A distributed disaster event-driven dynamic hierarchical early warning and local decision method

Abstract

The application discloses a kind of distributed disaster event-driven dynamic hierarchical early warning and local decision-making methods, specifically: multiple monitoring nodes detect abnormal signal to generate event data packet and broadcast to user terminal;Each user terminal independently receives multi-source event data, completes spatiotemporal consistency analysis, confidence calculation, risk classification and control strategy retrieval locally, does not rely on central node to make unified determination or issue control instruction;Terminal combines event space position, equipment current position and predicted position, and matches control strategy and executes hierarchical response from the spatial risk mapping database established in advance.The application adopts decentralized early warning decision architecture, each terminal as an independent decision center, reduces response delay, eliminates single point failure risk, improves system reliability and event determination accuracy, supports risk level dynamic adjustment and terminal local autonomous decision-making.

Description

Technical Field

[0001] This invention belongs to the field of disaster monitoring, early warning and emergency control technology, and specifically relates to a distributed disaster event-driven dynamic hierarchical early warning and local decision-making method. Background Technology

[0002] Disaster early warning systems are important infrastructure for ensuring public safety. Existing disaster early warning systems generally adopt a centralized architecture. Their basic working mode is as follows: each monitoring node collects raw sensor data and uploads it to the central server through a communication link. The central server completes data aggregation, event identification, and early warning information dissemination. Terminal devices passively receive the early warning results and execute response actions.

[0003] However, the above centralized architecture has the following prominent problems in practical applications: First, the response delay is relatively large: all raw data must be transmitted to the central server through the communication link before event identification and early warning decisions can be initiated. The delay introduced by data transmission and centralized computing is difficult to meet the needs of application scenarios with extremely high real-time requirements. For example, a delay of several seconds in earthquake early warning may cause response measures to fail to take effect in a timely manner.

[0004] Second, the system lacks reliability: In a centralized architecture, the central server is the single point of failure in the entire system. Once the central server fails due to hardware failure, communication interruption, or attack, the entire early warning system will lose its early warning and decision-making capabilities, and all terminal devices will be unable to obtain early warning information.

[0005] Third, terminal devices lack local decision-making capabilities: Under the existing architecture, terminal devices only execute early warning instructions from the central server and lack the ability to make autonomous decisions based on their own location, movement status and environmental information, thus failing to achieve targeted and differentiated responses.

[0006] Fourth, the false alarm rate and false negative rate are relatively high: Traditional systems often use single-point threshold triggering or simple multi-point threshold voting to determine events, failing to fully utilize the inherent correlation between multi-source event data in the time and space dimensions, resulting in prominent false alarm and false negative problems.

[0007] Fifth, the lack of a dynamic grading mechanism: the warning level of the existing system is usually determined once after the event is identified. It lacks the ability to dynamically adjust the risk level based on the new data added during the event's evolution and cannot reflect the real-time development of the disaster event.

[0008] Therefore, a new disaster early warning system architecture is urgently needed to overcome the inherent defects of the above-mentioned centralized architecture and achieve low latency, high reliability and distributed intelligent decision-making. Summary of the Invention

[0009] The technical problem to be solved by this invention is: how to achieve rapid identification of disaster events, multi-source data fusion verification, dynamic risk classification, and local autonomous decision-making control of terminals through a distributed architecture without relying on a central server for centralized decision-making, thereby reducing early warning response delay, improving system reliability, and reducing false alarm rate.

[0010] To address the aforementioned technical problems, this invention provides a distributed disaster event-driven dynamic hierarchical early warning and local decision-making method. In this method, each user terminal acts as an independent decision-making center, receiving event data broadcast from multiple monitoring nodes and independently performing spatiotemporal consistency analysis, confidence calculation, risk classification, and control strategy retrieval locally, without relying on a central node for unified judgment or unified issuance of control commands. The method includes the following steps: Step S1: Monitoring Node Event Generation: Multiple monitoring nodes are set up in the monitoring network. Each monitoring node includes a sensing unit, a data processing unit, and an event identification unit. The sensing unit is used to collect environmental physical quantity data in real time. The data processing unit performs preprocessing such as filtering and feature extraction on the collected data. The event identification unit determines whether an abnormal event has occurred according to a preset triggering algorithm. When a monitoring node detects an abnormal signal that meets the preset triggering conditions, it generates an event data packet. The event data packet includes: a location identifier, used to uniquely identify the monitoring node; a timestamp, recording the precise time of event triggering; an event type, identifying the category of the disaster event; initial characteristic parameters, recording the characteristic quantities of the abnormal signal; and a data quality marker, reflecting the reliability of the event data. The event data packet adopts a minimum necessary information structure, that is, it only contains information fields necessary for subsequent analysis by the user terminal and does not contain raw sensor waveform data, so as to reduce communication bandwidth requirements and improve transmission efficiency.

[0011] Step S2, Event Broadcast: The monitoring node broadcasts event data packets to multiple user terminals through the communication network. The broadcast transmission adopts a priority control strategy, setting the transmission priority according to the event type and severity to ensure that critical data is transmitted first. At the same time, a collision avoidance mechanism is adopted. When multiple monitoring nodes generate event data packets at the same time, they are sent sequentially according to the priority order of each monitoring node using a time-sharing transmission strategy to avoid communication channel congestion.

[0012] Step S3, Multi-source event reception and caching: The user terminal receives event data packets sent by multiple monitoring nodes and builds an event set locally according to the time sequence. The event set stores event data packets from different monitoring nodes. At the same time, a caching time window is set. Historical event data packets that exceed the time window will be automatically cleared to control local storage overhead.

[0013] Step S4, Spatiotemporal Consistency Analysis: The user terminal performs spatiotemporal consistency analysis on the event set, which includes the following three aspects.

[0014] (1) Time Consistency Judgment: Determine whether the timestamps of multiple event data packets in the event set fall within the time window determined based on the disaster propagation speed model. Specifically, let the event triggering time of the i-th monitoring node be... The distance from the disaster source to this node is If the propagation speed of the disaster wave is v, then the theoretical arrival time is... The calculation formula is: (II) In the formula, For the time of a disaster, events at that node are considered consistent in the time dimension if the following conditions are met: (III) In the formula, This is the preset time tolerance.

[0015] (2) Spatial propagation consistency judgment: Based on the spatial coordinates of each monitoring node and the corresponding event arrival time, verify whether the event arrival sequence conforms to the preset disaster propagation speed range. Specifically, for any two triggering nodes i and j, calculate their spatial distance. and time difference Determine whether the following conditions are true: (IV) In the formula, , Let be the event trigger time of the j-th monitoring node. and These represent the lower and upper limits of the range of disaster spread speed, respectively.

[0016] (3) Event expansion trend analysis: Within the preset time observation window, determine the spatial distribution relationship of the newly added response monitoring nodes and calculate the rate of change of the number of newly added response nodes. ,when When it is determined that the disaster event is expanding, When it is determined that the disaster event is trending towards stability, The time is used to determine the attenuation of disaster events, among which, This represents the total number of changes in the number of newly added response nodes. The time interval or span during which the number of nodes changes.

[0017] Step S5, Confidence Assessment: Calculate the confidence level of the disaster event based on multi-source event data. The calculation formula is as follows: The meanings of each parameter in the formula are as follows: The normalized value of the number of response nodes is calculated using the following formula: (V) In the formula, This represents the actual number of responding nodes. This represents the number of nodes expected to respond based on the disaster model. The time distribution consistency score reflects the degree of deviation between the trigger time and the theoretical arrival time of each event. The calculation formula is as follows: (VI) In the formula, the summation range is all response nodes. The smaller the deviation of each node, the better. The closer to 1; The spatial distribution consistency score reflects whether the spatial distribution of response nodes conforms to the disaster diffusion pattern. Specifically, the theoretical distance between each response node and the disaster source is calculated based on the event trigger time of each node and compared with the actual distance. The calculation formula is as follows: (VIII) In the formula, Let i be the actual distance from the i-th response node to the disaster source. The theoretical distance is calculated by working backward from the event trigger time of this node. Where v is the signal propagation speed, For signal arrival time, The signal generation time; To determine the maximum monitoring range of the network, the summation range includes all responding nodes; the smaller the deviation between the actual distance and the theoretical distance of each node, the better. The closer it is to 1.

[0018] The propagation pattern matching score reflects the degree of matching between the event arrival time sequence and the disaster propagation speed model. Specifically, the apparent propagation speed is calculated for each node pair consisting of all response nodes, and the proportion of node pairs falling within the preset propagation speed range is statistically analyzed. The calculation formula is as follows: In the formula, The total number of node pairs consisting of all responding nodes. - 1) / 2; To meet The number of node pairs under the condition, i.e. the number of node pairs whose apparent propagation speed falls within the preset disaster propagation speed range; , , , For the corresponding weight coefficients, satisfying It is obtained by pre-setting based on specific application scenarios or by training through historical data.

[0019] This is the correction amount for the outer perimeter expansion. As the disaster event continues to expand outwards and new outer perimeter nodes respond, To increase confidence, positive values ​​are used. This is the correction amount for intensity changes. It is taken as a positive value when the intensity of the disaster signal shows an increasing trend and a negative value when it shows a decreasing trend. This is the core area anomaly correction value, which is positive when monitoring nodes in the core area exhibit abnormal behavior. The abnormal behavior includes signal interruption, data anomaly, data loss, or node disconnection. The abnormal node characteristics are only used as auxiliary enhancement criteria and are not used as the sole basis for triggering disaster determination.

[0020] Step S6, Dynamic Classification: Based on the confidence level C of the disaster event and its changing trend, the disaster event is divided into four risk levels: Attention Level (… ), warning level ( ), emergency level ( ) and critical level ( ),in , , The preset level threshold is used; during the event evolution, when the confidence level rises to the threshold of the adjacent higher level, the level is upgraded; when the confidence level remains within the continuous observation period... When the risk level falls below the threshold corresponding to the lower level, it is downgraded, thereby achieving dynamic adjustment of the risk level.

[0021] Step S7, Local Decision-Making: The user terminal retrieves the corresponding control strategy from the preset risk-disposal database based on the event spatial location, the current location of the device, and the predicted location of the device. The predicted location of the device is predicted based on the current motion state of the controlled device and a kinematic model, using the following formula: (VII) In the formula, The predicted location of the controlled equipment. The current location of the controlled device. The current speed of the controlled device. The estimated remaining time for the disaster wave to reach the location of the controlled equipment; the risk-response database is a pre-established spatial risk mapping database, indexed by risk level and spatial location area, storing the corresponding control strategies; when there is no control strategy in the database that matches the current risk level and location, the default highest protection level control strategy is executed.

[0022] Step S8, Control Execution: After the user terminal independently completes the event judgment and control decision, it outputs control commands to the controlled equipment according to the retrieved control strategy. The control commands implement graded responses according to the risk level: the attention level corresponds to information notification, the warning level corresponds to speed limit or deceleration, the emergency level corresponds to emergency braking or shutdown, and the critical level corresponds to emergency stop of the entire system and personnel evacuation.

[0023] This invention also provides a distributed disaster event-driven dynamic hierarchical early warning and local decision-making network system, comprising: The monitoring node module, located in the monitoring network, includes a sensing unit, a data processing unit, and an event recognition unit, and is used to detect abnormal signals and generate event data packets.

[0024] The event broadcast module, connected to the monitoring node module, is used to broadcast event data packets to multiple user terminals via the communication network using a priority control strategy.

[0025] The user terminal module is used to receive event data packets sent by multiple monitoring node modules and to build an event set locally; The user terminal module is configured as an independent local decision-making center, which does not rely on an external central server and can independently complete event judgment and control decisions.

[0026] The spatiotemporal analysis module, located within the user terminal module, is used to perform temporal consistency judgment, spatial propagation consistency judgment, and event expansion trend analysis on a set of events.

[0027] The confidence module, located within the user terminal module, is used to calculate the confidence level of disaster events based on multi-source event data.

[0028] The hierarchical decision-making module, located within the user terminal module, is used to independently determine the risk level based on the confidence level of the disaster event and its changing trend, and to match and retrieve control strategies from a pre-established spatial risk mapping database based on the spatial location of the event, the current location of the equipment, and the predicted location of the equipment.

[0029] The local execution module connects to the controlled device and is used to output control commands to the controlled device according to the control strategy to achieve hierarchical response control.

[0030] Compared with the prior art, the present invention has the following beneficial effects: 1. Achieve a decentralized early warning architecture: In this invention, event identification is completed locally at the monitoring node, and spatiotemporal analysis and decision-making are completed locally at the user terminal. Each user terminal acts as an independent decision center, independently completing event judgment and control decisions, without relying on a central server for centralized processing or unified issuance of control commands. This eliminates the risk of single point of failure and significantly improves the overall reliability of the system.

[0031] 2. Reduced early warning response delay: Monitoring nodes directly broadcast event data packets to user terminals without the need for processing through a central server, significantly reducing the time overhead of data transmission and centralized computing; at the same time, the event data packets adopt the minimum necessary information structure, further reducing communication bandwidth requirements and transmission delay.

[0032] 3. Improve the accuracy of event judgment: By analyzing the spatiotemporal consistency of multi-source events, events are cross-validated from three dimensions: temporal consistency, spatial propagation consistency, and expansion trend. This effectively reduces the false alarm rate caused by single-point interference and the false alarm rate caused by insufficient information.

[0033] 4. Supports dynamic adjustment of risk levels: Based on real-time calculation and trend analysis of confidence levels, the risk level can be dynamically upgraded and downgraded, enabling early warning information to truly reflect the evolution of disaster events and providing a more accurate basis for decision-making.

[0034] 5. Enable local autonomous decision-making at the terminal: The user terminal makes local decisions based on its own location information and device movement status, and can implement differentiated and targeted control strategies to meet the personalized response needs of different locations and different device types. Attached Figure Description

[0035] Figure 1 This is a flowchart illustrating a distributed disaster event-driven dynamic hierarchical early warning and local decision-making method according to the present invention.

[0036] Figure 2 This is a structural block diagram of a distributed disaster event-driven dynamic hierarchical early warning and local decision-making system according to the present invention.

[0037] Figure 3 This is a schematic diagram illustrating the spatiotemporal consistency analysis of a distributed disaster event-driven dynamic hierarchical early warning and local decision-making method according to the present invention.

[0038] Figure 4 This is a schematic diagram illustrating the confidence calculation and dynamic grading principle of a distributed disaster event-driven dynamic grading early warning and local decision-making method according to the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are for illustrative purposes only and do not limit the scope of protection of this invention.

[0040] This embodiment uses earthquake early warning as an application scenario to provide a detailed description of the technical solution of the present invention. See also... Figure 1 and Figure 2 The distributed disaster event-driven dynamic hierarchical early warning and local decision-making method and system of this embodiment includes the following specific implementation methods.

[0041] I. Monitoring Network Deployment and Event Generation: Multiple earthquake monitoring nodes are deployed in earthquake-prone areas. Each monitoring node includes an earthquake sensor (sensing unit), an embedded data processor (data processing unit), and an event determination module (event recognition unit). The earthquake sensor uses an accelerometer to collect triaxial acceleration data in real time, with a sampling frequency of 200Hz.

[0042] The data processing unit performs bandpass filtering (passband frequency 0.1Hz~40Hz) on the acquired acceleration data and extracts P-wave features. The event recognition unit uses the STA / LTA (short-time average / long-time average) triggering algorithm to detect earthquake events (STA / LTA triggering algorithm is a widely used event detection method in the field of earthquake monitoring). The short-time window is 0.5s, the long-time window is 10s, and the trigger ratio threshold is set to 3.0. When the STA / LTA ratio exceeds the trigger threshold, it is determined that an earthquake event has been triggered, and an event data packet is generated.

[0043] The specific fields of the event data packet are as follows: the location identifier is the unique number of the monitoring node (e.g., "STA-0035"); the timestamp uses UTC time accurate to milliseconds (e.g., "2025-06-15T08:23:45.123Z"); the event type identifier is "EARTHQUAKE"; the initial motion characteristic parameters include the P-wave arrival time, P-wave initial motion amplitude, and initial motion period; the data quality label is automatically assigned according to the signal-to-noise ratio (SNR), with an SNR greater than 10dB marked as "high", 5~10dB marked as "medium", and less than 5dB marked as "low".

[0044] II. Event Broadcasting and Reception: The monitoring nodes broadcast event data packets to user terminals within the coverage area via 4G / 5G wireless communication networks or dedicated data links. The broadcast adopts a priority control strategy: event data packets with the event type "EARTHQUAKE" and initial amplitude exceeding the threshold have the highest priority and are sent first. When multiple monitoring nodes trigger simultaneously within a 100ms time window, they are sent in a time-division manner according to the order of distance between each node and the estimated location of the earthquake source from near to far, with a transmission time slot interval of 5ms for each node.

[0045] In this embodiment, the user terminal is a train-mounted earthquake early warning control device (a standard terminal device in the high-speed railway earthquake early warning system, deployed in the early warning system of the Japanese Shinkansen and the earthquake monitoring and early warning system along the Chinese high-speed railway; essentially, it is an embedded computing and communication terminal installed on the train, whose core function is to receive early warning information sent by earthquake monitoring stations along the line, and, in combination with the train's own position, speed and other operating status, execute control actions such as braking or speed limiting; the hardware typically includes a communication receiving module, an embedded processor, and an interface module with the train control system). This device receives event data packets broadcast by multiple monitoring nodes, stores them in the event cache unit according to the receiving time order, constructs a local event set, and sets the event cache time window to 120 seconds, automatically clearing historical event data packets that exceed 120 seconds.

[0046] III. Spatiotemporal Consistency Analysis: See Figure 3 The user terminal performs spatiotemporal consistency analysis on the event set, as detailed below.

[0047] (1) Time consistency judgment: The average propagation speed of earthquake P waves in the Earth's crust is approximately v = 6 km / s. Let the distance of the i-th monitoring node from the source be v. The theoretical arrival time of the P wave is calculated according to formula (II). The actual trigger time Compared with the theoretical arrival time, when the conditions in formula (III) are met... When the events of a node are consistent in time, it is determined that the events of that node are consistent in time.

[0048] (2) Spatial propagation consistency judgment: For any two triggering nodes i and j, calculate their spatial distance. and time difference According to formula (IV), when When the event propagation falls within the P-wave propagation speed range [5km / s, 8km / s], the event propagation of the pair of nodes is determined to be consistent. The proportion of consistent propagation among all node pairs is statistically analyzed and used as an evaluation index for spatial propagation consistency.

[0049] (3) Event expansion trend analysis: Set a 5-second time observation window, and count the number and spatial distribution of newly added response nodes in each window. If the newly added nodes show a distribution trend that conforms to the disaster expansion law in spatial location (e.g., along the direction of seismic wave propagation), and the rate of change of the number of newly added nodes is greater than zero, then the earthquake event is judged to be expanding, which is conducive to improving the confidence of event judgment.

[0050] IV. Confidence Calculation and Dynamic Grading: See Figure 4 Based on the results of the spatiotemporal consistency analysis, the user terminal calculates the confidence level C of the disaster event according to formula (I). In this embodiment, the weighting coefficient is taken as: , , , .

[0051] The specific calculation process is as follows: Normalized value of the number of response nodes Assuming that 12 monitoring nodes are expected to respond within 5 seconds of an earthquake ( ), actually received event data packets from 8 nodes ( According to formula (V), we get .

[0052] Consistency score of time distribution The absolute deviations between the actual trigger times and the theoretical arrival times of the eight response nodes were 0.3s, 0.5s, 0.8s, 0.2s, 1.1s, 0.4s, 0.6s, and 0.9s, respectively, with an average deviation of 0.6s. According to formula (VI), .

[0053] Spatial distribution consistency score According to formula (VIII), the deviation between the actual distance and the theoretical distance of each response node is used for calculation. In this embodiment, the maximum monitoring range distance of the monitoring network is... The actual distance between the 8 response nodes The theoretical distance calculated from the trigger time The absolute values ​​of the deviations were 2.1km, 3.5km, 5.2km, 1.8km, 7.0km, 2.6km, 4.2km, and 6.1km, respectively, with an average deviation of 4.06km. 4.06 / 200 = 0.980. Considering that the deviation direction of some nodes does not conform to the theoretical diffusion direction, after correction for directional consistency, we take... .

[0054] Propagation pattern matching score According to formula (IX), the 8 response nodes constitute a total of For each pair of nodes, calculate the apparent propagation velocity, where... If the propagation speed of one node pair falls within the P-wave velocity range [5km / s, 8km / s] (the other 6 node pairs have larger time stamp deviations due to signal interference), then... ,Pick .

[0055] Peripheral expansion correction Two new peripheral nodes responded within the observation window. Intensity change correction amount The intensity of the seismic signal is showing an increasing trend. Core area anomaly correction amount No abnormal behavior was observed at the core area monitoring nodes, including signal interruption, data anomalies, data loss, or node disconnection. .

[0056] Substituting into formula (I), we get: C = 0.30×0.667 + 0.25×0.700 + 0.25×0.750 + 0.20×0.800 + 0.05 +0.03 + 0 = 0.200 + 0.175 + 0.188 + 0.160 + 0.05 + 0.03 = 0.803 The dynamic grading threshold is set as follows: (Attention Level) (Warning level) (Emergency level), due to The current risk level is classified as emergency.

[0057] If the built-in confidence level continues to rise above 0.90 within the next 10 seconds, it will be dynamically upgraded to a critical level. If the confidence level remains above 0.90 during the continuous observation period... If the index falls below 0.60, it will be downgraded to a warning level.

[0058] V. Local Decision-Making and Control Execution: The user terminal (train-mounted early warning control device) makes local decisions based on the risk level and equipment location. Let the current position of the train be... The current driving speed is (Approximately 83.3 m / s), the estimated remaining time for the seismic wave to reach the train's location is... According to formula (VII), we get: The terminal retrieves the control strategy corresponding to the "emergency level" risk level and the predicted location from the risk-disposal database. If the predicted location is in the bridge section, the corresponding control strategy is "emergency braking and stopping in a safe area"; if the predicted location is in the general roadbed section, the corresponding control strategy is "emergency speed reduction to the speed limit".

[0059] Taking the train scenario in the embodiment as an example, the data structure of the risk-handling database is as follows: Table 1: Risk-Reaction Control Strategy Table The terminal outputs control commands to the train control system based on the retrieved control strategy to achieve graded response: Attention level: display warning information to the driver without interfering with train operation; Warning level: automatically limit speed to 200km / h and issue an alarm to the driver; Emergency level: emergency braking or speed limit to 80km / h; Critical level: emergency stop of the entire train and trigger passenger evacuation guidance.

[0060] The above embodiments are merely preferred embodiments of the present invention, and the scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principle of the present invention should also be considered within the scope of protection of the present invention. For example, applying the present invention to disaster scenarios with similar wave propagation characteristics, such as tsunami warning, explosion shock wave monitoring, and geological disaster warning along rail transit lines, only requires replacing the sensor type and the corresponding propagation velocity parameters; the essence of the technical solution is the same as this embodiment, and all should be considered within the scope of protection of the present invention.

Claims

1. A distributed disaster event-driven dynamic hierarchical early warning and local decision-making method, characterized in that, Each user terminal acts as an independent decision-making center, receiving event data broadcast from multiple monitoring nodes and independently performing spatiotemporal consistency analysis, confidence calculation, risk classification, and control strategy retrieval locally, without relying on a central node for unified judgment or unified issuance of control commands; the method includes the following steps: Step S1, Monitoring Node Event Generation: Multiple monitoring nodes are set up in the monitoring network. Each monitoring node includes a sensing unit, a data processing unit, and an event recognition unit. When the monitoring node detects an abnormal signal that meets the preset triggering conditions, it generates an event data packet. The event data packet includes a location identifier, a timestamp, an event type, initial characteristic parameters, and a data quality marker. Step S2, Event Broadcast: The monitoring node broadcasts the event data packet to multiple user terminals through the communication network. The broadcasting adopts a priority control strategy, giving priority to the transmission of key data. Step S3, Multi-source event reception and caching: The user terminal receives event data packets sent by multiple monitoring nodes and constructs an event set locally; Step S4, Spatiotemporal Consistency Analysis: The user terminal performs spatiotemporal consistency analysis on the event set, including time consistency judgment, spatial propagation consistency judgment, and event expansion trend analysis; Step S5, Confidence Assessment: Calculate the confidence level of the disaster event based on the multi-source event data in the event set. The factors for calculating the confidence level include the number of response nodes, the event time distribution, the event spatial distribution, and the degree of matching of the propagation pattern. Step S6, Dynamic Classification: Based on the confidence level of the disaster event and its changing trend, the disaster event is divided into multiple risk levels, and the risk levels are dynamically adjusted according to the real-time changes in the confidence level. Step S7, Local Decision: The user terminal retrieves the corresponding control strategy from the preset risk-disposal database based on the event spatial location, the current location of the device, and the predicted location of the device; Step S8, Control Execution: The user terminal outputs control commands to the controlled device according to the control strategy to achieve hierarchical response control.

2. The method according to claim 1, characterized in that, The event data packet adopts a minimum necessary information structure, which means that the event data packet only contains the fields necessary for the user terminal to perform spatiotemporal consistency analysis and confidence assessment, and does not contain the original sensor waveform data.

3. The method according to claim 1, characterized in that, In step S2, the broadcast transmission also employs a conflict avoidance mechanism, which includes: when multiple monitoring nodes generate event data packets simultaneously, a time-division sending strategy is adopted to send the event data packets sequentially according to the priority order of each monitoring node.

4. The method according to claim 1, characterized in that, In step S4, the time consistency judgment includes: judging whether the timestamps of multiple event data packets in the event set fall within the time window determined based on the disaster propagation speed model; the spatial propagation consistency judgment includes: verifying whether the event arrival sequence conforms to the preset disaster propagation speed range based on the spatial coordinates of each monitoring node and the corresponding event arrival time.

5. The method according to claim 1, characterized in that, In step S4, the event expansion trend analysis includes: within a preset time observation window, determining the spatial distribution relationship of newly added response monitoring nodes, and judging the expansion trend of the disaster event based on the spatial expansion trend of the newly added nodes and the rate of change of the number of newly added response nodes.

6. The method according to claim 1, characterized in that, In step S5, the confidence assessment further includes confidence correction based on the following enhancement factors: the response expansion status of peripheral monitoring nodes, the changing trend of disaster signal intensity, and the degree of anomaly of core area monitoring nodes; the formula for calculating the confidence level C is: in, To respond to the normalized value of the number of nodes, For consistency score of time distribution, The spatial distribution consistency score is used to determine the spatial distribution consistency score. The score is based on the degree of matching the propagation pattern. , , , The corresponding weight coefficients and ; For the peripheral expansion correction amount, This is the correction amount for intensity variation. This is the core area anomaly correction value, which is positive when monitoring nodes in the core area exhibit abnormal behavior. The abnormal behavior includes signal interruption, data anomaly, data loss, or node disconnection. The abnormal node characteristics are only used as auxiliary enhancement criteria and are not used as the sole basis for triggering disaster determination.

7. The method according to claim 1, characterized in that, In step S6, the multiple risk levels include: attention level, warning level, emergency level and critical level; when the confidence level rises to the threshold of the adjacent higher level, the risk level is upgraded; when the confidence level falls back to below the threshold of the lower level during the continuous observation period, the risk level is downgraded.

8. The method according to claim 1, characterized in that, In step S7, the device predicted location is based on the current motion state of the controlled device and a kinematic model to predict the location of the controlled device when the disaster wave arrives; the local decision-making also includes: when there is no control strategy matching the current risk level and location in the risk-disposal database, the default highest protection level control strategy is executed; The risk-management database is a pre-established spatial risk mapping database. The spatial risk mapping database uses the spatial location region and risk level of the disaster event as an index to store the control strategies corresponding to each spatial region. The user terminal performs matching and retrieval based on the spatial location of the disaster event, the current location of the equipment, and the predicted location of the equipment to obtain the control strategy corresponding to the current risk status.

9. A system applying the method of claim 1, characterized in that, include: The monitoring node module, set up in the monitoring network, includes a sensing unit, a data processing unit, and an event recognition unit. It is used to detect abnormal signals and generate event data packets containing point identifiers, timestamps, event types, initial characteristic parameters, and data quality markers. An event broadcasting module, connected to the monitoring node module, is used to broadcast the event data packets to multiple user terminals via a communication network using a priority control strategy. The user terminal module is used to receive event data packets sent by multiple monitoring node modules and to build an event set locally; The user terminal module is configured as an independent local decision-making center, which does not rely on an external central server and can independently complete event judgment and control decisions. The spatiotemporal analysis module is located within the user terminal module and is used to perform temporal consistency judgment, spatial propagation consistency judgment, and event expansion trend analysis on the event set. The confidence module, located within the user terminal module, is used to calculate the confidence level of a disaster event based on the number of response nodes, the event time distribution, the event spatial distribution, and the matching degree of the propagation pattern. The hierarchical decision-making module, located within the user terminal module, is used to independently determine the risk level based on the confidence level and changing trend of the disaster event, and to match and retrieve control strategies from a pre-established spatial risk mapping database based on the spatial location of the event, the current location of the equipment, and the predicted location of the equipment. The local execution module is connected to the controlled device and is used to output control commands to the controlled device according to the control strategy to achieve hierarchical response control.

10. The system according to claim 9, characterized in that, The user terminal module further includes an event caching unit, which is used to store received event data packets in a time sequence and automatically clear expired data when the preset caching time window is exceeded; the event broadcasting module further includes a conflict avoidance unit, which is used to schedule the transmission in a time-sharing manner according to a preset priority when multiple monitoring nodes trigger at the same time.

Images