Multi-terminal data transmission detection method and system based on artificial intelligence

Abstract

The invention discloses a multi-terminal data transmission detection method and system based on artificial intelligence, relates to the technical field of network communication, and aims to solve the problems of data consistency, security and efficiency in a current heterogeneous network, and the adopted scheme comprises the steps that a cloud connection layer is constructed, a plurality of edge nodes are configured, and each edge node is connected with a specified terminal; performing logic grouping on the edge nodes corresponding to the plurality of terminals with interaction requirements; each edge node is responsible for monitoring the real-time operation state of the terminal connected with the edge node and collecting monitoring data; the cloud connection layer receives the monitoring data, evaluates the consistency among the monitoring data in the same group based on an artificial intelligence prediction model, and obtains a difference value; and the cloud connection layer generates and sends an adjustment signal according to the difference value, and the edge node adjusts the corresponding terminal according to the adjustment signal so as to ensure that the data interaction between the associated terminals in the same group is kept consistent. The method is suitable for the current complex heterogeneous internet environment.

Description

Technical Field

[0001] The present invention relates to the field of network communication technology, and in particular to a multi-terminal data transmission detection method and system based on artificial intelligence. Background Art

[0002] As a cutting-edge science, artificial intelligence is committed to exploring and developing theories, methods and technologies to simulate and expand human intelligence. It is an important branch of computer science that focuses on analyzing the essence of intelligence and strives to create intelligent systems that can perceive, learn, reason and interact in a human-like manner.

[0003] Heterogeneous network transmission refers to the process of realizing data exchange in a comprehensive network system composed of different types of technologies and equipment. Such networks may include but are not limited to wired networks, wireless networks, sensor networks, and satellite networks. By integrating these different sub-networks, heterogeneous networks achieve seamless transmission and sharing of data and information in a diverse environment.

[0004] In a complex network environment, it is particularly important to ensure the security of communication between different terminals, data consistency and privacy protection. Especially in a heterogeneous network environment, since a large amount of data interaction between multiple terminals is used for decision-making, analysis and processing, and control operations, if there are inconsistencies or security vulnerabilities in the data between terminals, it may not only lead to incorrect analysis results and decision-making errors, but also may cause privacy issues due to the leakage of sensitive information (such as personal identity information, financial information, etc.).

[0005] Therefore, it is necessary to provide an artificial intelligence-based multi-terminal data transmission detection method and system to solve the above technical problems. Summary of the invention

[0006] In response to the needs and shortcomings of current technological development, the present invention provides an artificial intelligence-based multi-terminal data transmission detection method and system, which solves the problems of data consistency, security and efficiency in heterogeneous networks through the deep integration of edge computing and cloud computing.

[0007] In the first aspect, the present invention provides a multi-terminal data transmission detection method based on artificial intelligence, and the technical solution adopted to solve the above technical problems is as follows: A multi-terminal data transmission detection method based on artificial intelligence comprises the following steps: Build a cloud connectivity layer; In the cloud connection layer, several edge nodes are configured, each of which is connected to a designated terminal; Logically group the edge nodes corresponding to multiple terminals with interaction requirements; Each edge node is responsible for monitoring the real-time operating status of the terminals connected to it and collecting monitoring data; The cloud connection layer receives all monitoring data from edge nodes, evaluates the consistency between monitoring data in the same group based on the artificial intelligence prediction model, and obtains the difference value; The cloud connection layer generates and sends adjustment signals to the corresponding edge nodes based on the difference values; The edge node that receives the adjustment signal adjusts the corresponding terminal according to the instruction to ensure that the data interaction between the terminals in the same group that have an interactive relationship remains highly consistent.

[0008] Optionally, edge nodes corresponding to multiple terminals with interaction requirements are logically grouped, and the logical grouping process is as follows: The cloud connection layer identifies the interaction requirements between terminals based on the monitoring data and terminal interaction logs of previous terminals. Based on the above interaction requirements, a correlation graph is constructed. Each node in the above correlation graph represents an edge node, and an edge represents the interaction relationship between the terminals connected by two edge nodes. A clustering algorithm is used to group the nodes in the association graph to ensure that the interaction density between nodes within each group is the highest and the cross-group interaction is the least.

[0009] Optionally, monitoring data includes: Network performance indicators: bandwidth, latency, and throughput; Security indicators: traffic entropy value, encryption protocol compliance; Business indicator: task execution status code.

[0010] Optionally, after each edge node collects the monitoring data, it first performs preprocessing operations of data cleaning and normalization, and then sends it to the cloud connection layer.

[0011] Further optionally, an artificial intelligence prediction model is constructed based on monitoring data of different indicators. The artificial intelligence prediction model construction method includes: Build a database with three partitions, corresponding to network performance indicators, security indicators, and business indicators; Collect monitoring data from edge nodes and store them in corresponding partitions according to the type of monitoring data; Divide the historical monitoring data in the database into a training set and a test set; Use the training set to train different preset machine learning algorithms to generate a preliminary prediction model; Use the test set to verify and evaluate multiple trained prediction models and select the prediction model with the best performance; Three artificial intelligence prediction models suitable for network performance, security and business scenarios are obtained respectively.

[0012] Further optionally, the consistency between the monitoring data in the same group is evaluated based on the artificial intelligence prediction model, and the evaluation process is divided into two stages: edge computing stage and cloud computing stage; In the edge computing stage, the cloud connection layer assigns relatively simple evaluation tasks to edge nodes. These evaluation tasks belong to the group to which the edge nodes belong. The edge nodes are responsible for evaluating and generating consistency reports and difference values, and sending them back to the cloud connection layer. At the same time, the edge nodes also automatically generate adjustment signals based on the difference values ​​and send them to the task-related edge nodes other than themselves, so as to make immediate adjustments to the corresponding terminals. In the cloud computing stage, the cloud connection layer centrally handles relatively complex evaluation tasks, specifically involving security and business-level tasks.

[0013] Optionally, the consistency between the monitoring data in the same group is evaluated based on the artificial intelligence prediction model. During the evaluation process, the edge node and the cloud connection layer jointly participate in the decision-making and control process, including: Create a task queue, and all newly arrived tasks are added to the queue in order; The scheduler takes a task from the front of the queue and processes it until the queue is empty; Based on the principle of shortest expected execution time of tasks, tasks are preferentially assigned to edge nodes for processing to reduce latency.

[0014] Further optionally, the edge nodes and the cloud connection layer jointly participate in the decision-making and control process. In this process, the execution order is determined according to the priority of the tasks, wherein static priority allocation is pre-set based on the importance and urgency of the tasks.

[0015] In a second aspect, the present invention provides a multi-terminal data transmission detection system based on artificial intelligence, and the technical solution adopted to solve the above technical problems is as follows: A multi-terminal data transmission detection system based on artificial intelligence, based on the multi-terminal data transmission detection method based on artificial intelligence in the first aspect, its implementation structure includes: Cloud connectivity layer; Edge nodes: One end of several edge nodes is connected to the cloud connection layer, and the other end is connected to the designated terminal; A grouping unit, used to logically group edge nodes corresponding to multiple terminals with interaction requirements; The monitoring unit is configured on the edge node and is used to monitor the real-time operating status of the terminals connected to it and collect relevant data; Artificial intelligence prediction model is used to evaluate the consistency between monitoring data in the same group and obtain the difference value; An adjustment signal generating unit generates an adjustment signal according to the difference value and sends the adjustment signal to the corresponding edge node; The control unit is configured on the edge node, receives the adjustment signal, and implements necessary adjustment measures on the corresponding terminal to ensure that the data interaction between the terminals in the group that have an interactive relationship maintains a high degree of consistency.

[0016] Optionally, the construction process of the artificial intelligence prediction model involved includes: Build a database with three partitions, corresponding to network performance indicators, security indicators, and business indicators; Collect monitoring data from edge nodes and store them in corresponding partitions according to the type of monitoring data; Divide the historical monitoring data in the database into a training set and a test set; Use the training set to train different preset machine learning algorithms to generate a preliminary prediction model; Use the test set to verify and evaluate multiple trained prediction models and select the prediction model with the best performance; Three artificial intelligence prediction models suitable for network performance, security and business scenarios are obtained respectively.

[0017] The multi-terminal data transmission detection method and system based on artificial intelligence of the present invention have the following beneficial effects compared with the prior art: 1. The present invention is applicable to the current complex heterogeneous Internet environment. In such an environment, the network performance, security and business indicators of different terminals usually vary significantly. By deploying edge nodes to monitor the status of each terminal in real time and using advanced artificial intelligence prediction models to analyze and adjust the collected data, the reliability, consistency and privacy protection of communication and data interaction between different terminals can be effectively monitored and guaranteed. This not only improves the stability of the entire heterogeneous Internet network, but also ensures normal communication and efficient data exchange between terminals. 2. In the process of real-time monitoring of heterogeneous Internet terminals, the present invention cleverly combines edge computing with cloud computing, where edge nodes are responsible for performing some preliminary data processing and simple tasks, while more complex tasks are undertaken by the cloud connection layer. This approach not only reduces unnecessary data transmission and reduces network load, but also reduces the computing pressure of the cloud connection layer, thereby greatly shortening the data processing time and improving the real-time response capability and operating efficiency of the system; 3. The present invention can dynamically allocate tasks that need to be evaluated to the cloud connection layer or edge nodes. This allocation automatically adjusts the detection cycle according to the real-time load and energy consumption, thereby achieving the best balance between load, energy consumption and delayed response. It not only improves the precision and accuracy of the system, but also enhances its real-time performance, and further strengthens the stability and reliability of the system. In addition, continuous monitoring and in-depth analysis of terminal devices combined with artificial intelligence technology can significantly improve the overall performance and efficiency of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Attached Figure 1 It is a flow chart of a method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the technical solution, the technical problem solved and the technical effect of the present invention more clearly understood, the technical solution of the present invention is clearly and completely described below in conjunction with specific embodiments.

[0020] Embodiment 1: Refer to the attached Figure 1 , this embodiment proposes a multi-terminal data transmission detection method based on artificial intelligence, which includes the following steps: S1. Use Kubernetes-based control nodes and edge computing frameworks (such as KubeEdge) to build a cloud connection layer to achieve unified management of cloud and edge nodes.

[0021] Generally speaking, the cloud connection layer needs to integrate the following modules: Data synchronization and communication module: supports MQTT / HTTP protocol to ensure real-time data transmission between the cloud and edge nodes, and built-in TLS encryption to ensure communication security; AI model deployment and update: The cloud stores trained network performance, security, and business prediction models, and dynamically sends them to edge nodes through containerization technology; Elastic resource scheduling: Dynamically allocate computing resources based on edge node loads, such as optimizing cloud computing capabilities through Kubernetes automatic scaling mechanisms.

[0022] S2. In the cloud connection layer, several edge nodes are configured, and each edge node is connected to a designated terminal.

[0023] Specifically, each edge node is connected to the designated terminal through a dedicated communication protocol (such as Modbus or OPC UA), which ensures low latency and data integrity. At the same time, a strong password policy (≥12 mixed characters) is set, non-essential services are disabled, security patches are updated regularly, and security is enhanced through Huawei Cloud ECS key login.

[0024] All edge nodes are connected to the cloud connection layer for interaction and data transmission.

[0025] S3. Logically group the edge nodes corresponding to multiple terminals with interaction requirements. The logical grouping process is as follows: The cloud connection layer identifies the interaction needs between terminals based on the monitoring data of each terminal in the past and the terminal interaction logs (such as task execution frequency and data traffic pattern); Based on the above interaction requirements, a correlation graph is constructed. Each node in the above correlation graph represents an edge node, and an edge represents the interaction relationship between the terminals connected by two edge nodes. A clustering algorithm is used to group the nodes in the association graph to ensure that the interaction density between nodes within each group is the highest and the cross-group interaction is the least.

[0026] Subsequently, the group-specific AI prediction models constructed later (such as the AI ​​prediction model for network performance, the AI ​​prediction model for security, and the AI ​​prediction model for business scenarios) are sent to the edge nodes to support local incremental updates to reduce cloud dependence.

[0027] S4. Each edge node is responsible for monitoring the real-time operating status of the terminal connected to it and collecting monitoring data. The monitoring data includes: (1) Network performance indicators: bandwidth, latency, and throughput; (2) Security indicators: traffic entropy value, encryption protocol compliance; for traffic entropy value, based on NetFlow data flow statistics, calculate the Shannon entropy value of source IP, destination IP, and port number, and trigger an abnormal alarm when the entropy value suddenly increases (such as exceeding the baseline by 30%); for encryption protocol compliance, parse the SNI (Server Name Indication) and ALPN (Application Layer Protocol Negotiation) fields of TLS traffic to verify whether only TLS 1.2 / 1.3 protocols are allowed. Use a preset whitelist (such as the National Encryption SM2 / SM4 certificate) to check the validity of the certificate chain and reject self-signed or expired certificates; (3) Business indicators: task execution status code; for the task execution status code, the business log of the terminal is subscribed through HTTP API polling (interval of 1 second) or WebSocket long connection to extract the status code (such as 200 / success, 500 / server error), and supports custom business status code mapping (such as 0x0A01 means "control instruction execution timeout").

[0028] After each edge node collects the monitoring data, it first performs data cleaning and normalization preprocessing operations and then sends it to the cloud connection layer.

[0029] Data cleaning specifically includes: ① Missing value processing: For time series data such as bandwidth and delay, linear interpolation (window size = 3) is used to fill short-term missing values; if the continuous missing exceeds 5 cycles, it is marked as "data unavailable" and the terminal is triggered to reconnect; ② Outlier filtering: Based on the Z-score algorithm (threshold |Z|≥3), abnormal data is detected. For example, when the delay suddenly increases (such as jumping from 50ms to 500ms), the abnormal point is discarded and the alarm event is recorded; ③ Check the legitimacy of the business status code (if it is not the preset status code, it is considered invalid), and associate the terminal log to locate the source of the error.

[0030] Normalization specifically includes: ① Numerical indicators: Use Min-Max standardization to scale indicators such as bandwidth and delay to the [0,1] interval; use logarithmic transformation for skewed distribution data (such as traffic entropy); ② Categorical indicators: One-Hot encoding for labels such as encryption protocol compliance; ③ Keep the structure and form of each monitoring data in the same group consistent to facilitate subsequent analysis and processing.

[0031] Of course, in the actual preprocessing operation, after performing data cleaning and normalization operations, the edge node also performs real-time stream processing: using a lightweight stream processing engine (such as Apache Kafka Streams or Flink Edge), window aggregation (such as 10-second rolling window calculation of average delay) and complex event processing (CEP) are implemented at the edge node. In addition, the edge node also performs resource optimization strategies: a) Data sampling: In high-load scenarios, random sampling (sampling rate 50%) is enabled for non-critical indicators (such as historical throughput), and the Bernoulli sampling algorithm is used to reduce computing overhead; b) Compressed transmission: Use the LZ4 or Zstandard algorithm to compress the preprocessed data, with a compression ratio of up to 5:1, reducing the amount of data transmitted to the cloud.

[0032] S5. The cloud connection layer receives all monitoring data from edge nodes, and then evaluates the consistency between the monitoring data in the same group based on the artificial intelligence prediction model to obtain the difference value.

[0033] The artificial intelligence prediction model described in this step is constructed based on the monitoring data of different indicators. The specific construction process includes: Build a database with three partitions, corresponding to network performance indicators, security indicators, and business indicators; Collect monitoring data from edge nodes and store them in corresponding partitions according to the type of monitoring data; Divide the historical monitoring data in the database into a training set and a test set; Use the training set to train different preset machine learning algorithms to generate a preliminary prediction model; Use the test set to verify and evaluate multiple trained prediction models and select the prediction model with the best performance; Three artificial intelligence prediction models suitable for network performance, security and business scenarios are obtained respectively.

[0034] Based on different indicators, the corresponding artificial intelligence prediction models are used to evaluate the consistency between the monitoring data in the same group and obtain the difference value. This evaluation process can be divided into two stages: edge computing stage and cloud computing stage. In the edge computing stage, the cloud connection layer assigns relatively simple evaluation tasks to edge nodes. These evaluation tasks belong to the group where the edge nodes are located. The edge nodes are responsible for evaluating and generating consistency reports and difference values, and sending them back to the cloud connection layer; at the same time, the edge nodes also automatically generate adjustment signals based on the difference values ​​and send them to other edge nodes related to the task, so as to make immediate adjustments to the corresponding terminals. In the cloud computing stage, the cloud connection layer centrally handles relatively complex evaluation tasks, which specifically involve tasks at the security and business levels. This method in which edge nodes are responsible for performing some preliminary data processing and simple tasks, while more complex tasks are undertaken by the cloud connection layer, can not only reduce unnecessary data transmission and reduce network load, but also reduce the computing pressure of the cloud connection layer, thereby greatly shortening the data processing time and improving the real-time response capability and operation efficiency of the system.

[0035] It should be added that the consistency between the monitoring data in the same group is evaluated based on the artificial intelligence prediction model. During the evaluation process, the edge nodes and the cloud connection layer jointly participate in the decision-making and control process, including: Create a task queue, and all newly arrived tasks are added to the queue in order; The scheduler takes a task from the front of the queue and processes it until the queue is empty; Based on the principle of shortest expected execution time of tasks, tasks are preferentially assigned to edge nodes for processing to reduce latency.

[0036] The edge nodes and the cloud connection layer jointly participate in the decision-making and control process. In this process, the execution order is determined according to the priority of the task. Among them, static priority allocation is pre-set based on the importance and urgency of the task.

[0037] S6. The cloud connection layer generates and sends an adjustment signal to the corresponding edge node based on the difference value. Specifically: If the bandwidth difference is ≥20% or the delay difference is ≥30ms, the dynamic bandwidth allocation or routing optimization strategy is triggered; The traffic entropy value suddenly increases (such as Δ≥1.5), triggering traffic cleaning or protocol blocking strategies; Encryption protocol violations (such as SSLv3 detection) trigger the mandatory protocol upgrade policy.

[0038] Continuous status code errors (such as 5xx errors lasting for 10 seconds) trigger task rescheduling or redundant terminal switching strategies.

[0039] For the above strategies, the following strategy priority mechanism can be set: The priority of security adjustments (such as protocol blocking) is P0, that of network optimization is P1, and that of service adjustments is P2.

[0040] The conflicting strategies generate the final instructions through weighted voting (e.g., security weight = 0.6, network weight = 0.3).

[0041] Transmission protocol: Based on MQTT 5.0's QoS 2 (Exactly Once), it guarantees that the command will be delivered and supports offline caching (Retain Message). It also uses a two-way confirmation process: the cloud connection layer sends the command to the edge node, marked as UNACK status; after the edge node replies ACK, the command status changes to PROCESSING; after the execution is completed, the result code (0 = success, 1 = failure) is returned, and the cloud layer is updated to COMPLETED or FAILED.

[0042] S7. The edge node that receives the adjustment signal adjusts the corresponding terminal according to the instruction to ensure that data interaction between terminals in an interactive relationship within the same group maintains a high degree of consistency.

[0043] Embodiment 2: Based on the artificial intelligence-based multi-terminal data transmission detection method described in Embodiment 1, this embodiment proposes an artificial intelligence-based multi-terminal data transmission detection system, whose implementation structure includes: Cloud connectivity layer; Edge nodes: One end of several edge nodes is connected to the cloud connection layer, and the other end is connected to the designated terminal; A grouping unit, used to logically group edge nodes corresponding to multiple terminals with interaction requirements; The monitoring unit is configured on the edge node and is used to monitor the real-time operating status of the terminals connected to it and collect relevant data; Artificial intelligence prediction model is used to evaluate the consistency between monitoring data in the same group and obtain the difference value; An adjustment signal generating unit generates an adjustment signal according to the difference value and sends the adjustment signal to the corresponding edge node; The control unit is configured on the edge node, receives the adjustment signal, and implements necessary adjustment measures on the corresponding terminal to ensure that the data interaction between the terminals in the group that have an interactive relationship maintains a high degree of consistency.

[0044] The control unit at the edge node receives the adjustment signal and implements necessary adjustment measures on the corresponding terminal, which may include: 1. In terms of data processing and synchronization, data correction is performed. If the monitoring data shows that there are differences in the data of terminals in the same group, the control unit will correct the data based on the difference value evaluated by the artificial intelligence prediction model. For example, in a multi-sensor monitoring system, if the data collected by some sensor terminals deviates from that of other sensors in the same group, the control unit of the edge node will correct these deviation data to make them consistent with other normal data in the group.

[0045] 2. In terms of communication optimization, adjust the communication frequency or switch the communication link. Specifically, (a) the control unit will adjust the communication frequency of the terminal according to the real-time operating status and data interaction of the monitored terminal; if it is found that the data interaction volume of some terminals is large and the network bandwidth is limited, the control unit may appropriately reduce the communication frequency of some non-critical data to avoid network congestion and ensure that critical data can be transmitted in a timely and accurate manner; for example, in a smart factory, when the operating data transmission volume of production equipment is large, the control unit may reduce the transmission frequency of some equipment status monitoring data to give priority to ensuring the stable transmission of production data. (b) If the current communication link is detected to be faulty or unstable, the control unit will switch to the backup communication link in time. For example, in an IoT application, the edge node originally connects to the terminal via Wi-Fi, but if the Wi-Fi signal is interfered or interrupted, the control unit will automatically switch the terminal connection to Bluetooth or other available communication methods to ensure that data interaction is not interrupted.

[0046] 3. In terms of terminal device configuration, adjust parameters, or restart or reset the device. Specifically, (a) based on monitoring data and system requirements, the control unit may adjust some parameters of the terminal. For example, in an intelligent lighting system, the control unit can adjust the brightness parameters of the lighting terminal according to the ambient light intensity and personnel activities; in a video surveillance system, adjust the resolution parameters of the camera according to the image quality and storage capacity to achieve the best operating effect and data interaction quality. (b) When the terminal device fails or is abnormal, causing problems in data interaction, the control unit can restart or reset the terminal device. For example, when a smart terminal freezes or has a software failure, the control unit of the edge node can remotely send instructions to restart the terminal device to restore it to normal working state and ensure that data interaction with other terminals in the group can proceed normally.

[0047] The construction process of the artificial intelligence prediction model involved in this embodiment includes: Build a database with three partitions, corresponding to network performance indicators, security indicators, and business indicators; Collect monitoring data from edge nodes and store them in corresponding partitions according to the type of monitoring data; Divide the historical monitoring data in the database into a training set and a test set; Use the training set to train different preset machine learning algorithms to generate a preliminary prediction model; Use the test set to verify and evaluate multiple trained prediction models and select the prediction model with the best performance; Three artificial intelligence prediction models suitable for network performance, security and business scenarios are obtained respectively.

[0048] In summary, the artificial intelligence-based multi-terminal data transmission detection method and system of the present invention can effectively monitor and ensure the reliability, consistency and privacy protection of communication and data interaction between different terminals by deploying edge nodes to monitor the status of each terminal in real time, and using advanced artificial intelligence prediction models to analyze and adjust the collected data. It not only improves the stability of the entire heterogeneous Internet network, but also ensures normal communication and efficient data exchange between terminals.

[0049] The above specific examples are used to explain the principles and implementation methods of the present invention in detail. These examples are only used to help understand the core technical content of the present invention. Based on the above specific embodiments of the present invention, any improvements and modifications made by technicians in this technical field without departing from the principles of the present invention should fall within the scope of patent protection of the present invention.

Claims

1. A multi-terminal data transmission detection method based on artificial intelligence, characterized in that: The following steps are involved: Build a cloud connectivity layer; In the cloud connection layer, several edge nodes are configured, each of which is connected to a designated terminal; Logically group the edge nodes corresponding to multiple terminals with interaction requirements; Each edge node is responsible for monitoring the real-time operating status of the terminals connected to it and collecting monitoring data; The cloud connection layer receives all monitoring data from edge nodes, evaluates the consistency between monitoring data in the same group based on the artificial intelligence prediction model, and obtains the difference value; The cloud connection layer generates and sends adjustment signals to the corresponding edge nodes based on the difference values; The edge node that receives the adjustment signal adjusts the corresponding terminal according to the instruction to ensure that the data interaction between the terminals in the same group that have an interactive relationship remains highly consistent.

2. According to the artificial intelligence-based multi-terminal data transmission detection method of claim 1, it is characterized in that: The edge nodes corresponding to multiple terminals with interaction requirements are logically grouped. The logical grouping process is as follows: The cloud connection layer identifies the interaction requirements between terminals based on the monitoring data and terminal interaction logs of previous terminals. Based on the above interaction requirements, a correlation graph is constructed. Each node in the above correlation graph represents an edge node, and an edge represents the interaction relationship between the terminals connected by two edge nodes. A clustering algorithm is used to group the nodes in the association graph to ensure that the interaction density between nodes within each group is the highest and the cross-group interaction is the least.

3. The method for detecting multi-terminal data transmission based on artificial intelligence according to claim 1 is characterized in that: The monitoring data include: Network performance indicators: bandwidth, latency, and throughput; Security indicators: traffic entropy value, encryption protocol compliance; Business indicator: task execution status code.

4. The method for detecting multi-terminal data transmission based on artificial intelligence according to claim 3 is characterized in that: After each edge node collects the monitoring data, it first performs data cleaning and normalization preprocessing operations and then sends it to the cloud connection layer.

5. The method for detecting multi-terminal data transmission based on artificial intelligence according to claim 3 is characterized in that: Based on the monitoring data of different indicators, an artificial intelligence prediction model is constructed. The artificial intelligence prediction model construction method includes: Build a database with three partitions, corresponding to network performance indicators, security indicators, and business indicators; Collect monitoring data from edge nodes and store them in corresponding partitions according to the type of monitoring data; Divide the historical monitoring data in the database into a training set and a test set; Use the training set to train different preset machine learning algorithms to generate a preliminary prediction model; Use the test set to verify and evaluate multiple trained prediction models and select the prediction model with the best performance; Three artificial intelligence prediction models suitable for network performance, security and business scenarios are obtained respectively.

6. The method for detecting multi-terminal data transmission based on artificial intelligence according to claim 5 is characterized in that: The consistency between the monitoring data in the same group is evaluated based on the artificial intelligence prediction model. The evaluation process is divided into two stages: edge computing stage and cloud computing stage; In the edge computing stage, the cloud connection layer assigns relatively simple evaluation tasks to edge nodes. These evaluation tasks belong to the group to which the edge nodes belong. The edge nodes are responsible for evaluating and generating consistency reports and difference values, and sending them back to the cloud connection layer. At the same time, the edge nodes also automatically generate adjustment signals based on the difference values ​​and send them to the task-related edge nodes other than themselves, so as to make immediate adjustments to the corresponding terminals. In the cloud computing stage, the cloud connection layer centrally handles relatively complex evaluation tasks, specifically involving security and business-level tasks.

7. The method for detecting multi-terminal data transmission based on artificial intelligence according to claim 6 is characterized in that: Based on the artificial intelligence prediction model, the consistency between the monitoring data in the same group is evaluated. During the evaluation process, the edge nodes and the cloud connection layer jointly participate in the decision-making and control process, including: Create a task queue, and all newly arrived tasks are added to the queue in order; The scheduler takes a task from the front of the queue and processes it until the queue is empty; Based on the principle of shortest expected execution time of tasks, tasks are preferentially assigned to edge nodes for processing to reduce latency.

8. The method for detecting multi-terminal data transmission based on artificial intelligence according to claim 7 is characterized in that: The edge nodes and the cloud connection layer jointly participate in the decision-making and control process. In this process, the execution order is determined according to the priority of the task. Among them, static priority allocation is pre-set based on the importance and urgency of the task.

9. A multi-terminal data transmission detection system based on artificial intelligence, characterized in that: A multi-terminal data transmission detection method based on artificial intelligence according to any one of claims 1 to 8, wherein the implementation structure includes: Cloud connectivity layer; Edge nodes: One end of several edge nodes is connected to the cloud connection layer, and the other end is connected to the designated terminal; A grouping unit, used to logically group edge nodes corresponding to multiple terminals with interaction requirements; The monitoring unit is configured on the edge node and is used to monitor the real-time operating status of the terminals connected to it and collect relevant data; Artificial intelligence prediction model is used to evaluate the consistency between monitoring data in the same group and obtain the difference value; An adjustment signal generating unit generates an adjustment signal according to the difference value and sends the adjustment signal to the corresponding edge node; The control unit is configured on the edge node, receives the adjustment signal, and implements adjustment measures on the corresponding terminal to ensure that the data interaction between the terminals in the group that have an interactive relationship maintains a high degree of consistency.

10. The multi-terminal data transmission detection system based on artificial intelligence according to claim 9 is characterized in that: The construction process of the artificial intelligence prediction model includes: Build a database with three partitions, corresponding to network performance indicators, security indicators, and business indicators; Collect monitoring data from edge nodes and store them in corresponding partitions according to the type of monitoring data; Divide the historical monitoring data in the database into a training set and a test set; Use the training set to train different preset machine learning algorithms to generate a preliminary prediction model; Use the test set to verify and evaluate multiple trained prediction models and select the prediction model with the best performance; Three artificial intelligence prediction models suitable for network performance, security and business scenarios are obtained respectively.

Images