Edge computing method and cloud platform based on artificial intelligence
By combining cloud data servers and artificial intelligence in edge computing, verifying client devices and scheduling edge data servers to process data, the problem of low sensor data processing efficiency is solved, and efficient and real-time data processing and resource allocation is achieved, suitable for multiple Application scenarios.
Patent Information
- Application Number
- CN202411162925.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-08-23
AI Technical Summary
When existing edge computing methods process large-scale and heterogeneous sensor data, they have uneven allocation of computing resources, low data processing efficiency, and lack universality and scalability, making it difficult to meet real-time requirements.
Receive interactive information through cloud data servers, verify the legitimacy of client devices, and schedule edge data servers for data processing based on device feature information, realize edge computing, combine artificial intelligence to perform data preprocessing and analysis, and dynamically adjust computing resource allocation.
It significantly improves data processing efficiency, reduces data latency, optimizes computing resource allocation, enhances the scalability and universality of the system, and is suitable for multiple fields such as the Internet of Things and intelligent transportation.
Smart Images

Figure CN118689656B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data processing, and specifically to an edge computing method and a cloud platform based on artificial intelligence. Background Art
[0002] With the rapid development of IoT technology, various smart devices and sensors have been widely used in various fields, generating massive amounts of real-time data. If all this data is transmitted to the cloud for processing, it will not only increase the pressure on network bandwidth, but also cause high data latency, and fail to meet application scenarios with high real-time requirements. Therefore, edge computing, as a technology that pushes computing power and data storage to the edge of the network, has gradually become a key solution to this problem.
[0003] Edge computing can significantly reduce data transmission latency, improve data processing efficiency, and reduce reliance on the cloud by migrating computing tasks and data storage to the edge of the network closer to the data source. However, traditional edge computing methods often face problems such as uneven distribution of computing resources and low data processing efficiency when processing large-scale, heterogeneous sensor data. In addition, with the continuous advancement of artificial intelligence technology, combining artificial intelligence with edge computing to achieve intelligent data processing and analysis has become a hot topic in current research.
[0004] In the existing technology, although some methods combining edge computing with artificial intelligence have been proposed, most of these methods focus on single-scenario applications and lack universality and scalability. At the same time, these methods often find it difficult to adaptively adjust computing resource allocation when processing large-scale, dynamically changing sensor data, resulting in low data processing efficiency. Summary of the invention
[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and provide an edge computing method based on artificial intelligence, comprising the following steps:
[0006] Step 1: The cloud data server receives the interaction information sent by the client, obtains the first type of interaction characteristics according to the interaction information, and obtains the response status of the cloud data server according to the first type of interaction characteristics. If it is a response status, it proceeds to step 2; if it is a non-response status, it discards the received interaction information;
[0007] Step 2: Obtain the client type in the interaction information. If it is a Class I client, proceed to Step 3; if it is not a Class I client, proceed to Step 4.
[0008] Step 3: The cloud data server obtains the device legitimacy information associated with the first type of interactive feature, and verifies the device feature information in the device legitimacy information with the corresponding device feature information in the cloud data server. If the verification is successful, the process proceeds to step 5. If the verification fails, the device is determined to be an illegal device, and the cloud data server issues an alarm message.
[0009] Step 4: The cloud data server obtains the device legitimacy verification information associated with the first type of interactive feature, and verifies the obtained device legitimacy verification information with the device legitimacy verification information stored in the cloud data server. If the verification is successful, proceed to step 5;
[0010] Step 5: The cloud data server generates client connection information and sends it to the edge data server scheduling module. The edge data server scheduling module schedules the first edge server to connect with the corresponding client based on the client connection information sent by the cloud data server. The first edge data server executes the response to complete the edge computing data processing.
[0011] Furthermore, the cloud data server receives the interaction information, obtains the corresponding first type of interaction feature according to the interaction information, and obtains the response state of the cloud data server according to the corresponding first type of interaction feature, including:
[0012] The first interactive feature described is a request for edge data server computing power allocation; the edge data server is a server that provides edge computing support to the client; if there is an allocatable edge data server, the cloud data server responds, otherwise, the cloud data server does not respond.
[0013] Furthermore, the obtaining of the client type in the interaction information includes:
[0014] A client that matches the client database in the cloud data server is a first-class client; otherwise, it is a non-first-class client.
[0015] Furthermore, the cloud data server obtains the device legitimacy information associated with the first type of interactive feature, and verifies the feature information in the device legitimacy information with the corresponding feature information in the cloud data server, including:
[0016] When a class of clients sends interactive information, device legitimacy information is generated based on the historical trajectory data stored in the class of clients and sent to the cloud data server. The cloud data server compares the historical trajectory data included in the device legitimacy information with the historical trajectory data of the corresponding class of clients stored in the cloud data server. If they are consistent, the verification is successful; otherwise, the verification fails.
[0017] Furthermore, the cloud data server obtains the device legitimacy verification information associated with the first type of interactive feature, and verifies the obtained device legitimacy verification information with the verification information stored in the cloud data server, including:
[0018] The device legitimacy verification information is the identity information and identity authentication information of the non-class I client. If the identity authentication information is consistent with the identity authentication information stored in the cloud, the identity information of the non-class I client is verified successfully.
[0019] Furthermore, the cloud data server generates client connection information and sends it to the edge data server scheduling module, including:
[0020] If it is a Class I client, the cloud data server generates client connection information based on the location data of the Class I client and sends it to the edge server scheduling module;
[0021] If it is a non-class I client, the cloud data server generates client connection information based on the positioning data and client identity information of the non-class I client, and sends it to the edge server scheduling module.
[0022] Further, the edge data server scheduling module schedules the first edge server to connect with the corresponding client according to the client connection information sent by the received cloud data server, and the first edge data server executes the response to complete the edge computing data processing, including:
[0023] The edge data server scheduling module obtains the client location information of the client connection information, matches the obtained client location information with each edge data server in the edge data server list, and the edge data server with the smallest location distance from the client location information is the first edge data server to be matched. According to the data transmission delay between the remaining edge data servers and the first edge data server, an edge computing sequence is generated, and the client is connected to the first edge data server;
[0024] The first edge data server obtains the data segmentation feature, segments the to-be-processed data transmitted by the client according to the data segmentation feature, obtains multiple data blocks, and obtains a data block sequence; the first edge data server sends the preset data preprocessing model call information and the data blocks in the data block sequence to each edge data server in the edge computing sequence according to the edge computing sequence and the preset data preprocessing model call information;
[0025] The first edge data server receives the preprocessed data blocks returned by each edge data server, and splices them according to the data block sequence to obtain the preprocessed data of the data to be processed, thereby completing the data edge computing.
[0026] An edge computing cloud platform based on artificial intelligence, applying the edge computing method based on artificial intelligence, including: a cloud data server, an edge data server scheduling module, an edge data server, a communication module and a client;
[0027] The cloud data server and edge data server scheduling module are respectively connected to the communication module; the edge data server is connected to the edge data server scheduling module; and the client is communicatively connected to the communication module.
[0028] Preferably, the client includes a class I client, a non-class I client and a client management module; the class I client and the non-class I client are respectively connected to the client management module; and the client management module is connected to the communication module.
[0029] The beneficial effects of the present invention are:
[0030] Significantly improve data processing efficiency: By combining artificial intelligence and edge computing, the present invention can pre-process and analyze data at the edge of the network, reducing the time for data transmission to the cloud, thereby significantly improving data processing efficiency and meeting application scenarios with high real-time requirements.
[0031] Reduce data latency: Since computing tasks and data storage are migrated to the edge of the network closer to the data source, the present invention can reduce data transmission latency, make data processing faster, and improve the response speed of the system.
[0032] Optimize computing resource allocation: The present invention realizes the reasonable allocation and efficient utilization of computing resources by intelligently starting the edge computing terminal and dynamically adjusting the grouping window size according to the feedback data, thus avoiding the waste of computing resources and bottleneck problems.
[0033] Enhance the scalability and universality of the system: The method proposed in the present invention is not only suitable for single-scenario applications, but can also process large-scale, heterogeneous sensor data. It has strong scalability and universality and can be widely used in multiple fields such as the Internet of Things, smart transportation, and smart security. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 A flowchart of an edge computing method based on artificial intelligence;
[0035] Figure 2 This is a schematic diagram of the principle of an edge computing cloud platform based on artificial intelligence;
[0036] Figure 3 This is a client connection diagram. DETAILED DESCRIPTION
[0037] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings, but the protection scope of the present invention is not limited to the following.
[0038] The features and performance of the present invention are further described in detail below in conjunction with the embodiments.
[0039] like Figure 1 As shown, the edge computing method based on artificial intelligence includes the following steps:
[0040] Step 1: The cloud data server receives the interaction information sent by the client, obtains the first type of interaction characteristics according to the interaction information, and obtains the response status of the cloud data server according to the first type of interaction characteristics. If it is a response status, it proceeds to step 2; if it is a non-response status, it discards the received interaction information;
[0041] Step 2: Obtain the client type in the interaction information. If it is a Class I client, proceed to Step 3; if it is not a Class I client, proceed to Step 4.
[0042] Step 3: The cloud data server obtains the device legitimacy information associated with the first type of interactive feature, and verifies the device feature information in the device legitimacy information with the corresponding device feature information in the cloud data server. If the verification is successful, the process proceeds to step 5. If the verification fails, the device is determined to be an illegal device, and the cloud data server issues an alarm message.
[0043] Step 4: The cloud data server obtains the device legitimacy verification information associated with the first type of interactive feature, and verifies the obtained device legitimacy verification information with the device legitimacy verification information stored in the cloud data server. If the verification is successful, proceed to step 5;
[0044] Step 5: The cloud data server generates client connection information and sends it to the edge data server scheduling module. The edge data server scheduling module schedules the first edge server to connect with the corresponding client based on the client connection information sent by the cloud data server. The first edge data server executes the response to complete the edge computing data processing.
[0045] The cloud data server receives the interaction information, obtains the corresponding first type of interaction characteristics according to the interaction information, and obtains the response state of the cloud data server according to the corresponding first type of interaction characteristics, including:
[0046] The first interactive feature described is a request for edge data server computing power allocation; the edge data server is a server that provides edge computing support to the client; if there is an allocatable edge data server, the cloud data server responds, otherwise, the cloud data server does not respond.
[0047] The client type in the obtained interaction information includes:
[0048] A client that matches the client database in the cloud data server is a first-class client; otherwise, it is a non-first-class client.
[0049] The cloud data server obtains the device legitimacy information associated with the first type of interactive feature, and verifies the feature information in the device legitimacy information with the corresponding feature information in the cloud data server, including:
[0050] When a class of clients sends interactive information, device legitimacy information is generated based on the historical trajectory data stored in the class of clients and sent to the cloud data server. The cloud data server compares the historical trajectory data included in the device legitimacy information with the historical trajectory data of the corresponding class of clients stored in the cloud data server. If they are consistent, the verification is successful; otherwise, the verification fails.
[0051] The cloud data server obtains the device legitimacy verification information associated with the first type of interactive feature, and verifies the obtained device legitimacy verification information with the verification information stored in the cloud data server, including:
[0052] The device legitimacy verification information is the identity information and identity authentication information of the non-class I client. If the identity authentication information is consistent with the identity authentication information stored in the cloud, the identity information of the non-class I client is verified successfully.
[0053] The cloud data server generates client connection information and sends it to the edge data server scheduling module, including:
[0054] If it is a Class I client, the cloud data server generates client connection information based on the location data of the Class I client and sends it to the edge server scheduling module;
[0055] If it is a non-class I client, the cloud data server generates client connection information based on the positioning data and client identity information of the non-class I client, and sends it to the edge server scheduling module.
[0056] The edge data server scheduling module schedules the first edge server to connect with the corresponding client according to the client connection information sent by the cloud data server, and the first edge data server executes the response to complete the edge computing data processing, including:
[0057] The edge data server scheduling module obtains the client location information of the client connection information, matches the obtained client location information with each edge data server in the edge data server list, and the edge data server with the smallest location distance from the client location information is the first edge data server to be matched. According to the data transmission delay between the remaining edge data servers and the first edge data server, an edge computing sequence is generated, and the client is connected to the first edge data server;
[0058] The first edge data server obtains the data segmentation feature, segments the to-be-processed data transmitted by the client according to the data segmentation feature, obtains multiple data blocks, and obtains a data block sequence; the first edge data server sends the preset data preprocessing model call information and the data blocks in the data block sequence to each edge data server in the edge computing sequence according to the edge computing sequence and the preset data preprocessing model call information;
[0059] The first edge data server receives the preprocessed data blocks returned by each edge data server, and splices them according to the data block sequence to obtain the preprocessed data of the data to be processed, thereby completing the data edge computing.
[0060] like Figure 2 As shown, an edge computing cloud platform based on artificial intelligence, applying the edge computing method based on artificial intelligence, includes: a cloud data server, an edge data server scheduling module, an edge data server, a communication module and a client;
[0061] The cloud data server and edge data server scheduling module are respectively connected to the communication module; the edge data server is connected to the edge data server scheduling module; and the client is communicatively connected to the communication module.
[0062] like Figure 3 As shown, the client includes a class I client, a non-class I client and a client management module; the class I client and the non-class I client are respectively connected to the client management module; the client management module is connected to the communication module.
[0063] Specifically, the specific steps and implementation methods of this method will be explained in detail.
[0064] Step 1: Interactive information processing of cloud data server
[0065] As the hub of the entire system, the cloud data server first receives the interactive information sent by the client. This interactive information may include client requests, data or status updates, etc. Based on this information, the server extracts the first type of interactive features, which are usually related to the computing power allocation request of the edge data server. If there are allocatable edge data server resources, the cloud data server will enter the response state and prepare for subsequent operations; if there are insufficient resources or it cannot respond for other reasons, the received interactive information will be discarded and no operation will be performed.
[0066] Step 2: Client type identification
[0067] After confirming that a response is required, the cloud data server further analyzes the interaction information to identify the type of client. This step is achieved by matching the client information with the client database stored in the cloud. If a match is found, the client is identified as a Class I client; otherwise, it is considered a non-Class I client.
[0068] Step 3: Device legitimacy verification for a class of clients
[0069] For a type of client, the cloud data server will obtain the device legitimacy information associated with its interactive features. This type of information usually includes the client's historical trajectory data, etc. The server compares this information with the corresponding data stored in itself to verify the legitimacy of the device. If the information is consistent, the verification is passed and the next step is entered; if it is inconsistent, it is determined to be an illegal device and an alarm message is issued.
[0070] Step 4: Device legitimacy verification for non-class 1 clients
[0071] For non-class I clients, since they are not registered in the cloud data server, they cannot provide historical trajectory data for verification. At this time, the cloud data server will obtain the device legitimacy verification information associated with its interactive features, which usually includes the client's identity information and identity authentication information. The server compares this information with the information stored in itself to verify the identity of the client. If the identity authentication information is consistent, the verification is passed and proceed to the next step.
[0072] Step 5: Scheduling and connecting edge data servers
[0073] After the device legitimacy verification is passed, the cloud data server will generate client connection information and send it to the edge data server scheduling module. For a class I client, the connection information includes its location data; for a non-class I client, it also includes its identity information. Based on this information, the scheduling module selects the edge data server closest to the client as the first edge data server and generates an edge computing sequence. The client then establishes a connection with the first edge data server and starts data transmission and processing.
[0074] Once the client establishes a connection with the first edge data server, the server will obtain the data segmentation features and segment the data to be processed transmitted by the client according to these features to obtain multiple data blocks and form a data block sequence. The first edge data server will call information based on the edge computing sequence and the preset data preprocessing model, and send this information and data blocks to each edge data server in the sequence for preprocessing. After the processing is completed, each server will return the preprocessed data blocks to the first edge data server, which will splice them according to the data block sequence to obtain the preprocessed complete data, thereby completing the edge computing processing of the data.
[0075] In order to realize the above-mentioned edge computing method based on artificial intelligence, a cloud platform architecture including a cloud data server, an edge data server scheduling module, an edge data server, a communication module and a client is constructed. In this architecture, the cloud data server is responsible for receiving and processing the client's interactive information, verifying the legitimacy of the device, and generating client connection information. The edge data server scheduling module is responsible for scheduling the appropriate edge data server to connect with the client according to the instructions of the cloud data server. The edge data server is responsible for specific data processing tasks. The communication module is responsible for data transmission and communication between various components in the entire system. The client includes a class I client, a non-class I client and a client management module, and interacts with the entire system through the communication module.
[0076] The following are two specific examples of edge computing methods based on artificial intelligence:
[0077] Example 1: Intelligent security monitoring system
[0078] In an intelligent security monitoring system, it is necessary to monitor the video data sent back by multiple cameras in real time and perform anomaly detection, such as detecting intruders, fires, etc. Due to the large number of cameras and the need for real-time processing, an edge computing method based on artificial intelligence is adopted.
[0079] Implementation steps:
[0080] Step 1: The cloud data server receives interaction information from each camera, which includes the camera's location, device ID, and instructions for requesting video data processing.
[0081] Step 2: The cloud data server identifies the type of camera, determines it as a Class I client based on historical records, and obtains its device legitimacy information.
[0082] Step 3: The cloud data server verifies the legitimacy of the camera device and confirms that the camera is a legitimate device by comparing historical trajectory data and device feature information.
[0083] Step 4: The cloud data server generates client connection information, including the camera’s location data and device ID, and sends it to the edge data server scheduling module.
[0084] Step 5: The edge data server scheduling module selects the edge data server closest to the camera as the first edge data server and generates an edge computing sequence.
[0085] Data processing: The first edge data server receives the video data transmitted by the camera, divides the video data into multiple data blocks according to the data segmentation features, and uses the preset video analysis model to pre-process the data, such as frame difference, motion detection, etc. After the processing is completed, the results are sent back to the cloud data server for further analysis and storage.
[0086] By adopting an AI-based edge computing method, the intelligent security monitoring system can process video data sent back by a large number of cameras in real time and perform anomaly detection, greatly improving the efficiency and accuracy of monitoring.
[0087] Example 2: Intelligent traffic management system
[0088] In an intelligent traffic management system, traffic flow data from various intersections needs to be collected and processed in real time for traffic signal control, congestion prediction, etc. Since the amount of traffic flow data is large and needs to be processed in real time, an edge computing method based on artificial intelligence is adopted.
[0089] Implementation steps:
[0090] Step 1: The cloud data server receives interaction information from traffic flow detectors at each intersection. The interaction information includes the detector's location, device ID, and instructions for requesting processing of traffic flow data.
[0091] Step 2: The cloud data server identifies the type of traffic flow detector, determines it as a Class I client based on historical records, and obtains its device legitimacy information.
[0092] Step 3: The cloud data server verifies the legitimacy of the traffic flow detector and confirms that the detector is a legitimate device by comparing historical trajectory data and device feature information.
[0093] Step 4: The cloud data server generates client connection information, including the location data and device ID of the traffic flow detector, and sends it to the edge data server scheduling module.
[0094] Step 5: The edge data server scheduling module selects the edge data server closest to the intersection as the first edge data server and generates an edge computing sequence.
[0095] Data processing: The first edge data server receives the data transmitted by the traffic flow detector, divides the data into multiple data blocks according to the data segmentation characteristics, and uses the preset traffic flow analysis model to pre-process the data, such as traffic flow statistics, congestion analysis, etc. After the processing is completed, the results are sent back to the cloud data server for further analysis and storage.
[0096] By adopting AI-based edge computing methods, the intelligent traffic management system can collect and process traffic flow data from various intersections in real time, and perform traffic signal control and congestion prediction, greatly improving the efficiency and accuracy of traffic management and reducing the occurrence of traffic congestion and traffic accidents.
Claims
1. The edge computing method based on artificial intelligence is characterized by: The steps include: Step 1: The cloud data server receives the interaction information sent by the client, obtains the first type of interaction characteristics according to the interaction information, and obtains the response status of the cloud data server according to the first type of interaction characteristics. If it is a response status, it proceeds to step 2; if it is a non-response status, it discards the received interaction information; Step 2: Obtain the client type in the interaction information. If it is a Class I client, proceed to Step 3; if it is not a Class I client, proceed to Step 4. Step 3: The cloud data server obtains the device legitimacy information associated with the first type of interactive feature, and verifies the device feature information in the device legitimacy information with the corresponding device feature information in the cloud data server. If the verification is successful, the process proceeds to step 5. If the verification fails, the device is determined to be an illegal device, and the cloud data server issues an alarm message. Step 4: The cloud data server obtains the device legitimacy verification information associated with the first type of interactive feature, and verifies the obtained device legitimacy verification information with the device legitimacy verification information stored in the cloud data server. If the verification is successful, proceed to step 5; Step 5: The cloud data server generates client connection information and sends it to the edge data server scheduling module. The edge data server scheduling module schedules the first edge server to connect with the corresponding client according to the client connection information sent by the cloud data server. The first edge data server executes the response to complete the edge computing data processing. The cloud data server obtains the device legitimacy information associated with the first type of interactive feature, and verifies the feature information in the device legitimacy information with the corresponding feature information in the cloud data server, including: When a class of clients sends interactive information, the device legitimacy information is generated based on the historical trajectory data stored in the class of clients and sent to the cloud data server. The cloud data server compares the historical trajectory data included in the device legitimacy information with the historical trajectory data of the corresponding class of clients stored in the cloud data server. If they are consistent, the verification is passed, otherwise, the verification fails; The cloud data server obtains the device legitimacy verification information associated with the first type of interactive feature, and verifies the obtained device legitimacy verification information with the verification information stored in the cloud data server, including: The device legitimacy verification information is the identity information and identity authentication information of the non-class I client. If the identity authentication information is consistent with the identity authentication information stored in the cloud, the identity information verification of the non-class I client is successful. The cloud data server generates client connection information and sends it to the edge data server scheduling module, including: If it is a Class I client, the cloud data server generates client connection information based on the location data of the Class I client and sends it to the edge server scheduling module; If it is a non-class I client, the cloud data server generates client connection information based on the location data and client identity information of the non-class I client, and sends it to the edge server scheduling module; The edge data server scheduling module schedules the first edge server to connect with the corresponding client according to the client connection information sent by the cloud data server, and the first edge data server executes the response to complete the edge computing data processing, including: The edge data server scheduling module obtains the client location information of the client connection information, matches the obtained client location information with each edge data server in the edge data server list, and the edge data server with the smallest location distance from the client location information is the first edge data server to be matched. According to the data transmission delay between the remaining edge data servers and the first edge data server, an edge computing sequence is generated, and the client is connected to the first edge data server; The first edge data server obtains the data segmentation feature, segments the to-be-processed data transmitted by the client according to the data segmentation feature, obtains multiple data blocks, and obtains a data block sequence; the first edge data server sends the preset data preprocessing model call information and the data blocks in the data block sequence to each edge data server in the edge computing sequence according to the edge computing sequence and the preset data preprocessing model call information; The first edge data server receives the preprocessed data blocks returned by each edge data server, and splices them according to the data block sequence to obtain the preprocessed data of the data to be processed, thereby completing the data edge computing.
2. The edge computing method based on artificial intelligence according to claim 1, characterized in that: The cloud data server receives the interaction information, obtains the corresponding first type of interaction characteristics according to the interaction information, and obtains the response state of the cloud data server according to the corresponding first type of interaction characteristics, including: The first interactive feature described is a request for edge data server computing power allocation; the edge data server is a server that provides edge computing support to the client; if there is an allocatable edge data server, the cloud data server responds, otherwise, the cloud data server does not respond.
3. The edge computing method based on artificial intelligence according to claim 2, characterized in that: The client type in the obtained interaction information includes: A client that matches the client database in the cloud data server is a first-class client; otherwise, it is a non-first-class client.
4. The edge computing cloud platform based on artificial intelligence is characterized by: The edge computing method based on artificial intelligence according to any one of claims 1 to 3 is applied, comprising: a cloud data server, an edge data server scheduling module, an edge data server, a communication module and a client; The cloud data server and edge data server scheduling module are respectively connected to the communication module; the edge data server is connected to the edge data server scheduling module; and the client is communicatively connected to the communication module.
5. The artificial intelligence-based edge computing cloud platform according to claim 4, characterized in that: The client comprises a first-class client, a non-first-class client and a client management module; the first-class client and the non-first-class client are respectively connected to the client management module; the client management module is connected to the communication module.
