in real time. Edge nodes can pre-process these multimodal data and issue alarms in advance when railway IoT devices fail. At the same time, the scheduling and monitoring of train operations require rapid response. Edge nodes should be deployed near the railway control center and can quickly process multimodal data from train sensors, on-board equipment and track monitoring equipment, maintain real-time decision-making capabilities, and reduce data transmission delays. At the same time, edge nodes are deployed in various important parts of the train (such as on-board sensors, wheel sensors, and carriage vibration sensors) for preliminary data processing and fault detection. The multimodal data (including temperature, pressure, and vibration) obtained by these devices is crucial for timely judgment of the operating status of the train, and can avoid bandwidth waste to improve response speed.

[0079] It should be understood that when the railway IoT devices are in a densely populated area, it is necessary to deploy edge nodes around the railway IoT devices so that they can respond in a timely manner, thereby ensuring that the edge nodes can pre-process, denoise and make preliminary fault judgments on multimodal data; at the same time, when the operating state of the railway IoT device is at the state base value, the railway IoT device is in a static or low-load state, indicating that the processing work of the edge node is not complicated, and there is no need to increase the number of deployed edge nodes; when the operating state of the railway IoT device is at the state extreme value, the railway IoT device is in an active state or a large change in the surrounding environment of the railway IoT device is detected, indicating that the edge node needs to cope with high data processing requirements or has more processing tasks, or the amount of data generated by multimodal data is large, and the number of deployed edge nodes needs to be increased; at the same time, for more complex analysis, machine learning and big data storage tasks, they need to be transmitted to the cloud, and a high-speed and stable communication protocol (such as 5G or dedicated links) should be used between the cloud and the edge nodes to ensure the smoothness of collaborative computing.

[0080] For example, redundant nodes need to be deployed at key nodes such as track safety monitoring equipment and signal control systems. If an edge node fails or loses connection, the backup node can take over the task to ensure that key functions such as signal control and equipment monitoring are not interrupted. For important nodes such as stations and train dispatching centers, at least two or more backup nodes should be deployed to ensure that when an edge node fails, other nodes can continue to provide services, and hot backup and data synchronization are used to ensure that critical data is not lost.

[0081] The edge computing module is used to process the acquired multimodal data through edge computing strategies to obtain edge data. The edge computing strategies include performing spatiotemporal data synchronization on the acquired multimodal data, dividing the multimodal data after spatiotemporal data synchronization into modal blocks, judging whether to fuse the modal block pairs according to the attraction of the modal block pairs, and generating modal groups to obtain edge data.

[0082] Edge computing strategies include:

[0083] The acquired multimodal data is synchronized in time and space. The logic of time and space data synchronization includes interpolation logic and dynamic synchronization logic. The interpolation logic includes:

[0084] Obtaining integrity of multimodal data , configure the complete threshold , the integrity of the multimodal data is compared with the complete threshold to determine whether to perform temporal and spatial interpolation on the multimodal data. If , then there is no need to perform temporal and spatial interpolation on multimodal data;

[0085] like , it is necessary to perform temporal and spatial interpolation on the multimodal data.

[0086] Exemplarily, the integrity of multimodal data is obtained by comparing the number of multimodal data actually obtained in a certain timestamp with the number of all multimodal data that should be obtained. The completeness threshold is also determined based on the actual situation. When time interpolation is required, linear interpolation is performed through the multimodal data of known timestamps before and after to estimate the multimodal data corresponding to the missing timestamp. For situations where the multimodal data changes are more complex, spline interpolation is required for correction. When spatial interpolation is required, it is necessary to combine the coordinate position of the edge node corresponding to the multimodal data (for example, (2.5, 2)), find the coordinate positions of the four edge nodes corresponding to the edge node near the edge node (for example, (1, 1), (2, 2), (3, 1) and (4, 3)), calculate the distance between the edge node and the other four edge nodes, determine the weight index (usually a positive number, such as 1 or 2), calculate the weights of the other four edge nodes, and record the values ​​of the multimodal data corresponding to the four edge nodes (10, 20, 30 and 40 respectively), and estimate the value corresponding to the edge node by the inverse distance weighted method.

[0087] Dynamic synchronization logic includes:

[0088] The frequency of spatiotemporal data synchronization is determined according to whether the target IoT device is in a dense area. If the target IoT device is in a dense area, the frequency of spatiotemporal data synchronization is the frequency extreme value.

[0089] If the target IoT device is not in a densely populated area, the frequency of spatiotemporal data synchronization is the frequency base value.

[0090] For example, in railway IoT devices, for locations that require timely response, that is, dense areas, it is necessary to regularly obtain standard time source data and the deployment locations of edge nodes for dynamic synchronization to ensure the accuracy of timestamps and spatial data synchronization. For non-dense areas, it is recommended to perform dynamic synchronization at regular intervals (such as a few hours or a day) to ensure the temporal accuracy and spatial accuracy of multimodal data, such as achieving data synchronization of timestamps and spatial locations through GPS.

[0091] It should be understood that in the case of missing timestamps and spatial positions at the same time, the maintenance system can adopt a joint interpolation method to fill in the missing data by considering both the time and space dimensions. Specifically, weighted interpolation based on time difference and space difference can be used to repair these missing multimodal data. Time and space synchronization is crucial in the maintenance system. To ensure the correctness and reliability of multimodal data, it is necessary to handle the missing and deviation of timestamps through interpolation, calibration, dynamic synchronization, etc., as well as to ensure the precise synchronization of sensors in space. By reasonably setting time and space tolerances, performing real-time corrections and dynamic adjustments, the maintenance system can effectively handle the data synchronization problems of multimodal data, thereby providing an accurate basis for subsequent maintenance management.

[0092] like Figure 3 As shown, the multimodal data after the spatiotemporal data synchronization is divided into modal blocks, the attraction of the modal block pair is calculated, the attraction threshold is configured, and the attraction of the modal block pair is compared with the attraction threshold to determine whether the modal block pair is fused. If the attraction of the modal block pair is less than the attraction threshold, the modal block pair is not fused.

[0093] If the attraction of the modal block pair is greater than or equal to the attraction threshold, the modal block pair is fused to generate a modal group;

[0094] After each fusion of modal block pairs, the list of modal groups is updated until the fusion of modal block pairs is completed, the final modal group is generated, and the edge data is obtained.

[0095] It should be understood that after calculating the attraction value of the modal block pair, it is necessary to select an attraction threshold to screen the modal block pairs with stronger attraction. Only when the attraction value of the modal block pair is greater than or equal to the attraction threshold, the modal block pair needs to be fused, ensuring that only highly correlated and highly complementary modal block pairs are fused, avoiding the addition of irrelevant or redundant information, enhancing the fusion effect, and generating modal groups. After each fusion of modal block pairs, the list of modal groups is updated, the fused modal blocks are removed, and the modal groups are added until all modal block pairs are fused. As the iteration proceeds, the modal groups are gradually fused to form larger modal groups, avoiding a comprehensive combination of all modal blocks at the beginning, optimizing the computational efficiency, and finally obtaining several fused modal groups. The modal group containing the most modal blocks is selected as the final modal group to represent the edge data. This selection can ensure that the final modal group not only contains meaningful information, but also has good representativeness, can cover the diversity and richness related to the task, improve the quality of edge data through the above fusion method, and can better capture important information in multimodal data.

[0096] Exemplarily, if there are two modal groups A1 and A2, where A1 contains 10 modal blocks and A2 contains 14 modal blocks, then A2 is selected as the final modal group to obtain edge data.

[0097] The functional expression for the attraction of a modal block pair is as follows:

[0098] ;

[0099] In the formula, Represents a modal block pair and The attractiveness value, represents the time coefficient, represents a constant, represents the time decay factor, represents the time difference of the modal block pair, represents the spatial coefficient, represents the spatial attenuation factor, represents the spatial difference of the modal block pair, represents the correlation coefficient, Represents the data dependency of a modal block pair.

[0100] It should be noted that the attraction value of the modal block pair It refers to the modal block pair and The strength or correlation of the interaction, Indicates A modal block, Indicates modal blocks, affected by time, space and data correlation, and usually a non-negative number; the time coefficient It is used to adjust the contribution of time difference to attraction, and its value range is between 0 and 1; time decay factor is used to control the decay rate of time differences, the time decay factor It is usually a positive real number (usually between 0 and 10). The larger the time decay factor, the greater the influence of time difference on attraction. Usually a non-negative real number representing a difference in time; the spatial coefficient It is used to adjust the contribution of spatial differences to attractiveness, and its value range is between 0 and 1; spatial attenuation factor It is used to control the attenuation rate of spatial differences, the spatial attenuation factor It is usually a positive real number (usually between 0 and 10). The larger the spatial attenuation factor, the greater the impact of spatial differences on attraction. Usually a non-negative real number, indicating the difference in space; the correlation coefficient It is used to adjust the contribution of data correlation to attractiveness, and its value range is between 0 and 1; data correlation of modal block pairs is the value obtained by correlation analysis, indicating the modal block and The linear correlation of the physical quantity ranges from -1 to 1.

[0101] For example, the time difference of the modal block pair , Represents a modal block timestamp, Represents a modal block The smaller the time difference of the modal block pair is, the stronger their temporal consistency is, so that the attraction of the modal block pair is greater; the spatial difference of the modal block pair is , Represents a modal block The coordinate position of Represents a modal block If the edge nodes are close in position, the smaller the spatial distance of the modal block pair is, indicating that the multimodal data has spatial consistency, and the attraction of the modal block pair is greater; the data correlation of the modal block pair The stronger the correlation between the physical quantities (such as temperature, pressure, and humidity) passing through the modal block pair, the greater the attraction between the modal block pairs.

[0102] The data transmission module is used to transmit edge data to the cloud.

[0103] Exemplarily, before edge data is transmitted, all edge data is classified (including sensitive data and non-sensitive data) and each type of data is clearly marked, which helps to strictly manage edge data during subsequent processing and transmission. Sensitive data must be encrypted for storage and transmission, and only authorized users or edge nodes can decrypt and access it.

[0104] It should be understood that sensitive data should be transmitted only in dedicated and secure network channels. At the same time, sensitive data should be isolated from non-sensitive data at the physical or virtual level to avoid transmission through shared networks or insecure paths. This will ensure the security of sensitive data in edge nodes and prevent illegal access or leakage during transmission, storage and processing, thereby improving the security and privacy protection of edge data.

[0105] The maintenance management module is used to determine whether the target IoT device has any abnormalities based on edge data in order to maintain the target IoT device, and synchronously manage edge computing strategies and record maintenance operations.

[0106] The logic for determining whether the target IoT device is abnormal includes:

[0107] Obtain the trend of edge data, configure the trend threshold, and compare the trend of edge data with the trend threshold to determine whether the target IoT device is abnormal. If the trend of edge data is less than or equal to the trend threshold, the target IoT device is not abnormal.

[0108] If the trend of the edge data is greater than the trend threshold, the target IoT device is abnormal and needs to be maintained.

[0109] For example, the trend of edge data is obtained through the rate of change of edge data. When the trend of edge data exceeds the trend threshold, it may be a signal of an abnormality in the railway Internet of Things device. For example, the temperature of the railway Internet of Things device gradually rises and does not reach the safety threshold, but the growth rate exceeds 10°C / hour, which may indicate that the railway Internet of Things device is faulty. Another example is that the current of the railway Internet of Things device drops continuously by more than a certain proportion, which may be a power failure or line problem.

[0110] For example, when an abnormality occurs in a railway IoT device, it means that the change rate of the railway IoT device is large and it is operating under high load. The number of edge nodes and computing resources are increased in a timely manner to process the multimodal data of the railway IoT device, thereby completing the acquisition of edge data.

[0111] It should be understood that after obtaining the result that an abnormality has occurred in a railway IoT device, it is necessary to verify whether the result is correct to avoid false alarms. At this time, it is necessary to compare whether multiple railway IoT devices have abnormalities at the same time to determine whether the result that an abnormality has occurred in a railway IoT device is correct.

[0112] It should be understood that the strategy for maintaining railway IoT devices includes equipment resetting and component replacement. If the abnormality of the railway IoT device is minor and may be caused by a temporary edge node error or overload, consider performing a reset operation first (for example, when the temperature or voltage is abnormal, restart the device first to see if it returns to normal); for components or equipment that have been confirmed to be damaged, the faulty components should be replaced in time to ensure the stability of the maintenance system (for example, if the battery power is seriously low, the battery needs to be replaced). All maintenance operations should be recorded to form a complete maintenance file for subsequent analysis and decision-making. Once an abnormality occurs in a railway IoT device and maintenance is performed, the relevant personnel should be notified in time for follow-up.

[0113] Example 2

[0114] In this embodiment, a method flow chart of an IoT device maintenance method for edge computing big data processing is provided, such as Figure 4 As shown in the figure, the IoT device maintenance method for edge computing big data processing includes:

[0115] S1. Deploy edge nodes to obtain multimodal data of target IoT devices. The deployment logic of edge nodes includes determining the deployment location and number of edge nodes based on whether the target IoT devices are in a densely populated area and the operating status of the target IoT devices.

[0116] S2. Processing the acquired multimodal data through an edge computing strategy to obtain edge data. The edge computing strategy includes performing spatiotemporal data synchronization on the acquired multimodal data, dividing the spatiotemporal data synchronized multimodal data into modal blocks, determining whether to fuse the modal block pairs according to the attraction of the modal block pairs, and generating a modal group to obtain edge data.

[0117] S3, transmit edge data to the cloud;

[0118] S4. Determine whether the target IoT device has any abnormality based on the edge data to maintain the target IoT device, and synchronously manage the edge computing strategy and record the maintenance operation.

[0119] For the specific principles of the IoT device maintenance method for edge computing big data processing, please refer to the principles of the IoT device maintenance system for edge computing big data processing, which will not be repeated here.

[0120] Example 3

[0121] In this embodiment, a computer device is provided, including a memory and a processor, the memory is used to store instructions, and the processor is used to execute the instructions, so that the computer device executes the steps of implementing the above-mentioned Internet of Things device maintenance method for edge computing big data processing.

[0122] Example 4

[0123] In this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed, the steps of the above-mentioned Internet of Things device maintenance method for edge computing big data processing are implemented.

[0124] The computer-readable storage medium includes: a U disk, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and other media for storing program codes.

[0125] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. IoT equipment maintenance system for edge computing big data processing, characterized by: include: Data acquisition module, edge computing module, data transmission module and maintenance management module; The data acquisition module is used to deploy edge nodes to acquire multimodal data of target IoT devices. The deployment logic of the edge nodes includes determining the deployment location and number of edge nodes based on whether the target IoT devices are in a dense area and the operating status of the target IoT devices. The edge computing module is used to process the acquired multimodal data through an edge computing strategy to obtain edge data. The edge computing strategy includes performing spatiotemporal data synchronization on the acquired multimodal data, dividing the spatiotemporal data synchronized multimodal data into modal blocks, judging whether to fuse the modal block pairs according to the attraction of the modal block pairs, and generating a modal group to obtain edge data; The edge computing strategy includes: The acquired multimodal data is synchronized in time and space. The logic of the time and space data synchronization includes interpolation logic and dynamic synchronization logic. The interpolation logic includes: Obtaining integrity of multimodal data , configure the complete threshold , the integrity of the multimodal data is compared with the complete threshold to determine whether to perform temporal and spatial interpolation on the multimodal data. If , then there is no need to perform temporal and spatial interpolation on multimodal data; like , then it is necessary to perform temporal and spatial interpolation on the multimodal data; The dynamic synchronization logic includes: The frequency of spatiotemporal data synchronization is determined according to whether the target IoT device is in a dense area. If the target IoT device is in a dense area, the frequency of spatiotemporal data synchronization is the frequency extreme value. If the target IoT device is not in a densely populated area, the frequency of spatiotemporal data synchronization is the base frequency value; The edge computing strategy also includes: The multimodal data after spatiotemporal data synchronization is divided into modal blocks, the attraction of the modal block pair is calculated, the attraction threshold is configured, and the attraction of the modal block pair is compared with the attraction threshold to determine whether the modal block pair is fused. If the attraction of the modal block pair is less than the attraction threshold, the modal block pair is not fused. If the attraction of the modal block pair is greater than or equal to the attraction threshold, the modal block pair is fused to generate a modal group; After each fusion of modal block pairs, the list of modal groups is updated until the fusion of modal block pairs is completed, the final modal group is generated, and the edge data is obtained; The functional expression of the attraction of the modal block pair is as follows: ； In the formula, Represents a modal block pair and The attractiveness value, represents the time coefficient, represents a constant, represents the time decay factor, represents the time difference of the modal block pair, represents the spatial coefficient, represents the spatial attenuation factor, represents the spatial difference of the modal block pair, represents the correlation coefficient, Represents the data dependency of modal block pairs; The data transmission module is used to transmit edge data to the cloud; The maintenance management module is used to determine whether the target IoT device has an abnormality based on the edge data to maintain the target IoT device, and synchronously manage the edge computing strategy and record the maintenance operations.

2. The IoT device maintenance system for edge computing big data processing according to claim 1, characterized in that: The strategy for acquiring multimodal data of the target IoT device includes: By deploying edge nodes at the target IoT device, multimodal data of the target IoT device is obtained, and the frequency of obtaining multimodal data is dynamically adjusted according to the operating status of the target IoT device; The logic of dynamically adjusting the acquisition frequency of multimodal data includes: Obtain the operation evaluation results of the target IoT device , configure the evaluation threshold , compare the operation evaluation result of the target IoT device with the evaluation threshold to obtain the operation status of the target IoT device ,like ,but , that is, the operating state of the target IoT device is the state base value; like ,but , that is, the operating state of the target IoT device is the state extreme value, and the acquisition frequency of multimodal data is dynamically adjusted according to the operating state of the target IoT device. , then the frequency of acquiring multimodal data is reduced; like , then the frequency of acquiring multimodal data is increased.

3. The IoT device maintenance system for edge computing big data processing according to claim 2, characterized in that: The deployment logic of the edge node includes: Get the density of target IoT devices , configure the density threshold , compare the density of the target IoT device with the density threshold to determine whether the target IoT device is in a dense area. , then the target IoT device is not in a densely populated area; like , then the target IoT device is in a dense area; Comprehensively check whether the target IoT device is in a densely populated area and the operating status of the target IoT device Determine the deployment location and number of edge nodes. If the target IoT devices are located in a densely populated area and , then deploy the edge node to the target IoT device and increase the number of deployed edge nodes; If the target IoT device is in a densely populated area and , then deploy the edge nodes to the target IoT devices, keeping the number of deployed edge nodes unchanged; If the target IoT device is not in a densely populated area and , then increase the number of deployed edge nodes; If the target IoT device is not in a densely populated area and , the number of deployed edge nodes remains unchanged.

4. The IoT device maintenance system for edge computing big data processing according to claim 3, characterized in that: The logic for determining whether the target IoT device is abnormal includes: Obtain the trend of edge data, configure the trend threshold, and compare the trend of edge data with the trend threshold to determine whether the target IoT device is abnormal. If the trend of edge data is less than or equal to the trend threshold, the target IoT device is not abnormal. If the trend of the edge data is greater than the trend threshold, the target IoT device is abnormal and needs to be maintained.

5. The IoT device maintenance method for edge computing big data processing is implemented based on the IoT device maintenance system for edge computing big data processing according to any one of claims 1 to 4, characterized in that: include: S1. Deploy edge nodes to obtain multimodal data of target IoT devices. The deployment logic of edge nodes includes determining the deployment location and number of edge nodes based on whether the target IoT devices are in a densely populated area and the operating status of the target IoT devices. S2. Processing the acquired multimodal data through an edge computing strategy to obtain edge data. The edge computing strategy includes performing spatiotemporal data synchronization on the acquired multimodal data, dividing the spatiotemporal data synchronized multimodal data into modal blocks, determining whether to fuse the modal block pairs according to the attraction of the modal block pairs, and generating a modal group to obtain edge data. S3, transmit edge data to the cloud; S4. Determine whether the target IoT device has any abnormality based on the edge data to maintain the target IoT device, and synchronously manage the edge computing strategy and record the maintenance operation.

6. A computer device, characterized in that: include: A memory for storing instructions; A processor is used to execute the instructions so that the computer device executes the Internet of Things device maintenance method for edge computing big data processing as described in claim 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed, the Internet of Things device maintenance method for edge computing big data processing as described in claim 5 is implemented.

Images