Real-time risk monitoring and decision collaboration method for edge intelligent terminal

Abstract

The invention discloses a real-time risk monitoring and decision collaboration method for an edge intelligent terminal, and particularly relates to the field of industrial Internet of Things security monitoring and edge computing. The method comprises the following steps: collecting monitoring sub-region data through a multi-edge intelligent terminal, and calculating a local risk value; when the local risk value exceeds a threshold value, marking as a first edge intelligent terminal and broadcasting a collaboration request; the response edge intelligent terminal returns self risk data; the first edge intelligent terminal generates a final risk level through a weighted fusion algorithm according to a decision weight distribution strategy, and queries a rule base to generate a collaborative decision instruction; and sending a control instruction to the associated terminal based on the instruction, triggering equipment shutdown or data encryption operation, and feeding back an execution state. Through edge terminal localization calculation and collaborative decision, the problems of high response delay and high single-point fault risk of a centralized architecture are solved, millisecond risk response is realized, and the system stability and decision accuracy are improved.

Description

Technical Field

[0001] This invention relates to the field of industrial Internet of Things (IoT) security monitoring and edge computing technology, and more specifically, to a method for real-time risk monitoring and collaborative decision-making for edge intelligent terminals. Background Technology

[0002] In the field of industrial IoT security monitoring, existing technologies primarily rely on sensor nodes to collect environmental and equipment data, and then use centralized cloud platforms or servers for centralized risk analysis, assessment, and decision-making. Their function is to identify potential risks and ensure the safe operation of facilities; the core is to utilize central computing power for in-depth processing and global decision-making. The typical workflow involves sensor terminals collecting environmental and equipment parameters and transmitting them via network to a remote central platform. The central platform then analyzes the data using a risk assessment model, determines the risk level, generates decision instructions, and sends them back to the field for execution. The entire process depends on centralized data uploading and centralized processing and decision-making.

[0003] However, in actual use, it still has some drawbacks, such as the response delay caused by remote round-trip data transmission and processing, which cannot meet the risk scenarios that require millisecond-level response; the central node is a critical single point of failure; the existing centralized architecture has efficiency bottlenecks when integrating local information, and the failure of the central node may lead to system paralysis; the center needs to process global data to judge local risks, which is slow and cannot take advantage of the proximity of edge nodes to deal with them quickly. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art, the present invention provides a method for real-time risk monitoring and decision-making collaboration of edge intelligent terminals, which solves the problems mentioned in the background art through the following scheme.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for real-time risk monitoring and decision-making collaboration for edge intelligent terminals, comprising: S1: Set up multiple edge intelligent terminals, each edge intelligent terminal corresponding to a monitoring sub-area in the target area; S2: Each edge intelligent terminal collects environmental data and equipment operation data of the monitored sub-area in real time through its sensor module; Based on a pre-set risk assessment model and real-time collected data, the local risk value of each edge smart terminal is calculated through the data processing module. S3: Mark the edge intelligent terminal whose calculated local risk value exceeds the preset risk threshold as the first edge intelligent terminal. The first edge intelligent terminal generates and broadcasts collaborative decision request information through the communication module. The collaborative decision request information includes the identifier of the first edge intelligent terminal, the local risk value and related environmental parameters. S4: The edge intelligent terminal that receives the collaborative decision-making request information is the responding edge intelligent terminal; the responding edge intelligent terminal responds to the request through the communication module and sends its local risk value to the first edge intelligent terminal. The first edge intelligent terminal generates the final risk level based on a preset decision weight allocation strategy and a weighted fusion algorithm; Based on the final risk level, query the risk management rule base to generate collaborative decision-making instructions; S5: Based on the final risk level and collaborative decision-making instructions, the first edge intelligent terminal sends control instructions to itself and the responding edge intelligent terminal through the communication module. The control instructions trigger risk handling actions such as device shutdown or data encryption transmission, and feed back the decision results and execution status through its communication module.

[0006] Preferably, the construction of the risk assessment model includes: Feature extraction is performed on historical environmental data and equipment operation data to establish a risk feature vector; Based on the aforementioned risk feature vector, a risk classifier model is trained using a machine learning algorithm. The output of the risk classifier model is a risk probability value that represents the probability of an anomaly occurring.

[0007] Preferably, the local risk value is analyzed using the following methods: According to the formula: ,in The anomaly probability value output by the risk classifier model. This is the deviation between the sensor data and the reference value. The maximum allowable deviation threshold, The weighting coefficients are and satisfy the following conditions: .

[0008] Preferably, the generation method of the decision weight allocation strategy includes: (1) Assign weights based on the physical distance between each edge intelligent terminal and the location where the risk event occurs; (2) Adjust the weights dynamically based on the historical monitoring accuracy of each edge intelligent terminal.

[0009] (3) The final decision weight is generated by combining the distance weight and the accuracy weight.

[0010] Preferably, the calculation method for the final decision weight includes: According to the formula: ,in In response to edge intelligent terminals Distance weighting relative to the first edge intelligent terminal In response to edge intelligent terminals Distance to the first edge intelligent terminal The preset maximum association distance, For edge intelligent terminals Historical monitoring accuracy weighting The distance factor coefficient and .

[0011] Preferably, the specific analysis method for the final risk level includes: According to the formula: ,in The local risk value of the first edge intelligent terminal. For the first The local risk value of a response edge intelligent terminal. and The decision weight corresponding to the terminal, In response to the number of edge smart terminals.

[0012] Preferably, the risk threshold is set in the following ways: Multiple threshold levels are defined based on the security level of the target area; The threshold is dynamically reduced during the high-load operation phase of the equipment.

[0013] Preferably, the logic for generating the control commands includes: When the final risk level exceeds the preset second risk threshold, a device shutdown command is sent to the associated terminal. When the final risk level is between the first and second risk thresholds, a data encryption transmission instruction is sent to the terminal. Otherwise, maintain the terminal's normal operating status.

[0014] Preferably, the method further includes an exception handling mechanism: If no response is received from the edge smart terminal within the preset time, the backup communication protocol will be activated to resend the request. If the collaborative decision exceeds the preset timeout threshold, a single-point decision is executed based on the local risk value of the first edge intelligent terminal.

[0015] The technical effects and advantages of this invention are as follows: 1. Significantly reduced response latency, meeting millisecond-level decision-making requirements: Each edge intelligent terminal calculates its local risk value in real time through its data processing module. Only when the local risk value exceeds a threshold does the first edge intelligent terminal and the relevant responding edge intelligent terminal directly collaborate on a decision, eliminating the need for centralized processing on a remote central platform. The entire process of data processing and decision-making is completed locally between edge terminals, significantly shortening the data transmission and processing links, thereby achieving millisecond-level response. 2. Eliminating the risk of critical single point of failure and improving system stability: In existing technologies, the central node is a critical single point of failure. However, this invention adopts a distributed collaborative architecture of multiple edge intelligent terminals. The decision-making process is completed jointly by the first edge intelligent terminal and multiple responding edge intelligent terminals, eliminating the need for a single dependent central node. Even if some terminals fail, the remaining terminals can still work independently or collaboratively, eliminating the impact of single point of failure on the entire system and improving stability. 3. Fully leverage the potential of edge computing to achieve local collaborative decision-making and improve decision accuracy: The first edge intelligent terminal collaborates with the responding edge intelligent terminal, and based on a preset decision weight allocation strategy combined with the physical distance to the risk location and historical monitoring accuracy, generates the final risk level through a weighted fusion algorithm. This approach fully utilizes the local information of neighboring terminals, achieves local collaboration between edge nodes, avoids the shortcomings of central decision-making in integrating local information, improves decision accuracy, and fully leverages the potential of edge computing. 4. Leveraging the proximity advantage of edge nodes for rapid handling of localized risks: Edge intelligent terminals directly monitor sub-areas of the target area, enabling real-time collection and near-source processing of environmental and equipment operation data. When a localized risk arises, the first edge intelligent terminal can quickly collaborate with relevant response terminals to generate coordinated decision-making instructions and directly send control commands to itself and the response terminals to trigger response actions. This eliminates the need to wait for the central platform to process global data before making a decision, utilizing the advantages of edge nodes to achieve rapid handling of localized risks. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall structure of the present invention.

[0017] Figure 2 This is a schematic diagram of the dynamic collaborative decision-making process of the present invention.

[0018] Figure 3 This is a schematic diagram of the execution logic of the fault tolerance mechanism of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] refer to Figures 1-3 The method for real-time risk monitoring and collaborative decision-making of edge intelligent terminals, as shown, includes: S1: Set up multiple edge intelligent terminals, each edge intelligent terminal corresponding to a monitoring sub-area in the target area; S2: Each edge intelligent terminal collects environmental data and equipment operation data of the monitored sub-area in real time through its sensor module; Based on a pre-set risk assessment model and real-time collected data, the local risk value of each edge smart terminal is calculated through the data processing module. S3: Mark the edge intelligent terminal whose calculated local risk value exceeds the preset risk threshold as the first edge intelligent terminal. The first edge intelligent terminal generates and broadcasts collaborative decision request information through the communication module. The collaborative decision request information includes the identifier of the first edge intelligent terminal, the local risk value and related environmental parameters. S4: The edge intelligent terminal that receives the collaborative decision-making request information is the responding edge intelligent terminal; the responding edge intelligent terminal responds to the request through the communication module and sends its local risk value to the first edge intelligent terminal. The first edge intelligent terminal generates the final risk level based on a preset decision weight allocation strategy and a weighted fusion algorithm; Based on the final risk level, query the risk management rule base to generate collaborative decision-making instructions; S5: Based on the final risk level and collaborative decision-making instructions, the first edge intelligent terminal sends control instructions to itself and the responding edge intelligent terminal through the communication module. The control instructions trigger risk handling actions such as device shutdown or data encryption transmission, and feed back the decision results and execution status through its communication module.

[0021] Further explanation is needed: S1: Set up multiple edge intelligent terminals, each edge intelligent terminal corresponding to a monitoring sub-area in the target area.

[0022] It should be further explained that the target area is divided into multiple monitoring sub-areas, and an edge intelligent terminal is deployed in each monitoring sub-area. Through system configuration, the associated information is stored in the local storage module of each terminal.

[0023] S2: Each edge intelligent terminal collects environmental data and equipment operation data of the monitored sub-area in real time through its sensor module; Based on a pre-set risk assessment model and real-time collected data, the local risk value of each edge smart terminal is calculated through a data processing module.

[0024] The system collects environmental data in real time, including temperature, humidity, and dust concentration in the monitored sub-area; it also collects equipment operation data in real time, including voltage, current, speed, and vibration frequency in the monitored sub-area; and the data sampling frequency is set according to the monitoring requirements.

[0025] Calculate the local risk value: ,in The anomaly probability value output by the risk classifier model. This is the deviation between the sensor data and the reference value. The maximum allowable deviation threshold, The weighting coefficients are and satisfy the following conditions: .

[0026] It should be further explained that the risk assessment model extracts features from historical environmental data and equipment operation data to construct a risk feature vector, and then uses machine learning algorithms to train a risk classifier to output an anomaly probability value (between 0 and 1) that represents the probability of an anomaly occurring. Anomalies in equipment operation data can reflect risks more directly and reliably than environmental data.

[0027] Machine learning algorithms include support vector machines, random forests, and neural networks; risk classifier models include decision tree models and logistic regression models.

[0028] S3: Mark the edge intelligent terminal whose calculated local risk value exceeds the preset risk threshold as the first edge intelligent terminal. The first edge intelligent terminal generates and broadcasts collaborative decision request information through the communication module. The collaborative decision request information includes the identifier of the first edge intelligent terminal, the local risk value and related environmental parameters.

[0029] The "preset risk threshold" is a dynamically or statically set critical value, which is divided into multiple levels according to the security level of the target area (e.g., the threshold for high-risk areas is set to 0.6, and the threshold for medium-risk areas is set to 0.4), and the threshold is dynamically reduced during the high-load operation phase of the equipment (e.g., the threshold for high-risk areas is temporarily adjusted to 0.5); it is used to determine whether the current state of the terminal exceeds its independent processing capacity or security boundary. When the local risk value exceeds this threshold, the collaborative decision-making requirement is triggered. It should be further explained that the "collaborative decision-making request information" refers to direct communication between terminals via a wireless mesh network, employing a broadcast request plus directed response mode. The identifier in the collaborative decision-making request information is used to uniquely identify the terminal initiating the request, facilitating directed responses from other terminals. The local risk value quantifies the urgency or difficulty faced by the terminal and is a key basis for the recipient to assess whether and how to participate in the collaboration. Relevant environmental parameters (such as network status, geographical location, sensor data, task context, etc.) provide the background information required for decision-making, enabling potential assisting terminals to more accurately assess the suitability of their own resources, location, and capabilities, thereby making effective collaborative decisions.

[0030] S4: The edge intelligent terminal that receives the collaborative decision-making request information is a responding edge intelligent terminal; the responding edge intelligent terminal responds to the request through the communication module and sends its local risk value to the first edge intelligent terminal. The first edge intelligent terminal generates the final risk level based on a preset decision weight allocation strategy and a weighted fusion algorithm; Based on the final risk level, the risk management rule base is queried to generate collaborative decision-making instructions.

[0031] The steps for generating the final risk level are as follows: 1. Decision weight allocation strategies and collaborative decision instruction generation methods include: (1) Assign weights based on the physical distance between each edge intelligent terminal and the location where the risk event occurs; (2) Adjust the weights dynamically based on the historical monitoring accuracy of each edge intelligent terminal.

[0032] (3) The final decision weight is generated by combining the distance weight and the accuracy weight.

[0033] It needs further explanation regarding distance weight: the closer to the location where the risk occurs, the higher the weight. Calculation formula: Distance Weight ; Accuracy weight: Based on historical monitoring accuracy (e.g., TERM-002's accuracy rate over the past 30 days is 90%), the accuracy weight is 0.9. The accuracy weight is calculated as follows: ,in This represents the number of correct warnings issued in the past 24 hours (actual risk matches prediction). This represents the total number of warnings.

[0034] 2. The calculation methods for the final decision weight include: According to the formula: ,in In response to edge intelligent terminals Distance weighting relative to the first edge intelligent terminal In response to edge intelligent terminals Distance to the first edge intelligent terminal The preset maximum association distance, For edge intelligent terminals Historical monitoring accuracy weighting The distance factor coefficient and .

[0035] It needs to be further explained that, Historical data verification shows that: when At that time, the weighting of physical distance was too low, resulting in a 15% decrease in the accuracy of collaborative decision-making; when At that time, if the weight of nearby terminals is too high, the impact of edge node failures increases by 20%. In industrial scenarios, risks have significant spatial correlations. For example, the impact range of equipment failures and environmental anomalies usually decreases with increasing distance. Data collected by nearby edge intelligent terminals has a stronger correlation with current risk events and can better reflect the spatial distribution patterns of actual risks.

[0036] 3. The final risk level, and its specific analysis methods include: According to the formula: ,in The local risk value of the first edge intelligent terminal. For the first The local risk value of a response edge intelligent terminal. and The decision weight corresponding to the terminal, In response to the number of edge smart terminals.

[0037] It needs to be further explained that, , The primary terminal, as the source of risk triggering, has the strongest direct correlation between its own monitoring data and the current risk; therefore, it is assigned a fixed weight of 0.3 to reflect its fundamental decision-making role. The response terminal, as a spatially or logically related neighboring node, provides data that supplements information such as the risk spread range and the status of related devices. A weighted average of 0.7 is used to highlight the fusion value of collaborative data and avoid the potential bias of data from a single terminal. Risk data refers to the local risk values ​​sent by the response edge intelligent terminal to the primary edge intelligent terminal.

[0038] 4. Generate collaborative decision-making instructions by querying the risk management rule base based on the final risk level. The "risk management rule base" is a structured knowledge base or database that predefines specific collaborative decision-making instructions corresponding to different final risk level ranges.

[0039] S5: Based on the final risk level and collaborative decision-making instructions, the first edge intelligent terminal sends control instructions to itself and the responding edge intelligent terminal through the communication module. The control instructions trigger risk handling actions such as device shutdown or data encryption transmission, and feed back the decision results and execution status through its communication module.

[0040] The responding edge intelligent terminal is spatially or logically associated with the first edge intelligent terminal. Therefore, the responding edge intelligent terminal faces the same risks as the first edge intelligent terminal, necessitating risk mitigation actions via control commands. Control commands are operational commands that are specific to the collaborative decision-making commands generated by S4 and can directly drive terminal actuators or software modules. Their content precisely corresponds to the operation type (such as device shutdown or encrypted data transmission) defined for the final risk level in the risk mitigation rule base. The commands are explicitly issued to the responding edge intelligent terminal itself and all participating responding edge intelligent terminals, ensuring that collaborative actions are executed synchronously or sequentially within the relevant terminal group. The first risk threshold is 0.5, and the second risk threshold is 0.8.

[0041] It should be further explained that the second risk threshold of 0.8 serves as a high-risk threshold. When the final risk level exceeds this value, it indicates that the risk has reached a level requiring emergency handling (such as equipment failure potentially leading to a safety accident). Therefore, the highest level of risk control action, such as equipment shutdown, is triggered to quickly prevent the risk from escalating. The first risk threshold of 0.5 serves as a medium-risk threshold. When the final risk level is between 0.5 and 0.8, the risk has not yet reached the level of emergency shutdown, but there is a potential for data leakage or risk spread. Therefore, moderate-intensity handling actions, such as encrypted data transmission, are triggered to reduce the risk impact while ensuring equipment operation.

[0042] The exception handling mechanism: If no response is received from the edge smart terminal within the preset time, the backup communication protocol will be activated to resend the request. If the collaborative decision exceeds the preset timeout threshold, a single-point decision is executed based on the local risk value of the first edge intelligent terminal.

[0043] The process of initiating a backup communication protocol to resend a request includes the following key operations: 1. Intelligent Protocol Switching: When the primary communication protocol does not receive any response or data feedback from any responding terminal within a preset time (e.g., 3 seconds), the first edge intelligent terminal automatically switches to a pre-configured backup communication protocol stack. This stack will be dynamically selected based on environmental characteristics and urgency, for example, in extreme network outage situations, using a mobile terminal as a relay for direct device-to-device connection.

[0044] 2. Intelligent Adjustment Request: When retransmitting a request, the system will compress information based on the reason for the initial failure (e.g., retaining only the core identifier and risk value), increase transmission power, or broadcast to a specific subset of highly reliable terminals to improve the success rate. The number of retransmission attempts (e.g., a maximum of 3 times) and the interval (exponential backoff) are constrained by a preset strategy.

[0045] It should be further explained that the mechanism for performing single-point decisions based on local risk values ​​is activated only when the overall collaborative decision-making process times out (the total time from broadcast request to completion of fusion decision exceeds a preset threshold, such as 10 seconds). It is not equivalent to resending without response, but rather a final safety fallback strategy after the collaborative process fails.

[0046] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other. In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for real-time risk monitoring and collaborative decision-making at edge intelligent terminals, characterized in that, include: S1: Set up multiple edge intelligent terminals, each edge intelligent terminal corresponding to a monitoring sub-area in the target area; S2: Each edge intelligent terminal collects environmental data and equipment operation data of the monitored sub-area in real time through its sensor module; Based on a pre-set risk assessment model and real-time collected data, the local risk value of each edge smart terminal is calculated through the data processing module. S3: Mark the edge intelligent terminal whose calculated local risk value exceeds the preset risk threshold as the first edge intelligent terminal. The first edge intelligent terminal generates and broadcasts collaborative decision request information through the communication module. The collaborative decision request information includes the identifier of the first edge intelligent terminal, the local risk value and related environmental parameters. S4: The edge intelligent terminal that receives the collaborative decision-making request information is the responding edge intelligent terminal; the responding edge intelligent terminal responds to the request through the communication module and sends its local risk value to the first edge intelligent terminal. The first edge intelligent terminal generates the final risk level based on a preset decision weight allocation strategy and a weighted fusion algorithm; Based on the final risk level, query the risk management rule base to generate collaborative decision-making instructions; S5: Based on the final risk level and collaborative decision-making instructions, the first edge intelligent terminal sends control instructions to itself and the responding edge intelligent terminal through the communication module. The control instructions trigger risk handling actions such as device shutdown or data encryption transmission, and feed back the decision results and execution status through its communication module.

2. The method for real-time risk monitoring and decision-making collaboration of an edge intelligent terminal according to claim 1, characterized in that, The construction of the risk assessment model includes: Feature extraction is performed on historical environmental data and equipment operation data to establish a risk feature vector; Based on the aforementioned risk feature vector, a risk classifier model is trained using a machine learning algorithm. The output of the risk classifier model is a risk probability value that represents the probability of an anomaly occurring.

3. The method for real-time risk monitoring and decision-making collaboration of an edge intelligent terminal according to claim 2, characterized in that, The specific analysis methods for the local risk value include: According to the formula: ,in The anomaly probability value output by the risk classifier model. This is the deviation between the sensor data and the reference value. The maximum allowable deviation threshold, The weighting coefficients are and satisfy the following conditions: .

4. The method for real-time risk monitoring and decision-making collaboration of an edge intelligent terminal according to claim 1, characterized in that, The generation method of the decision weight allocation strategy includes: (1) Assign weights based on the physical distance between each edge intelligent terminal and the location where the risk event occurs; (2) Adjust the weights dynamically based on the historical monitoring accuracy of each edge intelligent terminal; (3) The final decision weight is generated by combining the distance weight and the accuracy weight.

5. The method for real-time risk monitoring and decision-making collaboration of an edge intelligent terminal according to claim 4, characterized in that, The calculation method for the final decision weight includes: According to the formula: ,in In response to edge intelligent terminals Distance weighting relative to the first edge intelligent terminal In response to edge intelligent terminals Distance to the first edge intelligent terminal The preset maximum association distance, In response to edge intelligent terminals Historical monitoring accuracy weighting The distance factor coefficient and .

6. The method for real-time risk monitoring and decision-making collaboration of an edge intelligent terminal according to claim 1, characterized in that, The specific analysis methods for the final risk level include: According to the formula: ,in The local risk value of the first edge intelligent terminal. For the first The local risk value of a response edge intelligent terminal. and The decision weight corresponding to the terminal, In response to the number of edge smart terminals.

7. The method for real-time risk monitoring and decision-making collaboration of an edge intelligent terminal according to claim 1, characterized in that, The risk threshold is set in the following ways: Multiple threshold levels are defined based on the security level of the target area; The threshold is dynamically reduced during the high-load operation phase of the equipment.

8. The method for real-time risk monitoring and decision-making collaboration of an edge intelligent terminal according to claim 1, characterized in that, The logic for generating the control commands includes: When the final risk level exceeds the preset second risk threshold, a device shutdown command is sent to the associated terminal. When the final risk level is between the first and second risk thresholds, a data encryption transmission instruction is sent to the terminal. Otherwise, maintain the terminal's normal operating status.

9. The method for real-time risk monitoring and decision-making collaboration of an edge intelligent terminal according to claim 1, characterized in that, It also includes an exception handling mechanism: If no response is received from the edge smart terminal within the preset time, the backup communication protocol will be activated to resend the request. If the collaborative decision exceeds the preset timeout threshold, a single-point decision is executed based on the local risk value of the first edge intelligent terminal.

Images