A seismic early warning analysis system and method based on intelligent devices

Abstract

The application relates to the technical field of earthquake monitoring and early warning, in particular to an earthquake early warning analysis system and method based on intelligent devices. The system comprises a terminal sensing and preliminary judgment module, a self-adaptive network transmission module, an edge aggregation and verification module and a cloud control fusion and calculation module. The terminal sensing and preliminary judgment module collects data through a sensor array, identifies P wave characteristics by using a dynamic weight adjustment model and generates an encrypted preliminary warning. The self-adaptive network transmission module establishes a dynamic self-organizing network based on a D2D protocol and realizes multi-hop relay transmission of the warning by electing a relay node. The edge aggregation and verification module performs space-time matching and cross verification on multiple warnings in a specific area, judges the authenticity of the event and generates graded early warning information. The cloud control fusion and calculation module fuses regional event data and professional network data, performs high-precision positioning and magnitude calculation and issues final early warning instructions. The application constructs a high-density sensing network by using a large number of intelligent devices, effectively eliminates the early warning blind area and realizes faster, more accurate and more robust earthquake early warning.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of earthquake monitoring and early warning, in particular, to an earthquake early warning analysis system and method based on intelligent devices. BACKGROUND

[0002] The earthquake early warning analysis system is a technical system that publishes early warning information to users before destructive seismic waves reach the target area by real-time monitoring of seismic wave propagation and rapid calculation of source parameters. The essence of the system is not to predict whether an earthquake will occur in advance, but to "monitor the earthquake that has occurred and warn before the destructive wave arrives", which can help the target area to take emergency measures in advance and reduce personnel casualties and property losses.

[0003] The current mainstream earthquake early warning analysis technology can be divided into two core methods according to the differences in "data source" and "analysis speed": the rapid early warning method based on "single station": using the P-wave initial motion signal received by a single seismic station to rapidly estimate the source parameters. Only relying on single station data, without waiting for multi-station data to be collected, the analysis speed is extremely fast, and it is suitable for monitoring "near earthquakes".

[0004] The precise early warning method based on "multi-station networking": collecting P-wave initial motion signals from multiple seismic stations, determining the epicenter location by "triangulation method", and correcting the magnitude by combining the amplitude data of multiple stations, which has higher analysis accuracy, but needs to wait for data transmission and collection, and is suitable for monitoring "distant earthquakes" or "strong earthquakes".

[0005] The existing technology has the following defects, which are specifically embodied in:

[0006] 1. There is a fatal blind area in early warning: the existing earthquake early warning system completely relies on a limited number of professional seismic monitoring stations at fixed locations. This sparse layout results in the epicenter area, which is most in need of early warning, being unable to obtain any effective early warning time before the arrival of destructive S waves due to being in the "blank" center of the station network, forming an unavoidable early warning blind area, which makes the system ineffective in the most critical area.

[0007] 2. The system communication is extremely fragile and dependent: the data transmission of the traditional system is severely dependent on centralized public communication networks (such as cellular networks and the Internet). However, strong earthquakes can easily damage the infrastructure of these networks (such as optical cables and base stations), resulting in the entire early warning system being paralyzed due to "information islands" at critical moments, and the early warning information cannot be collected and distributed, posing a serious challenge to the survivability and reliability of the system.

[0008] 3. Insufficient reliability and accuracy of early warning: due to single and limited data sources, the existing system has poor filtering capability for local interference (such as large vehicles and mechanical vibration), resulting in high risk of false reporting; at the same time, for complex source rupture process, only a small amount of P-wave data at the initial stage is easy to underestimate the magnitude of a large earthquake (i.e. "magnitude saturation" problem), causing inaccurate warning range and seriously affecting the actual effectiveness and credibility of the warning. SUMMARY

[0009] The purpose of the present application is to provide a seismic early warning analysis system and method based on intelligent devices to solve the problems raised in the background art.

[0010] To achieve the above-mentioned purpose, the present application aims to provide a seismic early warning analysis system based on intelligent devices, comprising: a terminal perception and preliminary judgment module: for continuously collecting local vibration data and environmental data through a sensor array, and performing real-time fusion and processing on the data, identifying the characteristics of the P-wave through a dynamic weight adjustment model, and generating and outputting an encrypted preliminary warning containing confidence and characteristic parameters.

[0011] An adaptive network transmission module: connected with the terminal perception and preliminary judgment module, based on D2D communication protocol, for automatically forming a dynamic self-organizing network among intelligent devices, forming a multi-hop transmission link through the election of relay nodes, and differentially encoding, aggregating and encrypting the preliminary warning for relay transmission.

[0012] An edge aggregation and verification module: connected with the adaptive network transmission module, for receiving a plurality of preliminary warnings from a specific geographic area within a preset time window, executing a spatio-temporal matching and multi-node cross-verification algorithm, and determining whether a regional seismic event is established, if the regional seismic event is established, immediately generating a hierarchical warning information containing the estimated magnitude and the range of the earthquake, and issuing it to the outside of the threatened area.

[0013] A cloud control fusion and calculation module: in communication with the edge aggregation and verification module, for gathering and fusing regional event data and professional seismic network data, performing high-precision seismic positioning and magnitude calculation based on multi-source information, determining the final three elements of the earthquake, and issuing precise warning instructions and linkage control signals to the public emergency system.

[0014] As a further improvement of the present technical solution, the dynamic weight adjustment model is implemented as follows: extracting the data collected by the sensor array, dividing the data collected by the sensor array into core perception dimensions and auxiliary correction dimensions, wherein the core perception dimensions include vibration waveform data and device posture data, and the auxiliary correction dimensions include environmental noise spectrum data and environmental air pressure fluctuation data.

[0015] The historical earthquake sample data is obtained from the database, and initial weight proportions higher than those of the auxiliary correction dimensions are assigned to the core perception dimensions based on offline training of the historical earthquake sample data, wherein the vibration waveform data is assigned the highest initial weight as the core basis for P-wave identification.

[0016] The weight optimization is realized by using the reinforcement learning framework, the environmental noise level, the device historical false alarm type and the sensor data quality are taken as the state space, the preset step increase and decrease of the feature weight of each dimension is taken as the action space, the P-wave identification accuracy improvement is taken as the positive reward, the false alarm and the missed alarm are taken as the negative punishment, the weight matrix is iteratively updated by using the time difference algorithm, the weight matrix is stored in low-precision quantization, only the feature channel with a contribution degree to the P-wave identification higher than a preset contribution degree threshold is reserved, and a dynamic weight adjustment model is obtained.

[0017] As a further improvement of the technical solution, the automatic formation of a dynamic self-organizing network among intelligent devices is realized by the following method: after the intelligent device is started, a detection frame containing the device unique identification and hardware feature code is broadcasted through the D2D communication protocol, the adjacent intelligent device receives the detection frame, verifies the legality of the sender identification, and returns a response frame containing its own identification and communication capability if the verification is passed, and a temporary communication link is established between the two parties.

[0018] Based on the signal strength parameter and the relative position information in the response frame, each intelligent device constructs a priority list of adjacent devices, forms a star-shaped subnetwork with the core device as the center, and realizes the interconnection of the whole network through the cross-connected intelligent devices.

[0019] The intelligent device periodically sends a state heartbeat packet to the connected node, which contains the current power, the remaining bandwidth and the position change amount, and if the receiving node does not receive the heartbeat packet continuously and the signal strength is below the threshold, it is determined that the connection is invalid, and the next intelligent device is selected from the priority list to rebuild the connection.

[0020] When the intelligent device detects that its own position moves more than a preset distance, it actively sends position update information to the original connection node and broadcasts a new detection frame to discover new intelligent devices around, predicts the possible disconnection based on the moving speed and direction, and establishes a preliminary connection with the newly discovered intelligent device in advance.

[0021] The relay forwarding amount of each node is counted in real time, and when the forwarding amount of a certain node exceeds a preset forwarding amount threshold, it automatically sends a load balancing request to the adjacent node, and the adjacent node responds to the request according to its own remaining resources to share part of the relay task.

[0022] As a further improvement of the technical solution, the elected relay node forms a multi-hop transmission link, and the specific implementation method is: after the intelligent device accesses the self-organizing network, it automatically detects its hardware capability and running state, and the device that meets the preset basic condition actively broadcasts candidate nomination information containing its capability parameters to the network.

[0023] After each intelligent device in the network receives the candidate nomination information, it quantitatively scores the candidate node based on a preset evaluation model, and the evaluation dimensions include communication coverage, historical forwarding reliability, resource redundancy, and location stability.

[0024] Each intelligent device sends a vote confirmation to the candidate node with the highest priority based on the scoring results, and when the candidate node accumulates a preset proportion of the total number of devices in the local network, it automatically becomes a formal relay node and broadcasts the election declaration to the entire network.

[0025] The term of office of the relay node is set to a dynamic length, and the relay node continuously sends a status heartbeat to the surrounding intelligent devices during the term of office. Before the end of the term of office, the intelligent devices within the current coverage range reevaluate based on their performance during the term of office, and if they meet the conditions for re-election, they automatically extend the term of office, otherwise a new round of election is initiated.

[0026] When the relay node has a sudden failure or its performance drops below a threshold, the intelligent devices directly connected to it immediately initiate an interim election and quickly select a replacement node from the standby candidate nodes.

[0027] As a further improvement of the technical solution, the spatiotemporal matching and multi-node cross-validation algorithm is executed, and the specific implementation method is: multiple preliminary alarms received within a preset time window and in a specific geographic area are analyzed, and the P-wave arrival time, geographic location coordinates of the intelligent device, vibration characteristic parameters, and confidence value contained in each alarm are extracted.

[0028] The difference between the P-wave arrival time of the first valid alarm in the region and the P-wave arrival time of other alarms is calculated, the standard deviation of all difference values is calculated, and if the standard deviation is less than a preset time threshold, it is determined that the time series matches, otherwise, the alarms that exceed the time deviation threshold are marked as suspicious data.

[0029] Based on the geographic location coordinates of each intelligent device and the P-wave arrival time, the back projection algorithm is used to calculate the candidate area of the theoretical epicenter, the spatial distance between the actual position of each intelligent device and the candidate area of the theoretical epicenter is calculated, the average distance and the dispersion coefficient are calculated, and if the dispersion coefficient is less than a preset spatial threshold, it is determined that the spatial distribution matches.

[0030] For the alarms that pass the spatiotemporal matching, the cosine similarity of the vibration characteristic parameters and the peak value deviation rate are calculated, and the alarms with a similarity higher than a preset feature threshold and a peak value deviation rate lower than a preset deviation threshold are retained to form an effective sample set.

[0031] The confidence values of the effective sample set are weighted and summed, if the weighted result is higher than a preset confidence threshold value, and the proportion of the number of the effective sample set in the total number of the initial received alarms exceeds a preset number threshold value, then it is determined that the regional seismic event is established, otherwise, it is determined that it is not a seismic event.

[0032] As a further improvement of the technical solution, the hierarchical early warning information is specifically implemented as follows:

[0033] Based on the regional event data matched in space and time and cross-verified by multiple nodes, a finite fault model is used to calculate and determine the estimated epicenter position, magnitude and seismic wave propagation speed parameters, combined with preset regional geological structure data and site effect parameters, the estimated intensity values of different locations in the affected area are calculated through the ground motion prediction equation, based on the estimated intensity values, the affected area is divided into three early warning levels, and the corresponding early warning instruction text and response suggestions are matched for each early warning level.

[0034] As a further improvement of the technical solution, the high-precision seismic positioning based on multi-source information is specifically implemented as follows: the regional event data and professional seismic network data are time-synchronized and calibrated, and abnormal values are removed, effective data with a time stamp error within a preset range are retained, a multi-source data fusion positioning model is constructed, the intelligent device data and the network data are distributed with weights according to the spatial distribution density, the iterative weighted least squares method is used to solve the epicenter coordinates, the distance residual of each intelligent device to the theoretical epicenter is taken as the optimization target, and the residual sum of squares is reduced through multiple iterations to obtain high-precision epicenter coordinates.

[0035] As a further improvement of the technical solution, the magnitude calculation is specifically implemented as follows: based on the intelligent device data, the equivalent regional magnitude is converted by using the vibration peak value parameter and the device distribution radius through an empirical formula, based on the professional network data, the surface wave magnitude calculation method is used, combined with the waveform amplitude recorded by the station and the epicentral distance correction value to determine the network magnitude, the two types of magnitude results are weighted and fused, and a deviation correction factor of historical seismic events is introduced to generate the final magnitude result.

[0036] The second aspect of the application provides a method of a seismic early warning analysis system based on intelligent devices, comprising: S1, terminal perception and preliminary judgment: used for continuously collecting local vibration data and environmental data through a sensor array, and performing real-time fusion and processing on the data, identifying the seismic P wave characteristics through a dynamic weight adjustment model, and generating and outputting an encrypted preliminary alarm containing confidence and characteristic parameters.

[0037] S2, adaptive network transmission: based on D2D communication protocol, used to automatically form a dynamic self-organizing network between intelligent devices, and through the election of relay nodes to form a multi-hop transmission link, to perform differential encoding, aggregation and encryption relay transmission of the preliminary alarm.

[0038] S3, edge aggregation and verification: used to receive multiple preliminary alarms from a specific geographic area within a preset time window, perform spatio-temporal matching and multi-node cross-verification algorithm, determine whether the regional seismic event is established, and if the regional seismic event is established, generate a hierarchical warning information containing the estimated magnitude and range, and issue it to the outside of the threat area.

[0039] S4, cloud control fusion and calculation: used to gather and fuse regional event data and professional seismic network data, perform high-precision seismic positioning and magnitude calculation based on multi-source information, determine the final seismic three elements, and issue precise warning instructions and linkage control signals to the public emergency system.

[0040] Compared with the prior art, the beneficial effects of the present application are:

[0041] 1. The present application innovatively uses a large number of ordinary intelligent devices (such as smart phones) to build a high-density, widely-distributed seismic sensing network, which enables a large number of devices in the epicenter area to become the first batch of detectors, realizing "epicenter self-warning" and fundamentally eliminating the warning blind area of traditional technology, thereby maximizing the risk avoidance time for all regions outside the epicenter.

[0042] 2. A high-resilience and self-healing communication lifeline is constructed: by introducing dynamic self-organizing network technology based on D2D protocol, the system does not rely on any fixed communication infrastructure that may be damaged. Intelligent devices can automatically network and transmit data through multi-hop relay, ensuring that even in the most extreme disaster conditions, warning information can still be reliably transmitted from the epicenter like a "relay race", greatly enhancing the robustness and survivability of the entire system.

[0043] 3. The present application reduces the false alarm rate to a very low level through the dual intelligent filtering mechanism of "terminal intelligent identification (dynamic weight model)" and "edge group cross-verification", and at the same time, the cloud end overcomes the limitations of a single data source by fusing massive intelligent device data and professional network data, realizes the calculation of source parameters from rapid preliminary estimation to final high-precision determination, ensures that the warning information is both "fast" and "accurate", and significantly improves the overall reliability of the system. BRIEF DESCRIPTION OF DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0045] Figure 1 For the system structure connection diagram of the present application.

[0046] Figure 2 For the method implementation step flow diagram of the present application. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only constitute some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the protection scope of the present application.

[0048] Embodiment: Please refer to Figure 1 As shown, a seismic early warning analysis system based on intelligent devices is provided, comprising: a terminal perception and preliminary judgment module: for continuously collecting local vibration data and environmental data through a sensor array, and performing real-time fusion and processing on the data, identifying the characteristics of P wave through a dynamic weight adjustment model, and generating and outputting an encrypted preliminary warning containing confidence and characteristic parameters.

[0049] In one specific embodiment, the dynamic weight adjustment model is specifically implemented as follows: extracting the data collected by the sensor array, and dividing the data collected by the sensor array into core perception dimensions and auxiliary correction dimensions, wherein the core perception dimensions include vibration waveform data and device posture data, and the auxiliary correction dimensions include environmental noise spectrum data and environmental air pressure fluctuation data.

[0050] The accelerometer collects vibration waveform data (for extracting P wave time domain / frequency domain characteristics), the gyroscope collects device posture data (for eliminating non-seismic shaking interference of the device), the microphone collects environmental noise spectrum data (for distinguishing natural vibration from artificial noise), and the barometer collects environmental air pressure fluctuation data (for assisting in judging whether the air pressure is abnormally caused by geological activity).

[0051] Obtain historical earthquake sample data from the database, and train offline based on the historical earthquake sample data, to assign an initial weight ratio of the core perception dimensions higher than that of the auxiliary correction dimensions, wherein the vibration waveform data is assigned the highest initial weight as the core basis for P wave identification.

[0052] The vibration waveform data weight proportion is the highest (as the core basis for P wave recognition), the device posture data is the second (for filtering non-seismic interference), the environmental noise and air pressure data are assigned basic weights as correction terms, and the initial weight matrix is generated by offline training of historical seismic sample data.

[0053] The weight optimization is realized by using a reinforcement learning framework, the environmental noise level, device historical false alarm types and sensor data quality are used as the state space, the preset step increase and decrease of each dimension feature weight are used as the action space, the P wave recognition accuracy improvement is used as the positive reward, and the false alarm and missed alarm are used as the negative punishment, the weight matrix is iteratively updated by using a time difference algorithm, the weight matrix is stored at a low precision, only the feature channels with a contribution degree to P wave recognition exceeding a preset contribution degree threshold are reserved, and a dynamic weight adjustment model is obtained.

[0054] It should be noted that the preset step is set by a professional, for example, based on the importance difference of feature dimensions, the step of core perception dimensions (such as vibration waveform) is set to a small value for fine tuning, and the step of auxiliary correction dimensions is set to a large value for rapid adaptation; at the same time, the step is dynamically adjusted in the initial stage and the convergence stage of model training, the step is slightly larger in the initial stage to speed up iteration, and the step is reduced in the later stage to accurately converge.

[0055] The adaptive network transmission module is connected with the terminal perception and preliminary judgment module, is constructed based on a D2D communication protocol, and is used for automatically establishing a dynamic self-organizing network among intelligent devices, forming a multi-hop transmission link by electing a relay node, and differentially encoding, aggregating and encrypting relay transmission of the preliminary alarm.

[0056] In one specific embodiment, the dynamic self-organizing network among intelligent devices is automatically established, and the specific implementation method is as follows: after the intelligent device is started, a probe frame containing a device unique identifier and a hardware feature code is broadcasted through a D2D communication protocol, a neighboring intelligent device receives the probe frame, verifies the legality of the identifier of the sender (based on a pre-stored device white list or a temporary certificate), returns a response frame containing its own identifier and communication capability if the verification is passed, and a temporary communication link is established between the two parties.

[0057] Each intelligent device constructs a neighboring device priority list based on the signal strength parameter and relative position information in the response frame, forms a star-shaped subnetwork with a core device as the center, and realizes interconnection of the whole network through cross-connected intelligent devices.

[0058] The intelligent device periodically (the period is dynamically adjusted according to network load) sends a state heartbeat packet to the connected node, the state heartbeat packet contains the current power, residual bandwidth and position change amount, if the receiving node continuously does not receive the heartbeat packet and the signal strength is below a threshold, it is determined that the connection is invalid, and the next intelligent device is selected from the priority list to rebuild the connection.

[0059] It should be noted that the signal strength threshold is set by professionals, for example, based on communication quality requirements and differences in device performance, with the core objective of ensuring that the data transmission error rate is lower than the preset error rate threshold. The basic threshold is determined in conjunction with the hardware specifications of the smart device's communication module. At the same time, the threshold is dynamically adjusted in conjunction with the network load. Under high load, the threshold is appropriately increased to prioritize the transmission of critical data, and under low load, the threshold is decreased to expand the connection coverage.

[0060] When a smart device detects that its location has moved beyond a preset distance, it actively sends location update information to the original connection node and broadcasts a new detection frame to discover new smart devices in the vicinity. Based on the movement speed and direction, it predicts the connection that may be lost and establishes a preliminary connection with the newly discovered smart device in advance.

[0061] It should be noted that the preset distance is set by professionals, for example, with the core setting of ensuring connection stability and network adjustment efficiency, and the basic value is determined by combining the device's communication coverage radius and the distance data of connection failure caused by historical movement; at the same time, it is dynamically adapted to the movement speed, with a smaller value set for high-speed movement to trigger adjustment in advance, and a larger value set for low-speed movement to reduce unnecessary network overhead.

[0062] The relay forwarding volume of each node is counted in real time. When the forwarding volume of a node exceeds the preset forwarding volume threshold, it automatically sends a load balancing request to the neighboring nodes. The neighboring nodes respond to the request based on their remaining resources (power, bandwidth) and share part of the relay task.

[0063] It should be noted that the preset forwarding threshold is set by professionals. For example, it is set with the node's resource carrying capacity and network transmission stability as the core, and the basic threshold is determined by combining the node's remaining power, bandwidth limit and hardware forwarding performance to ensure that the threshold is lower than the node's overload threshold. At the same time, the overall network load is dynamically adjusted. When the network is under high load, the threshold is appropriately reduced to divert traffic in advance, and when the network is under low load, the threshold is increased to reduce the overhead of frequent load scheduling.

[0064] Through the four steps of “device discovery - topology initialization - dynamic adjustment - self-healing optimization”, the technical implementation of “automatic construction of dynamic self-organizing network” is fully defined. It is both dependent on the sovereign adaptive network transmission module and its dynamic characteristics that distinguish it from traditional fixed networks (such as automatic adaptation to device movement and real-time link repair).

[0065] In one specific embodiment, the election relay node forms a multi-hop transmission link. The specific implementation method is as follows: after the smart device accesses the self-organizing network, it automatically detects its own hardware capabilities (communication module power, data processing performance) and operating status (remaining power, current load). Devices that meet the preset basic conditions actively broadcast candidate nomination information containing their own capability parameters to the network.

[0066] After receiving the candidate nomination information, each intelligent device in the network quantitatively scores the candidate nodes based on a preset evaluation model, and the evaluation dimensions include communication coverage, historical forwarding reliability, resource redundancy, and location stability.

[0067] The dimensions include communication coverage, which is calculated based on signal strength attenuation curve, historical forwarding reliability, which is calculated based on success forwarding rate in the near period, resource redundancy, which is calculated based on remaining power and idle bandwidth proportion, and location stability, which is calculated based on displacement change in the near period.

[0068] According to the scoring results, each intelligent device sends a vote confirmation to the candidate node with the highest priority, and when the number of valid votes received by the candidate node reaches a preset proportion of the total number of local network devices, the candidate node automatically becomes the formal relay node and broadcasts the election declaration to the entire network.

[0069] The term of the relay node is set as a dynamic length (adjusted according to network load fluctuations), and the relay node continuously sends status heartbeat to surrounding intelligent devices during the term. Before the end of the term, the intelligent devices within the current coverage range reevaluate the performance (forwarding delay, packet loss rate) of the relay node during the term, and if the relay node meets the re-election conditions, the term is automatically extended, otherwise a new round of election is started.

[0070] When the relay node experiences sudden failure (such as power consumption, communication interruption) or performance drops below the threshold, the intelligent devices directly connected to the relay node immediately initiate a temporary election and quickly select a replacement node from the standby candidate nodes.

[0071] Edge aggregation and verification module: connected with the adaptive network transmission module, used to receive multiple preliminary alarms from a specific geographic area within a preset time window, execute spatio-temporal matching and multi-node cross-verification algorithm, determine whether the regional earthquake event is true, if the regional earthquake event is true, immediately generate hierarchical warning information containing estimated magnitude and impact range, and release it outside the threat area.

[0072] The "preset time window" is a fixed time interval set according to the propagation speed of seismic waves (especially P waves) on the ground, and the core purpose is to ensure that the received multiple preliminary alarms "originate from the same earthquake event", rather than independent interference at different times (such as accidental device shaking, traffic vibration). The "specific geographic area" is a circular / polygonal geographic range drawn by the edge server based on the "device location of the first preliminary alarm", and the core purpose is to ensure that the received preliminary alarms "come from the core area affected by the earthquake", and to exclude false alarms from distant irrelevant devices. "Multiple preliminary alarms" means that at least a preset number (such as 3 or more) of preliminary alarms need to be received under the above "time window + geographic area" constraints, and the core purpose is to reduce the risk of single device false alarm (such as accidental triggering of alarms due to device falling or collision) through "multi-device cross-verification".

[0073] In one specific embodiment, the execution of the space-time matching is performed by a multi-node cross-validation algorithm, which is implemented by: parsing a plurality of preliminary alarms received within a preset time window and in a specific geographic area, and extracting the P-wave arrival time, geographic location coordinates of the intelligent device, vibration characteristic parameters, and confidence value contained in each alarm.

[0074] Taking the P-wave arrival time of the first valid alarm in the region as the reference, the difference between the P-wave arrival time of other alarms and the reference time is calculated, and the standard deviation of all the differences is counted. If the standard deviation is less than a preset time threshold, it is determined that the time sequence matches, otherwise, the alarm that exceeds the time deviation threshold is marked as suspicious data.

[0075] It should be noted that the preset time threshold is set by professionals, for example, it is set based on the laws of seismic wave propagation and the distribution of regional devices, combined with the propagation speed of P-waves in the earth's crust (3-6 km / s) and the average distribution density of intelligent devices in the region to calculate the basic threshold value, to ensure that the alarm time difference caused by the same earthquake is within a reasonable range. At the same time, reference is made to the standard deviation data of the P-wave arrival time of multiple intelligent devices in historical earthquake events for calibration, taking into account the need to reduce misjudgment and accurately select the same earthquake alarm.

[0076] Based on the geographic location coordinates of each intelligent device and the P-wave arrival time, a theoretical epicenter candidate area is calculated using a back projection algorithm, the spatial distance between the actual position of each intelligent device and the theoretical epicenter candidate area is counted, the average distance and the dispersion coefficient are calculated, and if the dispersion coefficient is less than a preset spatial threshold, it is determined that the spatial distribution matches.

[0077] It should be noted that the preset spatial threshold is set by professionals, for example, it is set based on the need to ensure the accuracy of epicenter positioning and the effectiveness of spatial matching, combined with the positioning error characteristics of the back projection algorithm, the distribution density of regional intelligent devices, and the dispersion coefficient data of historical earthquake epicenter positioning to determine the basic value; At the same time, it is associated with dynamic calibration of estimated magnitude, and the threshold value is appropriately reduced at high magnitude to strictly select accurate matching data, and the threshold value is slightly increased at low magnitude to adapt to the positioning error range. The back projection algorithm is a prior art and will not be described in detail here.

[0078] For the alarms that pass the space-time matching, the cosine similarity of the vibration characteristic parameters and the peak value deviation rate are calculated, and the alarms with a similarity higher than a preset feature threshold and a peak value deviation rate lower than a preset deviation threshold are retained to form an effective sample set.

[0079] It should be noted that the preset feature threshold, the preset deviation threshold are set by professionals, for example, the preset feature threshold (cosine similarity) is set based on the consistency of vibration feature parameters, and is determined in combination with the cosine similarity distribution data of historical earthquake P waveforms, so as to ensure that the alarm with highly matched waveform features is screened out; the preset deviation threshold (peak deviation rate) is set based on the reasonable fluctuation range of historical earthquake vibration peak value, and the calibration deviation of the sensor accuracy of the equipment is also referred to, and the feature matching accuracy and the inclusiveness of small errors of the equipment are considered together. The cosine similarity and the peak deviation rate of the calculated vibration feature parameters are prior art, and will not be described in detail here.

[0080] The confidence values of the effective sample set are weighted and summed, if the weighted result is higher than the preset confidence threshold, and the proportion of the number of the effective sample set in the total number of the received alarms exceeds the preset number threshold, it is determined that the regional seismic event is established, otherwise, it is determined that it is not a seismic event.

[0081] It should be noted that the preset confidence threshold and the preset number threshold are set by professionals, for example, based on statistical analysis of a large amount of historical seismic event data and interference event data. The core purpose is to achieve the best balance between maximizing the success rate of early warning and minimizing the false alarm risk. The specific values are finally determined through repeated testing and optimization under different geographical regions and network environments, and can be dynamically adjusted according to the system operation feedback.

[0082] In one specific embodiment, the hierarchical early warning information is specifically implemented by: based on the regional event data matched in space and time and verified by multiple nodes, using a finite fault model, calculating and determining the estimated epicenter position, magnitude and seismic wave propagation speed parameters, combining the preset regional geological structure data and site effect parameters, calculating the estimated intensity values of different locations in the affected area through the seismic motion prediction equation, based on the estimated intensity values, dividing the affected area into three early warning levels, and matching the corresponding early warning instruction text and response suggestions for each early warning level.

[0083] It should be noted that the calculation and determination of the estimated epicenter position first extracts the P-wave arrival time and device coordinates of multiple intelligent devices from regional event data; then according to the preset geometric parameters such as the trend, dip angle and sliding angle of the fault, the fault grid is constructed and several sub-faults are divided; finally, through the inversion algorithm, the residual error of the observation value and the model calculation value of the P-wave arrival time of each device is minimized to determine the equivalent rupture initiation point on the fault, and the projection of the point on the ground is the estimated epicenter position; the calculation and determination of the estimated magnitude first calculate the fault rupture area according to the fault parameters obtained by inversion and width; then extract the peak vibration acceleration of each device from the regional event data, combine the distance between the device and the estimated epicenter, and correct the equivalent peak acceleration at the source through the empirical formula; finally, based on the fault rupture area and the equivalent peak acceleration, substitute into the magnitude calculation formula to obtain the estimated magnitude; the calculation and determination of the estimated seismic wave propagation velocity parameter uses the P-wave arrival time difference of multiple intelligent devices and the distance between the devices to invert the propagation velocity of the seismic wave (mainly P-wave) in the regional medium. First, select two devices with known distance and calculate the P-wave arrival time difference; then assume that the seismic wave propagates in a homogeneous medium, and preliminarily estimate the propagation velocity; finally, combined with regional geological structure data (such as stratum lithology distribution), the velocity is corrected layer by layer (different velocities at different depths), and the final seismic wave propagation velocity parameter is obtained.

[0084] The preset regional geological structure data mainly refers to the key information of known active fault distribution, fault type (such as strike-slip, thrust) and regional stress field, which is used to constrain the possible rupture mode of the finite fault model.

[0085] The preset site effect parameter refers to the site amplification effect coefficient drawn based on the local soil structure, geological conditions (such as soft soil thickness, Vs30 wave velocity), which is used to correct the amplification or weakening effect of bedrock ground motion when propagating to the ground surface.

[0086] First-level warning (slight impact): When the estimated magnitude M < 3.0, or 3.0 ≤ M < 4.0 and the distance D > 30 kilometers, the warning information contains the guidance content of "slight vibration reminder" and "no need to take emergency measures";

[0087] Second-level warning (moderate impact): When 4.0 ≤ M < 5.0 and the distance D ≤ 30 kilometers, or 5.0 ≤ M < 6.0 and the distance D > 50 kilometers, the warning information contains "moderate vibration warning", "avoid nearby (such as away from windows and furniture)", and "vibration predicted arrival time";

[0088] Three-level early warning (serious impact): When the estimated magnitude M≥5.0 and the distance D≤50 km, or M≥6.0, the early warning information includes "emergency warning", "immediately evacuate to open areas (or indoor load-bearing wall roots)", "predicted arrival time of vibration", and "location of nearby emergency shelters";

[0089] Based on the real-time geographic location of intelligent devices, only devices outside the "threat area" that may be affected by the vibration are pushed the corresponding level of early warning information - first-level early warning is pushed to within 50 km of the epicenter, second-level early warning is pushed to within 100 km of the epicenter, and third-level early warning is pushed to within 200 km of the epicenter, to avoid pushing invalid information to areas with no impact.

[0090] Cloud control fusion and calculation module: in communication with the edge aggregation and verification module, for converging and fusing regional event data and professional seismic network data, performing high-precision seismic positioning and magnitude calculation based on multi-source information, determining the final seismic three elements, and issuing accurate early warning instructions and linkage control signals to the public emergency system.

[0091] In one specific embodiment, the high-precision seismic positioning based on multi-source information is implemented by: time synchronization calibration and outlier rejection of regional event data and professional seismic network data, retaining effective data with a time stamp error within a preset range, constructing a multi-source data fusion positioning model, assigning weights to intelligent device data and network data according to spatial distribution density, using an iterative weighted least squares method to solve the epicenter coordinates, taking the distance residual of each intelligent device to the theoretical epicenter as the optimization objective, and reducing the residual sum of squares through multiple iterations to obtain high-precision epicenter coordinates.

[0092] It should be noted that the preset range is determined based on the typical synchronization error of intelligent device clock and national standard time, and the statistical distribution characteristics of network transmission delay. By analyzing the time stamp error distribution of effective seismic events and interference events in historical data, the preset range is set to a reasonable interval that can cover most real events while excluding obvious abnormal data (such as ±500 milliseconds to ±2 seconds). This range can be differentiated according to different regional communication conditions and device performance, and dynamically optimized and adjusted with accumulated data from system operation.

[0093] In one specific embodiment, the magnitude calculation is implemented by: based on intelligent device data, using vibration peak parameters and device distribution radius to convert regional equivalent magnitude through an empirical formula, based on professional network data, using surface wave magnitude calculation method, combining waveform amplitude recorded by the station and epicentral distance correction value to determine network magnitude, weighting and fusing the two types of magnitude results, introducing a bias correction factor of historical seismic events to generate the final magnitude result.

[0094] It should be noted that the empirical formula, surface wave magnitude calculation method is prior art, here does not carry on too much repetition.

[0095] Reference Figure 2 As shown, a method for providing a smart device-based earthquake early warning analysis system is provided, comprising: S1, terminal perception and preliminary judgment: for continuously collecting local vibration data and environmental data through a sensor array, and performing real-time fusion and processing on the data, identifying the characteristics of the P wave through a dynamic weight adjustment model, generating and outputting an encrypted preliminary warning containing confidence and characteristic parameters.

[0096] S2, adaptive network transmission: based on D2D communication protocol construction, for automatically forming a dynamic self-organizing network among smart devices, forming a multi-hop transmission link through the election of relay nodes, and performing differential encoding, aggregation and encrypted relay transmission on the preliminary warning.

[0097] S3, edge aggregation and verification: for receiving multiple preliminary warnings from a specific geographic area within a preset time window, performing spatiotemporal matching and multi-node cross-verification algorithms, determining whether the regional earthquake event is valid, and if the regional earthquake event is valid, immediately generating hierarchical warning information containing estimated magnitude and range, and issuing it to the threat area outside.

[0098] S4, cloud control fusion and calculation: for gathering and fusing regional event data and professional seismic network data, performing high-precision earthquake location and magnitude calculation based on multi-source information, determining the final three elements of the earthquake, and issuing precise warning instructions and linkage control signals to the public emergency system.

[0099] The basic principles, main features and advantages of the present application are shown and described above. Those skilled in the art should understand that the present application is not limited to the above embodiments, and the above embodiments and descriptions in the specification are only preferred examples of the present application and are not intended to limit the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application.

Claims

1. A smart device based earthquake early warning analysis system, characterized in that, The application relates to a terminal sensing and preliminary judgment module, a self-adaptive network transmission module, an edge aggregation and verification module and a cloud control fusion and calculation module. The terminal sensing and preliminary judgment module is used for continuously collecting local vibration data and environment data through a sensor array, performing real-time fusion and processing on the data, identifying a seismic P wave feature through a dynamic weight adjustment model, and generating and outputting an encrypted preliminary warning containing a confidence level and feature parameters. The self-adaptive network transmission module is connected with the terminal sensing and preliminary judgment module, is constructed based on a D2D communication protocol, is used for automatically establishing a dynamic self-organizing network among intelligent devices, forming a multi-hop transmission link through election of relay nodes, and performing differential encoding, aggregation and encrypted relay transmission on the preliminary warning. The method for automatically establishing the dynamic self-organizing network among the intelligent devices comprises the following steps: After the intelligent device is started, a detection frame containing a device unique identifier and a hardware feature code is broadcasted through the D2D communication protocol, adjacent intelligent devices receive the detection frame, the legality of the identifier of the sending party is verified, and a response frame containing the identifier and communication capacity of the adjacent intelligent devices is returned after the verification is passed, so that a temporary communication link is established between the two parties. Each intelligent device constructs a priority list of adjacent devices based on signal strength parameters and relative position information in the response frame, forms a star-shaped subnetwork with a core device as the center, and realizes all-network interconnection through cross-connected intelligent devices. The intelligent device periodically sends a state heartbeat packet to the connected nodes, the state heartbeat packet contains current power, residual bandwidth and position change, and if the receiving node does not continuously receive the heartbeat packet and the signal strength is below a threshold value, it is determined that the connection is invalid, and the next intelligent device is selected from the priority list to reestablish the connection. When the intelligent device detects that the position moves more than a preset distance, the intelligent device actively sends position update information to the original connection node and broadcasts a new detection frame to discover new intelligent devices around, judges possible disconnected connections based on the moving speed and direction, and establishes a preliminary connection with the new discovered intelligent device in advance. The relay forwarding amount of each node is counted in real time, and when the relay forwarding amount of a node exceeds a preset forwarding amount threshold value, the node automatically sends a load balancing request to adjacent nodes, the adjacent nodes respond to the request according to the remaining resources of the adjacent nodes, and part of the relay tasks are shared. The edge aggregation and verification module is connected with the self-adaptive network transmission module, is used for receiving multiple preliminary warnings from a specific geographical area within a preset time window, executing a space-time matching and multi-node cross-verification algorithm, judging whether a regional seismic event is established, generating graded warning information containing an estimated magnitude and a propagation range if the regional seismic event is established, and publishing the graded warning information to an area outside a threat area. The cloud control fusion and calculation module is in communication with the edge aggregation and verification module, is used for converging and fusing regional event data and professional seismic network data, performing high-precision seismic positioning and magnitude calculation based on multiple information sources, determining final seismic three elements, and issuing accurate warning instructions and linkage control signals to a public emergency system.

2. The earthquake early warning analysis system based on smart devices according to claim 1, wherein, The dynamic weight adjustment model comprises the following steps: The data collected by the sensor array is divided into core sensing dimensions and auxiliary correction dimensions, wherein the core sensing dimensions include vibration waveform data and device posture data, and the auxiliary correction dimensions include environment noise spectrum data and environment air pressure fluctuation data. The historical earthquake sample data is obtained from a database, and initial weight proportions higher than those of auxiliary correction dimensions are assigned to core perception dimensions based on offline training of the historical earthquake sample data, wherein vibration waveform data is taken as a core basis for P-wave identification and is assigned the highest initial weight; The weight optimization is achieved by using a reinforcement learning framework, the environmental noise level, device historical false alarm types, and sensor data quality are taken as a state space, the preset step increase and decrease of each dimension feature weight are taken as an action space, the P-wave identification accuracy improvement is taken as a positive reward, and false alarms and omissions are taken as negative punishments, the weight matrix is iteratively updated by using a time difference algorithm, the weight matrix is stored at a low precision, only feature channels with a contribution to P-wave identification exceeding a preset contribution threshold are retained, and a dynamic weight adjustment model is obtained. 3.The earthquake early warning analysis system based on smart devices of claim 1, wherein, The election relay node forms a multi-hop transmission link, and the specific implementation method is as follows: After the intelligent device accesses the self-organizing network, the hardware capability and running state of the device are automatically detected, and the device that meets the preset basic condition actively broadcasts candidate nomination information containing the capability parameters of the device to the network; After each intelligent device in the network receives the candidate nomination information, the candidate node is quantitatively scored based on a preset evaluation model, and the evaluation dimensions include communication coverage, historical forwarding reliability, resource redundancy, and location stability; Each intelligent device sends a vote confirmation to the candidate node with the highest priority according to the scoring result, and when the candidate node accumulates a valid vote number reaching a preset proportion of the total number of local network devices, the candidate node automatically becomes an official relay node and broadcasts an election declaration to the whole network; The term of the relay node is set as a dynamic length, the relay node continuously sends a state heartbeat to the surrounding intelligent devices during the term, and before the end of the term, the intelligent devices in the current coverage range reevaluate the performance of the relay node during the term, and the relay node is automatically extended if the relay node meets the re-election condition, otherwise, a new round of election is started; When the relay node has a sudden failure or its performance drops below a threshold, the intelligent devices directly connected to the relay node immediately initiate a temporary election to quickly select a replacement node from the standby candidate nodes.

4. The earthquake early warning analysis system based on smart devices according to claim 1, wherein, The spatiotemporal matching and multi-node cross-validation algorithm is executed, and the specific implementation method is as follows: A plurality of preliminary alarms received within a preset time window and in a specific geographic region are analyzed, and P-wave arrival times, geographic location coordinates of intelligent devices, vibration characteristic parameters, and confidence values contained in each alarm are extracted; The P-wave arrival time of the first valid alarm in the region is taken as a reference, the difference between the P-wave arrival times of other alarms and the reference time is calculated, the standard deviation of all difference values is counted, if the standard deviation is less than a preset time threshold, it is determined that the time sequence matches, otherwise, the alarms exceeding the time deviation threshold are marked as suspicious data; Based on the geographic location coordinates of the intelligent devices and the P-wave arrival times, a back projection algorithm is used to calculate a candidate area of a theoretical epicenter, the spatial distances between the actual positions of the intelligent devices and the candidate area of the theoretical epicenter are counted, the average distance and the dispersion coefficient are calculated, and if the dispersion coefficient is less than a preset spatial threshold, it is determined that the spatial distribution matches. For the alarm matched in space-time, the cosine similarity of the vibration characteristic parameter and the peak deviation rate are calculated, the alarm with the similarity higher than a preset characteristic threshold and the peak deviation rate lower than a preset deviation threshold is retained to form an effective sample set; The confidence values of the effective sample set are weighted and summed, if the weighted result is higher than a preset confidence threshold and the proportion of the number of the effective sample set in the total number of the initial received alarms exceeds a preset number threshold, it is determined that the regional seismic event is established, otherwise, it is determined that the regional seismic event is not established.

5. The smart device based earthquake early warning analysis system of claim 1, wherein, The hierarchical early warning information is specifically implemented as follows: Based on the regional event data matched in space-time and cross-verified by multiple nodes, a finite fault model is used to calculate and determine the estimated epicenter position, magnitude and seismic wave propagation velocity parameters, combined with preset regional geological structure data and site effect parameters, the estimated intensity values of different locations in the affected area are calculated by using a seismic motion prediction equation, based on the estimated intensity values, the affected area is divided into three early warning levels, and corresponding early warning instruction texts and response suggestions are matched for each early warning level.

6. The smart device based earthquake early warning analysis system of claim 1, wherein, The high-precision seismic positioning based on multiple source information is specifically implemented as follows: The regional event data and professional seismic network data are time-synchronized, calibrated and abnormal values are removed, the effective data with a time stamp error within a preset range are retained, a multi-source data fusion positioning model is constructed, the intelligent device data and the network data are weighted according to the spatial distribution density, the iterative weighted least squares method is used to solve the epicenter coordinates, the distance residual of each intelligent device to the theoretical epicenter is taken as the optimization target, the residual sum of squares is reduced through multiple iterations, and the high-precision epicenter coordinates are obtained. 7.The earthquake early warning analysis system based on smart devices of claim 1, wherein, The magnitude calculation is specifically implemented as follows: Based on the intelligent device data, the regional equivalent magnitude is converted by using the vibration peak value parameter and the device distribution radius through an empirical formula, based on the professional network data, the surface wave magnitude calculation method is used, combined with the waveform amplitude recorded by the station and the epicentral distance correction value to determine the network magnitude, the two types of magnitude results are weighted and fused, the deviation correction factor of the historical seismic event is introduced, and the final magnitude result is generated.

8. A method for implementing the smart device based earthquake early warning analysis system of any of claims 1-7, characterized in that, It includes: S1, terminal perception and preliminary judgment: used for continuously collecting local vibration data and environmental data through a sensor array, and performing real-time fusion and processing on the data, identifying the characteristics of the P wave through a dynamic weight adjustment model, and generating and outputting an encrypted preliminary alarm containing confidence and characteristic parameters; S2, adaptive network transmission: based on the D2D communication protocol, used for automatically forming a dynamic self-organizing network among intelligent devices, forming a multi-hop transmission link through the election of relay nodes, and differentially encoding, aggregating and encrypting the preliminary alarm for relay transmission; S3, edge aggregation and verification: used for receiving multiple preliminary alarms from a specific geographic area within a preset time window, performing space-time matching and multi-node cross-verification algorithm, determining whether a regional seismic event is established, if the regional seismic event is established, generating hierarchical early warning information containing estimated magnitude and propagation range immediately, and publishing outside the threatened area; S4, cloud control fusion and calculation: used for gathering and fusing regional event data and professional seismic network data, performing high-precision earthquake positioning and magnitude calculation based on multi-source information, determining final earthquake three elements, and issuing accurate early warning instructions and linkage control signals to the public emergency system.

Images