Communication Data Fusion Method and System Based on Edge Computing

Abstract

The present invention discloses a communication data fusion method and system based on edge computing. The method includes a data acquisition step, a data preprocessing step, a data fusion step, an anomaly detection step, a node feedback I step, a calculation I step, a judgment I step, a calculation II step, and a node feedback II step. The present invention relates to the technical field of edge computing. Specifically, it relates to a communication data fusion method and system based on edge computing. The technical problem to be solved by the present invention is to provide a communication data fusion method and system based on edge computing, improve the accuracy of on-site information acquisition, facilitate the accurate judgment of risks existing at nodes by staff, and improve the accuracy and reliability of risk assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of edge computing, and more specifically, to a communication data fusion method and system based on edge computing. Background Art

[0002] In the development of smart city big data, the safety supervision of urban underground spaces has received increasing attention. To prevent the occurrence of dangerous events such as leaks and collapses, it is necessary to install various sensors on underground pipelines to monitor their real-time status and ensure the normal operation of the pipelines. Since most pipelines are buried underground and not easily observable directly, manholes are usually used as nodes to install various sensors.

[0003] Currently, the data measured by various sensors in a node is usually directly uploaded to the cloud for calculation to obtain various on-site information such as temperature, humidity, and soot content. Through cloud computing technology, it is judged whether there are potential safety hazards at the node, and then the results are transmitted to the management end through the Internet and mobile networks.

[0004] For the above technical solution, since various sensors generate a large amount of data during operation, it occupies a lot of resources during the data transmission process, which easily makes some on-site information unable to be accurately obtained, resulting in the staff being unable to accurately judge the risks existing at the node. If the sensors are set to only send warning signals when the measured value exceeds the upper limit, the staff will have difficulty understanding the specific on-site information at the node, affecting the accuracy of risk assessment. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a communication data fusion method and system based on edge computing, which can improve the accuracy of obtaining on-site information, facilitate the staff to accurately judge the risks existing at the node, and improve the accuracy and reliability of risk assessment.

[0006] In a first aspect, the present invention provides a communication data fusion method based on edge computing, and the following technical solution is adopted to achieve the invention purpose:

[0007] The communication data fusion method based on edge computing includes the following steps:

[0008] Data acquisition: A communication device is set in each of the N nodes, and a near-field communication connection is established between the communication device and multiple sensors with different functions installed in the same node where the communication device is located, to obtain the measurement data of multiple sensors and the identification information of the communication device within one detection period;

[0009] Data preprocessing: Convert the measurement data of the sensors into the on-site information of the corresponding nodes, convert the identification information of the devices into the IP addresses of the corresponding nodes, and store them in the database;

[0010] Data fusion: Extract the on-site information of N nodes and group and associate them according to the preset information types stored in the database. After weighted data fusion of each group of on-site information, convert it to obtain characteristic values, and set the characteristic upper limits of each characteristic value according to historical data;

[0011] Abnormality detection: If the characteristic value of one of the nodes exceeds the characteristic upper limit, record the measurement data of this node as abnormal data;

[0012] Node feedback I: Extract the IP address and on-site information of the node corresponding to the abnormal data in the database and send them to the management end.

[0013] As a further limitation of this technical solution, a calculation I step is also set after the node feedback I step;

[0014] Calculation I: Obtain the emergency range radius and the number of emergency nodes according to historical data. Perform actual coordinate transformation on the IP address of the abnormal data node, and obtain the number of IP addresses with abnormal data within the emergency range radius of the actual coordinates in a detection period, which is recorded as the risk number;

[0015] Judgment I: If the risk number is greater than or equal to the number of emergency nodes, extract the IP address and on-site information with abnormal data and send them to the management end, otherwise record them in the database.

[0016] As a further limitation of this technical solution, a calculation II step and a node feedback II step are also set after the judgment I step;

[0017] Calculation II: Obtain the characteristic values of I nodes within the emergency range radius of the actual coordinates of the abnormal data node in a detection period. Obtain the preset sampling interval of the detection period according to historical data, and calculate the similarity score between the characteristic value of the i-th node and the characteristic value of the abnormal data node according to the detection period and the preset sampling interval ;

[0018] Node feedback II: The edge computing device extracts the IP addresses and similarity scores of I nodes and sends them to the management end.

[0019] As a further limitation of this technical solution, in the node feedback II step, the calculation model is: , where H is the total number of sampling time points, h is the number of sampling time points, T is the detection period, t is the preset sampling interval, is the characteristic value at the h-th sampling time point, is the characteristic value at the (h - 1)-th sampling time point, is the average value of the characteristic values of H sampling time points.

[0020] As a further limitation of this technical solution, in the step of obtaining data, the sampling frequencies, detection accuracies, and distribution difference degrees of measurement data of different functional sensors are also obtained;

[0021] The data fusion step is updated as follows: extracting the on-site information of N nodes and grouping and associating them according to the preset information types stored in the database, obtaining the fusion weight values through weighted summation of the sampling frequencies, detection accuracies, and distribution difference degrees of measurement data of different functional sensors, performing weighted data fusion on each group of on-site information according to the fusion weight values and then converting to obtain characteristic values, and setting the characteristic upper limits of each characteristic value according to historical data.

[0022] In a second aspect, the present invention provides a communication data fusion system based on edge computing, and the following technical solution is adopted to achieve the invention purpose:

[0023] The communication data fusion system based on edge computing includes the following modules:

[0024] Communication device: The output end is connected to the input end of the acquisition module. One communication device is provided in each of the N nodes, which is used to establish a near-field communication connection with multiple sensors with different functions installed in the same node where it is located, obtain the measurement data of multiple sensors within one detection period, and transmit it to the acquisition module;

[0025] Acquisition module: The input end is connected to the output end of the communication device, and the output end is connected to the input end of the preprocessing module, which is used to obtain the measurement data of multiple sensors within one detection period and the identification information of the communication device and transmit it to the preprocessing module;

[0026] Preprocessing module: The input end is connected to the output end of the acquisition module, and the output end is connected to the input end of the data fusion module, which is used to convert the measurement data of the sensors into the on-site information of the corresponding nodes, convert the identification information of the devices into the IP addresses of the corresponding nodes, and store them in the database;

[0027] Data fusion module: The input end is connected to the output end of the preprocessing module, and the output end is connected to the input end of the detection module, which is used to extract the on-site information of N nodes and group and associate them according to the preset information types stored in the database, perform weighted data fusion on each group of on-site information and then convert to obtain characteristic values, and set the characteristic upper limits of each characteristic value according to historical data;

[0028] Detection module: The input end is connected to the output end of the data fusion module, and the output end is connected to the input end of the feedback module I, which is used to record the measurement data of a node as abnormal data if the characteristic value of one of the nodes exceeds the characteristic upper limit;

[0029] Feedback module I: Its input end is connected to the output end of the detection module, and it is used to extract the IP addresses and on-site information of the nodes corresponding to the abnormal data in the database and send them to the management end.

[0030] As a further limitation of this technical solution, it further includes calculation module I;

[0031] Calculation module I: Its input end is connected to the output end of feedback module I, and its output end is connected to the input end of judgment module I. It is used to obtain the emergency range radius and the number of emergency nodes according to historical data, perform actual coordinate transformation on the IP addresses of the abnormal data nodes, and obtain the number of IP addresses with abnormal data within the emergency range radius of the actual coordinates in a detection period, which is recorded as the risk number;

[0032] Judgment module I: Its input end is connected to the output end of calculation module I, and it is used to extract the IP addresses and on-site information with abnormal data and send them to the management end if the risk number is greater than or equal to the number of emergency nodes, otherwise record them in the database.

[0033] As a further limitation of this technical solution, it further includes calculation module II and feedback module II;

[0034] Calculation module II: Its input end is connected to the output end of judgment module I, and its output end is connected to the input end of feedback module II. It is used to obtain the characteristic values of I nodes within the emergency range radius of the actual coordinates of the abnormal data nodes in a detection period, obtain the preset sampling interval of the detection period according to historical data, and calculate the similarity scores between the characteristic values of the i-th node and the characteristic values of the abnormal data nodes respectively according to the detection period and the preset sampling interval ;

[0035] Feedback module II: Its input end is connected to the output end of calculation module II, and it is used for the edge computing device to extract the IP addresses and similarity scores of I nodes and send them to the management end.

[0036] As a further limitation of this technical solution, in feedback module II, the calculation model is: , where H is the total number of sampling time points, h is the number of sampling time points, T is the detection period, t is the preset sampling interval, is the characteristic value of the h-th sampling time point, is the characteristic value of the (h - 1)-th sampling time point, is the average value of the characteristic values of H sampling time points.

[0037] As a further limitation of this technical solution, the communication device is also used to obtain the sampling frequencies, detection precisions and distribution difference degrees of measurement data of different functional sensors and transmit them to the acquisition module;

[0038] The acquisition module is further configured to acquire the sampling frequencies, detection accuracies, and distribution difference degrees of measurement data of different functional sensors and transmit them to the preprocessing module;

[0039] The data fusion module is updated to: the input end is connected to the output end of the preprocessing module, and the output end is connected to the input end of the detection module. It is used to extract the on-site information of N nodes and group and associate them according to the preset information types stored in the database. The fusion weight value is obtained by weighted summation through the sampling frequencies, detection accuracies, and distribution difference degrees of different functional sensors. The grouped on-site information is subjected to weighted data fusion according to the fusion weight value and then converted to obtain characteristic values. The characteristic upper limits of each characteristic value are set according to historical data.

[0040] Compared with the prior art, the advantages and positive effects of the present invention are:

[0041] 1. Install various sensors with the manhole as a node and establish a near-field communication connection. By preprocessing the measurement data, various on-site information at the node is obtained, and the corresponding IP address of the node is associated and stored with it. Data fusion conversion is performed according to the preset information types in the database to obtain characteristic values. By comparing the characteristic values with the preset characteristic upper limits, abnormal data is identified and extracted. Furthermore, the risk hazards of the pipeline are judged through the fused data. With such a setting, a large amount of data generated by the operation of various sensors is preprocessed near the data source and then transmitted and further processed, occupying less resources during the data transmission process, which can improve the accuracy of on-site information acquisition, facilitate the staff to accurately judge the risks existing at the node, and improve the accuracy and reliability of risk assessment.

[0042] 2. Since there is a certain correlation between the measurement data and on-site information between adjacent nodes, the emergency range radius and the number of emergency nodes preset according to the intelligent fire protection system in the database are obtained, and the number of risk nodes with abnormal data, that is, the number of nodes, within the emergency range radius of the actual coordinates of the abnormal data node is calculated. When the number of risks is higher than the number of emergency nodes, the IP addresses and on-site information of all abnormal data are sent, which can further perform correlation analysis on the pipeline conditions around the abnormal data node and eliminate system failures, improving the accuracy and reliability of risk assessment.

[0043] 3. Obtain the characteristic values of I nodes within the emergency range radius of the actual coordinates of the abnormal data node in one detection period, calculate the similarity scores between the abnormal data node and other nodes according to data such as the preset sampling interval, and send the results to the management terminal. With such settings, it is possible to further conduct correlation analysis on the pipeline conditions around the abnormal data node and eliminate system failures, which is conducive to the staff accurately judging the risks existing at the node and improving the accuracy and reliability of risk assessment. Brief Description of the Drawings

[0044] The drawings described herein are used to provide a further understanding of the present application and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation of the present application. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts. In the drawings:

[0045] Figure 1 is the flowchart of Embodiment 1 of the present invention;

[0046] Figure 2 is the system diagram of Embodiment 2 of the present invention. Detailed Embodiments

[0047] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the attached Figure 1 - Figure 2 drawings. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0048] It should be noted that the orientation terms such as left, right, up, down, front, and back in the embodiments of the present invention are only relative concepts to each other or are referenced based on the normal use state of the product, that is, the traveling direction of the product, and should not be considered as having limitations.

[0049] When a component is referred to as "being located" or "being provided on" another component, it can be on another component or there may be an intermediate component at the same time. When a component is referred to as "being connected to" another component, it can be directly connected to another component or there may be an intermediate component at the same time.

[0050] Embodiment 1: A communication data fusion method based on edge computing includes the following steps:

[0051] S1. Data acquisition: Install a communication device in each of the N nodes, establish a near-field communication connection between the communication device and multiple sensors with different functions installed in the same node, acquire the measurement data of the multiple sensors and the identification information of the communication device within a detection period, and obtain the sampling frequency, detection accuracy, and distribution difference degree of the measurement data of the sensors with different functions.

[0052] S2. Data preprocessing: Convert the measurement data of the sensors into the on-site information of the corresponding nodes, convert the identification information of the devices into the IP addresses of the corresponding nodes, and store them in the database.

[0053] S3. Data fusion: Extract the on-site information of the N nodes and group and associate them according to the preset information types stored in the database. Obtain the fusion weight values by performing weighted summation through the sampling frequency, detection accuracy, and distribution difference degree of the measurement data of the sensors with different functions. Convert the grouped on-site information into characteristic values after weighted data fusion according to the fusion weight values, and set the characteristic upper limits of the characteristic values according to the historical data.

[0054] S4. Anomaly detection: If the characteristic value of one of the nodes exceeds the characteristic upper limit, record the measurement data of that node as abnormal data.

[0055] S5. Node feedback I: Extract the IP addresses and on-site information of the nodes corresponding to the abnormal data in the database and send them to the management terminal.

[0056] S6. Calculation I: Obtain the emergency range radius and the number of emergency nodes according to the historical data. Perform actual coordinate transformation on the IP addresses of the abnormal data nodes, and obtain the number of IP addresses with abnormal data within the emergency range radius of the actual coordinates within a detection period, which is recorded as the risk number.

[0057] S7. Judgment I: If the risk number is greater than or equal to the number of emergency nodes, extract the IP addresses and on-site information with abnormal data and send them to the management terminal; otherwise, record them in the database.

[0058] S8. Calculation II: Obtain the characteristic values of the I nodes within the emergency range radius of the actual coordinates of the abnormal data nodes within a detection period. Obtain the preset sampling interval of the detection period according to the historical data, and calculate the similarity scores between the characteristic values of the i-th node and the characteristic values of the abnormal data nodes respectively according to the detection period and the preset sampling interval. ;

[0059] S9. Node feedback II: The edge computing device extracts the IP addresses and similarity scores of the I nodes. , and sends them to the management terminal. The calculation model of , where H is the total number of sampling time points, h is the number of sampling time points, T is the detection period, and t is the preset sampling interval. is the characteristic value at the h-th sampling time point. is the characteristic value at the (h - 1)-th sampling time point. is the average value of the characteristic values of H sampling time points.

[0060] The working principle of the communication data fusion method based on edge computing in this embodiment is as follows:

[0061] Install various sensors with the manhole as a node, establish a near-field communication connection between the sensors and the communication device in the same node, obtain the measurement data of multiple sensors and the identification information of the communication device. There is an association between the measurement data provided by multiple sensors. Thus, through preprocessing the data, various on-site information at the node can be obtained, such as index information like temperature and humidity, soot content, gas monitoring data, etc., and associate the corresponding IP address of the node with it and store it in the database that is in near-field connection with the communication device of the node. Group and associate according to the preset information types in the database, such as distance, number, area, etc. By obtaining the sampling frequency, detection accuracy, and distribution difference degree of measurement data of different functional sensors, calculate the fusion weight value after weighted summation, and perform weighted summation on each group of on-site information according to the fusion weight value, so as to convert and obtain the characteristic value, improve the accuracy of abnormal data recognition. By comparing the characteristic value with the preset characteristic upper limit, identify and extract the abnormal data, and then judge the risk hidden danger of the pipeline through the fused data.

[0062] With such a setting, a large amount of data generated by the operation of various sensors is preprocessed near the data source and then transmitted and further processed. The resources occupied during the data transmission process are less, which can improve the accuracy of on-site information acquisition, facilitate the staff to accurately judge the risks existing at the node, and improve the accuracy and reliability of risk assessment.

[0063] There is a certain association between the measurement data and on-site information between adjacent nodes. Obtain the preset emergency range radius and the number of emergency nodes according to the intelligent fire protection system in the database, calculate the number of nodes with abnormal data, that is, the number of nodes, within the emergency range radius of the actual coordinates of the abnormal data node. When the number of risks is higher than the number of emergency nodes, send the IP addresses and on-site information of all abnormal data.

[0064] Obtain the characteristic values of I nodes within the emergency range radius of the actual coordinates of the abnormal data node in one detection period, calculate the similarity scores between the abnormal data node and other nodes according to data such as the preset sampling interval, and send the results to the management terminal. With such settings, it is possible to further conduct correlation analysis on the pipeline conditions around the abnormal data node and eliminate system failures, which is conducive to the staff accurately judging the risks existing at the node and improving the accuracy and reliability of risk assessment.

[0065] Embodiment 2: A communication data fusion system based on edge computing, including the following modules:

[0066] Communication device: The output end is connected to the input end of the acquisition module. A communication device is set in each of the N nodes, which is used to establish a near-field communication connection with multiple sensors with different functions installed in the same node, obtain the measurement data of multiple sensors in one detection period, and transmit it to the acquisition module;

[0067] Acquisition module: The input end is connected to the output end of the communication device, and the output end is connected to the input end of the preprocessing module. It is used to obtain the measurement data of multiple sensors in one detection period and the identification information of the communication device, and transmit it to the preprocessing module;

[0068] Preprocessing module: The input end is connected to the output end of the acquisition module, and the output end is connected to the input end of the data fusion module. It is used to convert the measurement data of the sensor into the on-site information of the corresponding node, convert the identification information of the device into the IP address of the corresponding node, and store it in the database;

[0069] Data fusion module: The input end is connected to the output end of the preprocessing module, and the output end is connected to the input end of the detection module. It is used to extract the on-site information of N nodes, group and associate them according to the preset information types stored in the database, perform weighted data fusion on each group of on-site information and then convert it to obtain characteristic values, and set the characteristic upper limits of each characteristic value according to historical data;

[0070] Detection module: The input end is connected to the output end of the data fusion module, and the output end is connected to the input end of feedback module I. It is used to record the measurement data of a node as abnormal data if the characteristic value of one of the nodes exceeds the characteristic upper limit;

[0071] Feedback module I: The input end is connected to the output end of the detection module, and the output end is connected to the input end of calculation module I. It is used to extract the IP address and on-site information of the node corresponding to the abnormal data in the database and send them to the management terminal.

[0072] Calculation Module I: Its input end is connected to the output end of the Feedback Module I, and its output end is connected to the input end of the Judgment Module I. It is used to obtain the emergency range radius and the number of emergency nodes according to historical data, perform actual coordinate transformation on the IP addresses of abnormal data nodes, and obtain the number of IP addresses with abnormal data within the emergency range radius of the actual coordinates in a detection period, which is recorded as the number of risks.

[0073] Judgment Module I: Its input end is connected to the output end of the Calculation Module I, and its output end is connected to the input end of the Calculation Module II. It is used to extract the IP addresses and on-site information with abnormal data and send them to the management end if the number of risks is greater than or equal to the number of emergency nodes, otherwise record them in the database.

[0074] Calculation Module II: Its input end is connected to the output end of the Judgment Module I, and its output end is connected to the input end of the Feedback Module II. It is used to obtain the characteristic values of I nodes within the emergency range radius of the actual coordinates of abnormal data nodes in a detection period, obtain the preset sampling interval of the detection period according to historical data, and calculate the similarity scores between the characteristic values of the i-th node and the characteristic values of abnormal data nodes according to the detection period and the preset sampling interval ;

[0075] Feedback Module II: Its input end is connected to the output end of the Calculation Module II. It is used for the edge computing device to extract the IP addresses and similarity scores of I nodes , and send them to the management end, The calculation model is: , where H is the total number of sampling time points, h is the number of sampling time points, T is the detection period, t is the preset sampling interval, is the characteristic value at the h-th sampling time point, is the characteristic value at the (h - 1)-th sampling time point, is the average value of the characteristic values of H sampling time points.

[0076] The working principle of the communication data fusion system based on edge computing in this embodiment is:

[0077] The sensor establishes a near-field communication connection with the communication device in the same node. The acquisition module acquires the measurement data of multiple sensors and the identification information of the communication device. There is an association between the measurement data provided by the multiple sensors. The preprocessing module preprocesses the data to obtain various on-site information at the node, and associates the corresponding IP address of the node with it and stores it in the database that is in near-field connection with the communication device of the node. The data fusion module performs grouped association according to the preset information types in the database, such as distance, number, area, etc. By obtaining the sampling frequencies, detection accuracies, and distribution difference degrees of measurement data of different functional sensors, a weighted sum is performed to obtain a fusion weight value, and each grouped on-site information is weighted and summed according to the fusion weight value, so as to be converted into a characteristic value, improving the accuracy of abnormal data recognition. The detection module compares the characteristic value with the preset characteristic upper limit, thereby identifying and extracting the abnormal data. The feedback module I feeds back the information related to the abnormal data to the management end, and then judges the risk hidden danger of the pipeline through the fused data;

[0078] Due to the certain association between the measurement data and on-site information among adjacent nodes, the calculation module I acquires the preset emergency range radius and emergency node number according to the intelligent fire protection system in the database, and calculates the number of risk nodes with abnormal data, that is, the number of nodes, within the emergency range radius of the actual coordinates of the abnormal data node. When the number of risks is higher than the number of emergency nodes, the judgment module I sends the IP addresses and on-site information of all abnormal data to the management end;

[0079] The calculation module II acquires the characteristic values of I nodes within the emergency range radius of the actual coordinates of the abnormal data node in a detection cycle, and calculates the similarity scores between the abnormal data node and other nodes according to data such as the preset sampling interval. The feedback module II sends the result to the management end, which is beneficial to further perform correlation analysis on the pipeline conditions around the abnormal data node and eliminate system failures;

[0080] With such a setting, a large amount of data generated by the operation of various sensors is preprocessed near the data source and then transmitted and further processed. The resources occupied during the data transmission process are less, which can improve the accuracy of on-site information acquisition, facilitate the staff to accurately judge the risks existing at the node, and improve the accuracy and reliability of risk assessment.

[0081] The above-disclosed are only specific embodiments of the present invention. However, the present invention is not limited thereto, and any changes that can be thought of by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. A communication data fusion method based on edge computing, characterized in that: The following steps are involved: Acquiring data: A communication device is set up in each of the N nodes, and a near-field communication connection is established between the communication device and multiple sensors with different functions installed in the same node, so as to obtain the measurement data of multiple sensors in a detection cycle and the identification information of the communication device, and obtain the sampling frequency, detection accuracy and distribution difference of the measurement data of sensors with different functions; Data preprocessing: convert the sensor's measurement data into the field information of the corresponding node, convert the device's identification information into the IP address of the corresponding node, and store it in the database; Data fusion: extract the field information of N nodes and group and associate them according to the preset information types stored in the database. The fusion weight value is obtained by weighted summing the sampling frequency, detection accuracy and distribution difference of the measurement data of sensors with different functions. The field information of each group is weightedly fused according to the fusion weight value and converted to obtain the feature value. The feature upper limit of each feature value is set according to the historical data. Anomaly detection: If the characteristic value of one of the nodes exceeds the characteristic upper limit, the measurement data of the node is recorded as abnormal data; Node feedback I: Extract the IP address and site information of the node corresponding to the abnormal data in the database and send it to the management end.

2. The communication data fusion method based on edge computing according to claim 1, characterized in that: After the node feedback I step, there is also a calculation I step; Calculation I: Obtain the emergency range radius and the number of emergency nodes based on historical data, transform the IP addresses of abnormal data nodes into actual coordinates, and obtain the number of IP addresses with abnormal data within the emergency range radius of the actual coordinates in a detection cycle, which is recorded as the risk number; Judgment I: If the number of risks is greater than or equal to the number of emergency nodes, the IP address and on-site information of the abnormal data are extracted and sent to the management end, otherwise it is recorded in the database.

3. The communication data fusion method based on edge computing according to claim 2 is characterized in that: After the judgment step I, there are also a calculation step II and a node feedback step II; Calculation II: Obtain the characteristic values ​​of I nodes within the emergency range radius of the actual coordinates of the abnormal data node in a detection cycle, obtain the preset sampling interval of the detection cycle based on historical data, and calculate the similarity score between the characteristic value of the i-th node and the characteristic value of the abnormal data node based on the detection cycle and the preset sampling interval. ; Node Feedback II: The edge computing device extracts the IP address and similarity score of I nodes , and send it to the management end.

4. The communication data fusion method based on edge computing according to claim 3 is characterized in that: In the node feedback II step, The calculation model is: , where H is the total number of sampling time points, h is the number of sampling time points, T is the detection period, and t is the preset sampling interval. is the characteristic value of the hth sampling time point, is the characteristic value of the h-1th sampling time point, is the average value of the characteristic values ​​at H sampling time points.

5. The communication data fusion system based on edge computing is characterized by: Includes the following modules: Communication device: The output end is connected to the input end of the acquisition module. A communication device is set in each of the N nodes to establish a near-field communication connection with multiple sensors with different functions installed in the same node, obtain the measurement data of multiple sensors in a detection cycle and pass it to the acquisition module, obtain the sampling frequency, detection accuracy and distribution difference of the measurement data of sensors with different functions and pass it to the acquisition module; Acquisition module: the input end is connected to the output end of the communication device, and the output end is connected to the input end of the preprocessing module, which is used to acquire the measurement data of multiple sensors in a detection cycle and the identification information of the communication device and pass them to the preprocessing module, and acquire the sampling frequency, detection accuracy and distribution difference of the measurement data of sensors with different functions and pass them to the preprocessing module; Preprocessing module: The input end is connected to the output end of the acquisition module, and the output end is connected to the input end of the data fusion module, which is used to convert the measurement data of the sensor into the field information of the corresponding node, convert the identification information of the device into the IP address of the corresponding node, and store it in the database; Data fusion module: The input end is connected to the output end of the preprocessing module, and the output end is connected to the input end of the detection module. It is used to extract the field information of N nodes and group and associate them according to the preset information types stored in the database. The fusion weight value is obtained by weighted summing the sampling frequency, detection accuracy and distribution difference of the measurement data of sensors with different functions. The field information of each group is weightedly fused according to the fusion weight value and then converted to obtain the feature value. The feature upper limit of each feature value is set according to the historical data. Detection module: the input end is connected to the output end of the data fusion module, and the output end is connected to the input end of the feedback module I, and is used to record the measurement data of one node as abnormal data if the characteristic value of one node exceeds the characteristic upper limit; Feedback module I: The input end is connected to the output end of the detection module, and is used to extract the IP address and field information of the node corresponding to the abnormal data in the database and send it to the management end.

6. The communication data fusion system based on edge computing according to claim 5, characterized in that: Also includes a computing module I; Calculation module I: The input end is connected to the output end of the feedback module I, and the output end is connected to the input end of the judgment module I, and is used to obtain the emergency range radius and the number of emergency nodes according to historical data, transform the IP address of the abnormal data node into actual coordinates, and obtain the number of IP addresses with abnormal data within the emergency range radius of the actual coordinates in a detection cycle, which is recorded as the risk number; Judgment module I: The input end is connected to the output end of calculation module I, and is used to extract the IP address and on-site information of abnormal data and send it to the management end if the number of risks is greater than or equal to the number of emergency nodes, otherwise it is recorded in the database.

7. The communication data fusion system based on edge computing according to claim 6, characterized in that: Also includes a calculation module II and a feedback module II; Calculation module II: The input end is connected to the output end of judgment module I, and the output end is connected to the input end of feedback module II, which is used to obtain the characteristic values ​​of I nodes within the emergency range radius of the actual coordinates of the abnormal data node in a detection cycle, obtain the preset sampling interval of the detection cycle according to historical data, and calculate the similarity score between the characteristic value of the i-th node and the characteristic value of the abnormal data node according to the detection cycle and the preset sampling interval. ; Feedback module II: The input end is connected to the output end of the calculation module II, and is used by the edge computing device to extract the IP address and similarity score of I nodes , and send it to the management end.

8. The communication data fusion system based on edge computing according to claim 7, characterized in that: In feedback module II, The calculation model is: , where H is the total number of sampling time points, h is the number of sampling time points, T is the detection period, and t is the preset sampling interval. is the characteristic value of the hth sampling time point, is the characteristic value of the h-1th sampling time point, is the average value of the characteristic values ​​at H sampling time points.

Images