Road network abnormal event identification and risk evaluation early warning method based on vibration perception

Abstract

The invention discloses a road network abnormal event identification and risk evaluation early warning method based on vibration perception, and relates to the technical field of urban road safety monitoring. Comprising the steps of collecting road vibration data of a road network through a sensor; preprocessing the road vibration data through Kalman filtering and multi-dimensional feature extraction; abnormal event recognition is carried out on the processed road vibration data through a differentiation algorithm, wherein abnormal events comprise vehicle overload, illegal construction and road structure damage; and carrying out risk scoring on the abnormal events through dynamic weight distribution, dividing the risk scores according to the risk levels, and carrying out graded response. According to the invention, through differential algorithm design, a mixed model architecture and a hierarchical response mechanism are adopted, abnormal event identification, risk level quantitative evaluation and differential early warning processing are realized, the safety state of an urban road network is monitored in real time, and early warning information is provided.

Description

Technical Field

[0001] The present invention relates to the technical field of urban road safety monitoring, and in particular to a method for identifying abnormal road network events and conducting risk assessment and early warning based on vibration perception. Background Art

[0002] With the acceleration of urbanization, urban roads face multiple safety threats, including overloaded vehicles, illegal construction disturbances, and aging and cracking structures. Traditional manual inspections and single-point sensor monitoring, due to incomplete coverage and poor real-time performance, are unable to meet these demands. Existing technologies face three bottlenecks: Multi-source data (traffic flow, vibration, and geology) lacks spatiotemporal synchronization, making it difficult to construct a unified risk model; traditional algorithms (such as SVM) are poorly adaptable to complex operating conditions and are susceptible to interference from environmental noise; static weight assessment models cannot dynamically reflect the evolution of risks; and while vibration sensing technology offers the advantages of non-contact and full road coverage, its engineering application still faces challenges: MEMS sensors are susceptible to temperature drift, low-frequency signals (<10Hz) are easily drowned out by traffic noise, and there is a lack of multi-scale feature fusion algorithms.

[0003] Therefore, there is an urgent need for a road network abnormal event identification and risk assessment warning method based on vibration perception to solve the above problems. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for identifying abnormal road network events and conducting risk assessment and early warning based on vibration perception, so as to solve the problems raised in the above-mentioned background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] A road network abnormal event identification and risk assessment early warning method based on vibration perception, the method comprising:

[0007] Step S100: collecting road vibration data of the road network through sensors;

[0008] Step S200: pre-processing the road vibration data through Kalman filtering and multi-dimensional feature extraction;

[0009] Step S300: using a differentiation algorithm to identify abnormal events on the processed road vibration data, where the abnormal events include vehicle overloading, illegal construction, and road structure damage;

[0010] Step S400: Risk scoring of abnormal events is performed through dynamic weight allocation, and risk scores are divided according to risk levels and graded responses are given.

[0011] Furthermore, the specific implementation process of collecting road vibration data of the road network includes:

[0012] The vibration signals of the road network are monitored in real time by sensors, and the vibration signals of the road network are synchronized with GPS through the PTP protocol to generate road vibration data. The sensors include acceleration sensors, seismic wave sensors and temperature sensors, among which: the temperature sensor is used to collect ambient temperature and correct the output signal of the sensor through an adaptive baseline correction algorithm.

[0013] It should be explained that the protection level of the acceleration sensor shell is IP67, and it uses a MEMS three-axis accelerometer with a range of ±20g and a noise density of It is installed on the top of the light pole through a magnetic base, 6-8m from the ground, and the installation position should avoid the area directly run over by vehicles.

[0014] The seismic wave sensor is a Geospace GS-11D model, featuring a low-frequency geophone with a frequency response of 0.1-100 Hz. It is buried 30 cm to the side of the roadbed, with a distance of 5 m or less from the accelerometer. A PT1000 thermistor is used for temperature compensation, collecting ambient temperature in real time and correcting the sensor output signal. The GPS is a u-blox NEO-M8N model. The edge computing node is an NVIDIA Jetson Xavier NX.

[0015] The adaptive baseline correction algorithm in the anti-interference design can monitor and compensate for the impact of ambient temperature changes on sensor output in real time, ensuring the accuracy and stability of sensor data; the multi-source data synchronization process uses the PTP protocol to achieve precise time synchronization between sensors, ensuring that the data collected by different sensors are consistent in time, with the error controlled within 1ms. Combined with the absolute timestamp provided by the GPS module, the time and location of vibration events can be accurately determined; the LoRaWAN gateway deployed on the edge computing node supports multiple self-organizing network transmissions, and can achieve large-scale wireless coverage in urban obstructed environments, with a coverage radius of up to 2km, effectively solving data transmission problems and ensuring that the data collected by the sensors can be transmitted to the data center in a timely and stable manner for processing and analysis.

[0016] Furthermore, the specific implementation process of preprocessing the road vibration data includes:

[0017] The road vibration data collected by the acceleration sensor is preprocessed using Kalman filtering, and features of the road vibration data are extracted from the time domain, frequency domain, and time-frequency domain. Dimensionality reduction is performed using principal component analysis, and the first three principal components are extracted as fusion features.

[0018] Among them, the Kalman filter process includes:

[0019] Construct the state vector:

[0020] x k=[a k ,a' k ] T

[0021] Among them, x k represents the state vector, a k Represents the real-time measurement value of the acceleration sensor, a' k Indicates the rate of change of acceleration;

[0022] According to the state equation x k|k-1 =Fx k-1 Predict the current state, where x k|k-1 represents the predicted state at the current moment k, x k-1 represents the state of the previous moment k-1, and F represents the state x of the previous moment k-1. k-1 Mapped to the predicted state x at the current time k k|k-1 The state transition matrix of

[0023] State transition matrix Where Δt represents the sampling interval;

[0024] Predict the covariance at the current moment: P k|k-1 =FP k-1 F T +Q, where P k|k-1 represents the predicted covariance of the current moment k, P k-1 represents the covariance of the previous moment k-1, Q represents the process noise covariance matrix used to characterize the dynamic uncertainty of the system, Q = diag(0.01,0.1);

[0025] Combined observations z k , calculate the Kalman gain D k =P k|k-1 H T (HP k|k-1 H T +R k ) -1 , where D k represents the Kalman gain at the current moment k, H represents the observation matrix, and H=[1,0], R k represents the observation noise covariance;

[0026] The observation noise covariance R k For dynamic adjustment, the formula is:

[0027] R k =αR k-1 +(1-α)(z k -Hx k|k-1 ) 2

[0028] Where α represents the smoothing factor, α=0.95, R k-1 represents the observation noise covariance of the previous moment k-1; z k represents the observation value at the current moment k;

[0029] Update the state estimate x k =x k|k-1 +D k (z k -Hx k|k-1 ) and the covariance matrix P * k =(ID k H)P k|k-1 , output smoothed acceleration and acceleration change rate, where x k represents the updated state estimate, P * k represents the updated covariance, and I represents the identity matrix.

[0030] The specific implementation process of the multi-dimensional feature extraction includes:

[0031] Extract time domain features on the time axis and construct a time domain feature set. The time domain features include statistical features and waveform indicators, where statistical features include mean, peak, variance, root mean square, kurtosis and skewness, and waveform indicators include peak factor C for quantifying signal impact strength. r and the impulse factor L used to identify transient abnormal events r ;

[0032] The formula for crest factor is as follows: Among them, C r represents the peak factor, Pvl represents the peak value, and Rms represents the root mean square. The formula for the pulse factor is as follows: Among them, L r represents the impulse factor, Me represents the mean;

[0033] Convert the road vibration data to the frequency domain, obtain the spectrum through Fourier transform, and extract the frequency domain features:

[0034] The 1 / 3 octave energy distribution is calculated by Fourier transform. The center frequency of the band covers 10-500Hz. The energy E of each band is m for:

[0035]

[0036] Among them, E m Indicates the energy of the mth frequency band, Band m represents the frequency band center frequency coverage of the mth frequency band, and X(f) represents the power spectrum density value of the fth frequency point of the mth frequency band;

[0037] Extract the power spectrum entropy PSE of the low frequency band (0.1-10Hz) that represents the signal spectrum complexity f :

[0038]

[0039] Among them, PSE f Represents the power spectrum entropy of the low frequency band, M represents the total number of frequency points in the low frequency band, p i represents the normalized probability distribution of the low-frequency power spectrum, E i Represents the energy of the i-th frequency point in the low frequency band, E h Represents the energy of the hth frequency point in the low frequency band;

[0040] Combine time and frequency information to extract time-frequency domain features:

[0041] The wavelet basis function is used to decompose the wavelet into 5 layers, and the energy proportion of the detail coefficients of each layer is extracted:

[0042]

[0043] Among them, RA j Indicates the ratio of the energy of the detail coefficient of the jth layer to the energy of the total detail coefficient, d j (t) represents the detail coefficient of the jth layer, s j (t) represents the energy of detail coefficients at layer j, and t represents the index of wavelet coefficients;

[0044] The time-frequency energy matrix is constructed, the dimension is reduced through principal component analysis, and the first three principal components are extracted as fusion features.

[0045] It should be explained that principal component analysis (PCA) is a classic unsupervised linear dimensionality reduction method that projects high-dimensional data into a low-dimensional space by extracting the principal components (i.e., the directions with the largest variance) from the data, while retaining as much key information as possible. The specific implementation process includes: centering (subtracting the mean) and scaling (dividing by the standard deviation) each feature to eliminate dimensionality effects; calculating the covariance matrix to reflect the correlation between features; decomposing the covariance matrix to obtain eigenvalues and corresponding eigenvectors; sorting the eigenvalues from large to small, and the corresponding eigenvectors are the principal component directions; selecting the eigenvectors corresponding to the specified number of largest eigenvalues based on the cumulative variance contribution rate (such as retaining 95% of the variance) or business requirements (such as visualization needs to be reduced to 2-3 dimensions); projecting the original data onto the selected principal components to obtain the reduced-dimensional data.

[0046] Furthermore, the specific implementation process of identifying abnormal events on the processed road vibration data using the differentiation algorithm includes:

[0047] Abnormal events include overloaded vehicles, illegal construction, and damaged road structures;

[0048] To address the differences in vibration characteristics of different abnormal events, a differentiated algorithm was designed. An optimized SVM was used to handle the nonlinear classification problem of overloaded vehicles, and model adaptability was improved through hybrid kernel functions and incremental learning. For unstructured vibration signals from illegal construction, an improved K-Medoids clustering method combined with dynamic time warping (DTW) was proposed to address the noise sensitivity of traditional K-Means. For the temporal characteristics of road structure damage, a Bi-LSTM+Attention network was constructed, integrating transfer learning to achieve trend prediction with small samples. The three algorithms, from the perspectives of pattern classification, cluster matching, and temporal prediction, provide full-scenario analysis capabilities for multi-source road vibration data:

[0049] (1) Use support vector machine to detect overloaded vehicles; Support vector machine (SVM) is a classification method based on statistical learning theory. Its core idea is to find an optimal hyperplane in the feature space to separate data points of different categories as much as possible;

[0050] Feature selection was performed on the processed road vibration data. The time domain root mean square value (reflecting the energy of the signal), the energy proportion of the 50-200 Hz frequency domain (this frequency band is related to the vibration characteristics of overloaded vehicles), and the energy of the third wavelet layer detail coefficient were extracted as input features.

[0051] The features are mapped to a high-dimensional feature space for processing through a hybrid kernel function. The specific formula of the hybrid kernel function is:

[0052]

[0053] Among them, xp q ,xp j represents the normalized eigenvector, xp q T xp j represents vector dot product, ||xp q -xp j || 2 represents the square of the Euclidean distance, γ poly and γ rbf They represent the weights of the polynomial kernel and the radial basis kernel respectively, and σ represents the parameter of the radial basis kernel;

[0054] Find the optimal hyperplane, the specific formula is:

[0055] yp im (w T xl im +b)≥1-ξ im ,ξim ≥0,im=1,...,n

[0056] Among them, w represents the normal vector of the hyperplane, b represents the intercept, C represents the penalty parameter, ξ i represents the slack variable, xl im and yp im Represent the feature vector and category label of the im-th sample respectively; im represents the total number of hyperplane samples;

[0057] Deploy an online support vector machine (SVM) to incrementally train the model based on newly collected road vibration data and update model parameters each cycle. When new data arrives, the SVM's online learning mechanism is used to incrementally train the model, enabling it to adapt to dynamic changes in vehicle load distribution.

[0058] (2) Detect illegal construction through K-Medoids clustering combined with dynamic time warping:

[0059] The low-frequency seismic wave signal was resampled from 10Hz to 100Hz to obtain richer detailed information. At the same time, Gaussian noise (SNR = 20dB) was added to expand the dataset size and enhance the generalization ability of the model.

[0060] Divide the dataset into K clusters, select the actual points in the dataset as cluster centers, and use Manhattan distance as the similarity metric. Compared with Euclidean distance, Manhattan distance is more robust to outliers. During the clustering process, optimize the selection of medoids through continuous iteration to minimize the objective function.

[0061] Define a construction feature template, such as the 1-3 Hz periodic pulse signal generated by an impact drill. Use the dynamic time warping algorithm to obtain the optimal alignment path between the two time series and determine the similarity between the signal to be detected and the construction feature template.

[0062] When the similarity between the signal to be detected and the construction feature template exceeds the similarity threshold, it is determined to be an illegal construction event;

[0063] (3) Detect road structure damage by building a Bi-LSTM+Attention network:

[0064] Long Short-Term Memory (LSTM) is a special type of Recurrent Neural Network (RNN). By introducing a gating mechanism, it solves the gradient vanishing and gradient exploding problems of traditional RNNs and can effectively process long-term series data. LSTM units contain an input gate, a forget gate, and an output gate.

[0065] The input layer receives the root mean square value of the time series vibration within a unit time; a two-layer bidirectional LSTM is built, with each layer containing 64 LSTM units. Weights are assigned to different time steps through the self-attention mechanism, and the output layer outputs the vibration prediction value of the future time node;

[0066] It should be explained that the input layer receives the 60-second time series vibration RMS value; a two-layer bidirectional LSTM (Bi-LSTM) is built, with each layer containing 64 LSTM units. Bi-LSTM can simultaneously process forward and reverse time series information, enhancing the ability to capture contextual information; an attention mechanism is introduced to weight key time steps and highlight important information; and the output layer predicts the vibration trend for the next 10 seconds.

[0067] Huber is used as the loss function, and its formula is:

[0068]

[0069] Among them, y represents the true value, represents the predicted value, δ represents the threshold parameter, δ = 0.1;

[0070] Pre-training is performed based on a public bridge vibration dataset to obtain common vibration features. During the fine-tuning phase, the parameters of the two-layer bidirectional LSTM layer are frozen, and only the Attention layer and the fully connected layer are trained. Road structure damage data is used to output road structure damage results.

[0071] The specific implementation process of assigning risk scores to abnormal events through dynamic weight distribution, dividing the risk scores according to risk levels and responding in a graded manner includes: constructing a dynamic weight distribution model and a graded response system to achieve closed-loop management from risk quantification to disposal. First, the entropy weight method-AHP fusion model is used to solve the problem of dynamic adjustment of weights of three types of risks: overload, construction, and structural damage, and a time decay factor is introduced to reflect the timeliness of risks; secondly, a four-level early warning response mechanism is designed to match differentiated disposal strategies through threshold division (<30 / 30-60 / 60-90 / ≥90); finally, relying on the GIS visualization dashboard, multi-dimensional information such as spatial location, event type, and historical trends is integrated to form a full-link technical architecture of "data collection-feature analysis-risk assessment-visualized decision-making" to support the refined management of urban road safety;

[0072] The dynamic weight allocation in risk assessment takes into account the time factor and subjective and objective weights. The risk weight is adjusted based on the time decay factor. Taking the overload event weight as an example, the formula is w overload =w0e -λt, where w0 is the initial weight, λ = 0.01 / hour is the decay coefficient, and t is time. The time decay factor reflects the timeliness of risk; as time passes, the impact of risk gradually decreases.

[0073] The entropy weight method is used to calculate the objective weight of the indicator. The basic principle of the entropy weight method is to determine the weight according to the degree of variation of the indicator. The greater the degree of variation, the higher the weight.

[0074] Based on the identification results of abnormal events, the objective weight of the indicator is calculated using the entropy weight method:

[0075]

[0076] Among them, pp uo Indicates the proportion of the u-th sample under the o-th indicator, g uo represents the index of the u-th sample after standardization under the o-th index, g Ψo represents the index of the Ψth sample after standardization under the oth index, zi represents the number of samples; e o represents the entropy value of the oth indicator; pp τo Indicates the proportion of the τth sample under the oth index, HL entropy,o represents the objective weight of the oth indicator, Oi represents the number of indicators, e η represents the entropy value of the ηth indicator;

[0077] Combined with the hierarchical analysis method to determine the subjective weight, construct the judgment matrix, compare the indicators pairwise, and obtain the relative weight HL AHP,o ;

[0078] The final weight is calculated by integrating subjective and objective weights: LG o =βHL entropy,o +(1-β)HL AHP,o , among which LG o represents the final weight of the oth indicator, and β represents the balance coefficient;

[0079] The risk score is obtained based on the final weight, such as: Among them, PFS represents risk score, X * uo represents the standardized value of the oth indicator of the uth sample;

[0080] Classifying the risk score response according to risk levels, including low risk, medium risk, high risk, and very high risk;

[0081] When the risk level is low (score <30), the relevant information is generated and stored in the database, and no alert is issued, so as to facilitate subsequent analysis and tracing;

[0082] When the risk level is medium risk (30≤score<60), the risk information will be sent to maintenance personnel in the jurisdiction via SMS push, so that they can understand the road conditions in a timely manner and take corresponding preventive measures.

[0083] When the risk level is high (60≤score<90), an audible and visual alarm is triggered, such as a road LED warning screen, to remind surrounding people to pay attention to road risks in an intuitive way, and at the same time attract the attention of relevant departments.

[0084] When the risk level is extremely high (score ≥ 90), not only will the sound and light alarm be triggered, but the traffic signal system will also be linked to restrict vehicle traffic to prevent potential dangers from further expanding, and the emergency management platform will be reported in a timely manner so that relevant departments can coordinate and deal with emergencies.

[0085] It's also worth explaining that users can drill down and analyze multiple dimensions through the dashboard, such as vibration energy, event type, and historical trends. For example, by viewing a vibration energy heat map, one can intuitively understand the distribution of vibration intensity in different road areas and quickly locate potential risk areas. Filtering by event type clearly reveals the location and frequency of various abnormal events, providing a basis for decision-making in road safety management. Analyzing historical trends helps identify changing patterns in road safety conditions and enable targeted preventive and maintenance measures to be taken in advance.

[0086] A road network abnormal event identification and risk assessment and warning system based on vibration sensing includes a sensor deployment and data acquisition module, a data preprocessing module, an abnormal event module, and a risk assessment and warning module:

[0087] The sensor deployment and data acquisition module is used to collect road vibration data. Accelerometers, seismic wave sensors, and temperature sensors are deployed to monitor the vibration signals of the road network in real time. A temperature sensor collects ambient temperature and uses an adaptive baseline correction algorithm to correct the sensor output signal. The module synchronizes the road network's vibration signals with GPS using the PTP protocol to generate road vibration data.

[0088] The data preprocessing module is used to preprocess the road vibration data by improving the Kalman filter parameter design and multi-dimensional feature extraction; the road vibration data collected by the acceleration sensor is preprocessed by Kalman filtering, and the features of the road vibration data are extracted from the time domain, frequency domain, and time-frequency domain through multi-dimensional feature extraction.

[0089] The abnormal event recognition module is used to identify abnormal events in processed road vibration data using a differentiation algorithm; detect overloaded vehicles using an optimized support vector machine; detect illegal construction using K-Medoids clustering combined with dynamic time warping; and detect road structure damage using a Bi-LSTM+Attention network.

[0090] The risk assessment and early warning module is used to assign risk scores to abnormal events through dynamic weight allocation, divide the risk scores according to the risk level and respond in a graded manner.

[0091] A terminal device includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The memory stores the computer program capable of running on the processor. When the processor loads and executes the computer program, a road network abnormal event identification and risk assessment and early warning method based on vibration perception is adopted.

[0092] A storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to execute the vibration-sensing-based road network abnormal event identification and risk assessment early warning method.

[0093] Compared with the existing technology, the beneficial effects achieved by the present invention are: by optimizing SVM, improving K-Medoids clustering and Bi-LSTM network, overloaded vehicle classification, illegal construction pattern matching and road structure damage trend prediction can be achieved;

[0094] By integrating the entropy weight method-AHP dynamic weight model with the hierarchical response mechanism, quantitative assessment of risk levels and differentiated early warning and disposal can be achieved;

[0095] It also has a high level of intelligence, can monitor the safety status of urban road networks in real time, provide early warning information, and is suitable for road risk assessment in a variety of complex scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0097] Figure 1 It is a schematic diagram of the steps of the road network abnormal event identification and risk assessment early warning method based on vibration perception of the present invention. DETAILED DESCRIPTION

[0098] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0099] See also Figure 1 The present invention provides a technical solution: a method for identifying abnormal road network events and assessing risk based on vibration perception, the method comprising:

[0100] Collect road vibration data of the road network through sensors;

[0101] Preprocessing of road vibration data through Kalman filtering and multi-dimensional feature extraction;

[0102] Using a differentiation algorithm to identify abnormal events on the processed road vibration data, the abnormal events include vehicle overloading, illegal construction, and road structure damage;

[0103] Abnormal events are scored for risk through dynamic weight allocation, and risk scores are divided and responded to in a graded manner according to risk levels.

[0104] Specifically, the specific implementation process of collecting road vibration data of the road network includes:

[0105] The vibration signals of the road network are monitored in real time by sensors, and the vibration signals of the road network are synchronized with GPS through the PTP protocol to generate road vibration data. The sensors include acceleration sensors, seismic wave sensors and temperature sensors, among which: the temperature sensor is used to collect ambient temperature and correct the output signal of the sensor through an adaptive baseline correction algorithm.

[0106] Specifically, the specific implementation process of preprocessing the road vibration data includes:

[0107] The road vibration data collected by the acceleration sensor is preprocessed using Kalman filtering, and features of the road vibration data are extracted from the time domain, frequency domain, and time-frequency domain. Dimensionality reduction is performed using principal component analysis, and the first three principal components are extracted as fusion features.

[0108] Among them, the Kalman filter process includes:

[0109] Construct the state vector:

[0110] x k =[a k ,a' k ] T

[0111] Among them, x k represents the state vector, a k Represents the real-time measurement value of the acceleration sensor, a' k Indicates the rate of change of acceleration;

[0112] According to the state equation x k|k-1 =Fx k-1 Predict the current state, where x k|k-1 represents the predicted state at the current moment k, x k-1 represents the state of the previous moment k-1, and F represents the state x of the previous moment k-1. k-1 Mapped to the predicted state x at the current time k k|k-1 The state transition matrix of

[0113] State transition matrix Where Δt represents the sampling interval;

[0114] Predict the covariance at the current moment: P k|k-1 =FP k-1 F T +Q, where P k|k-1 represents the predicted covariance of the current moment k, P k-1 represents the covariance of the previous moment k-1, Q represents the process noise covariance matrix used to characterize the dynamic uncertainty of the system, Q = diag(0.01,0.1);

[0115] Combined observations z k , calculate the Kalman gain D k =P k|k-1 H T (HP k|k-1 H T +R k ) -1 , where D k represents the Kalman gain at the current moment k, H represents the observation matrix, and H=[1,0], R k represents the observation noise covariance;

[0116] The observation noise covariance R k For dynamic adjustment, the formula is:

[0117] R k =αR k-1 +(1-α)(z k -Hx k|k-1 ) 2

[0118] Where α represents the smoothing factor, α=0.95, R k-1 represents the observation noise covariance of the previous moment k-1; z krepresents the observation value at the current moment k;

[0119] Update the state estimate x k =x k|k-1 +D k (z k -Hx k|k-1 ) and the covariance matrix P * k =(ID k H)P k|k-1 , output smoothed acceleration and acceleration change rate, where x k represents the updated state estimate, P * k represents the updated covariance, and I represents the identity matrix.

[0120] Specifically, the specific implementation process of the multi-dimensional feature extraction includes:

[0121] Extract time domain features on the time axis and construct a time domain feature set. The time domain features include statistical features and waveform indicators, where statistical features include mean, peak, variance, root mean square, kurtosis and skewness, and waveform indicators include peak factor C for quantifying signal impact strength. r and the impulse factor L used to identify transient abnormal events r ;

[0122] The formula for crest factor is as follows: Among them, C r represents the peak factor, Pvl represents the peak value, and Rms represents the root mean square. The formula for the pulse factor is as follows: Among them, L r represents the impulse factor, Me represents the mean;

[0123] Convert the road vibration data to the frequency domain, obtain the spectrum through Fourier transform, and extract the frequency domain features:

[0124] The 1 / 3 octave energy distribution is calculated by Fourier transform. The center frequency of the band covers 10-500Hz. The energy E of each band is m for:

[0125]

[0126] Among them, E m Indicates the energy of the mth frequency band, Band m represents the frequency band center frequency coverage of the mth frequency band, and X(f) represents the power spectrum density value of the fth frequency point of the mth frequency band;

[0127] Extract the power spectrum entropy PSE of the low frequency band (0.1-10Hz) that represents the signal spectrum complexity f:

[0128]

[0129] Among them, PSE f Represents the power spectrum entropy of the low frequency band, M represents the total number of frequency points in the low frequency band, p i represents the normalized probability distribution of the low-frequency power spectrum, E i Represents the energy of the i-th frequency point in the low frequency band, E h Represents the energy of the hth frequency point in the low frequency band;

[0130] Combine time and frequency information to extract time-frequency domain features:

[0131] The wavelet basis function is used to decompose the wavelet into 5 layers, and the energy proportion of the detail coefficients of each layer is extracted:

[0132]

[0133] Among them, RA j Indicates the ratio of the energy of the detail coefficient of the jth layer to the energy of the total detail coefficient, d j (t) represents the detail coefficient of the jth layer, s j (t) represents the energy of detail coefficients at layer j, and t represents the index of wavelet coefficients;

[0134] The time-frequency energy matrix is constructed, the dimension is reduced through principal component analysis, and the first three principal components are extracted as fusion features.

[0135] Specifically, the specific implementation process of identifying abnormal events on processed road vibration data using a differentiation algorithm includes:

[0136] Abnormal events include overloaded vehicles, illegal construction, and damaged road structures;

[0137] Based on the differences in vibration characteristics of different abnormal events, a differentiated algorithm was designed, using an optimized support vector machine to detect overloaded vehicles; illegal construction was detected through K-Medoids clustering combined with dynamic time warping; and road structure damage was detected by constructing a Bi-LSTM+Attention network.

[0138] Optimized support vector machine is used to detect overloaded vehicles;

[0139] Feature selection is performed on the processed road vibration data, and the root mean square value in the time domain, the energy proportion of 50-200Hz in the frequency domain, and the energy of the detail coefficient of the third layer of the wavelet are extracted as input features;

[0140] The features are mapped to a high-dimensional feature space for processing through a hybrid kernel function. The specific formula of the hybrid kernel function is:

[0141]

[0142] Among them, xp q ,xp j represents the normalized eigenvector, xp q T xp j represents vector dot product, ||xp q -xp j || 2 represents the square of the Euclidean distance, γ poly and γ rbf They represent the weights of the polynomial kernel and the radial basis kernel respectively, and σ represents the parameter of the radial basis kernel;

[0143] Find the optimal hyperplane, the specific formula is:

[0144] yp im (w T xl im +b)≥1-ξ im ,ξ im ≥0,im=1,...,n

[0145] Among them, w represents the normal vector of the hyperplane, b represents the intercept, C represents the penalty parameter, ξ i represents the slack variable, xl im and yp im Represent the feature vector and category label of the im-th sample respectively; im represents the total number of hyperplane samples;

[0146] Deploy an online support vector machine, incrementally train the model based on newly collected road vibration data in each cycle, and update the model parameters;

[0147] Detect illegal construction through K-Medoids clustering combined with dynamic time warping:

[0148] Resample the low-frequency seismic wave signal and add Gaussian noise to expand the data set;

[0149] Divide the data set into K clusters, select the actual points in the data set as cluster centers, use Manhattan distance as the similarity metric, and optimize the selection of cluster centers through continuous iteration to minimize the objective function;

[0150] Define a construction feature template, use the dynamic time warping algorithm to obtain the optimal alignment path between the two time series, and obtain the similarity between the signal to be detected and the construction feature template;

[0151] When the similarity between the signal to be detected and the construction feature template exceeds the similarity threshold, it is determined to be an illegal construction event;

[0152] By building a Bi-LSTM+Attention network, we can detect road structure damage:

[0153] The input layer receives the root mean square value of the time series vibration within a unit time; a two-layer bidirectional LSTM is built, with each layer containing 64 LSTM units. Weights are assigned to different time steps through the self-attention mechanism, and the output layer outputs the vibration prediction value of the future time node;

[0154] Huber is used as the loss function, and its formula is:

[0155]

[0156] Among them, y represents the true value, represents the predicted value, δ represents the threshold parameter, δ = 0.1;

[0157] Pre-training is performed based on a public bridge vibration dataset to obtain common vibration features. During the fine-tuning phase, the parameters of the two-layer bidirectional LSTM layer are frozen, and only the Attention layer and the fully connected layer are trained. Road structure damage data is used to output road structure damage results.

[0158] Specifically, the specific implementation process of assigning risk scores to abnormal events through dynamic weight allocation, dividing the risk scores according to risk levels, and responding in a graded manner includes:

[0159] Based on the identification results of abnormal events, the objective weight of the indicator is calculated using the entropy weight method:

[0160]

[0161] Among them, pp uo Indicates the proportion of the u-th sample under the o-th indicator, g uo represents the index of the u-th sample after standardization under the o-th index, g Ψo represents the index of the Ψth sample after standardization under the oth index, zi represents the number of samples; e o represents the entropy value of the oth indicator; pp τo Indicates the proportion of the τth sample under the oth index, HL entropy,o represents the objective weight of the oth indicator, Oi represents the number of indicators, e η represents the entropy value of the ηth indicator;

[0162] Combined with the hierarchical analysis method to determine the subjective weight, construct the judgment matrix, compare the indicators pairwise, and obtain the relative weight HL AHP,o ;

[0163] The final weight is calculated by integrating subjective and objective weights: LGo =βHL entropy,o +(1-β)HL AHP,o , among which LG o represents the final weight of the oth indicator, and β represents the balance coefficient;

[0164] Obtain risk score based on final weight;

[0165] The risk scores are divided into responses according to the risk level, which includes low risk, medium risk, high risk and very high risk.

[0166] When the risk level is low (score <30), the relevant information is generated and stored in the database, and no alert is issued, so as to facilitate subsequent analysis and tracing;

[0167] When the risk level is medium risk (30≤score<60), the risk information will be sent to maintenance personnel in the jurisdiction via SMS push, so that they can understand the road conditions in a timely manner and take corresponding preventive measures.

[0168] When the risk level is high (60≤score<90), an audible and visual alarm is triggered, such as a road LED warning screen, to intuitively remind surrounding people to pay attention to road risks and attract the attention of relevant departments.

[0169] When the risk level is extremely high (score ≥ 90), not only will the sound and light alarm be triggered, but the traffic signal system will also be linked to restrict vehicle traffic to prevent potential dangers from further expanding, and the emergency management platform will be reported in a timely manner so that relevant departments can coordinate and deal with emergencies.

[0170] A road network abnormal event identification and risk assessment and warning system based on vibration sensing includes a sensor deployment and data acquisition module, a data preprocessing module, an abnormal event module, and a risk assessment and warning module:

[0171] The sensor deployment and data acquisition module is used to collect road vibration data. Accelerometers, seismic wave sensors, and temperature sensors are deployed to monitor the vibration signals of the road network in real time. A temperature sensor collects ambient temperature and uses an adaptive baseline correction algorithm to correct the sensor output signal. The module synchronizes the road network's vibration signals with GPS using the PTP protocol to generate road vibration data.

[0172] The data preprocessing module is used to preprocess the road vibration data by improving the Kalman filter parameter design and multi-dimensional feature extraction; the road vibration data collected by the acceleration sensor is preprocessed by Kalman filtering, and the features of the road vibration data are extracted from the time domain, frequency domain, and time-frequency domain through multi-dimensional feature extraction.

[0173] The abnormal event recognition module is used to identify abnormal events in processed road vibration data using a differentiation algorithm; detect overloaded vehicles using an optimized support vector machine; detect illegal construction using K-Medoids clustering combined with dynamic time warping; and detect road structure damage using a Bi-LSTM+Attention network.

[0174] The risk assessment and early warning module is used to assign risk scores to abnormal events through dynamic weight allocation, divide the risk scores according to the risk level and respond in a graded manner.

[0175] Specifically, the above-mentioned sensor layout and data acquisition module, data preprocessing module, abnormal event identification module and risk assessment and warning module can be embedded in a computer processing system. The computer calls the above-mentioned modules to complete the tasks of abnormal event identification and risk assessment and warning based on the above-mentioned vibration perception-based road network abnormal event identification and risk assessment and warning method; the above-mentioned sensor layout and data acquisition module, data preprocessing module, abnormal event identification module and risk assessment and warning module can perform operations according to the specific steps given in the vibration perception-based road network abnormal event identification and risk assessment and warning method.

[0176] It should be noted that it should be understood that the division of the various modules of the above system is only a division of logical functions. In actual implementation, they can be fully or partially integrated into a physical entity, or they can be physically separated. Moreover, these modules can all be implemented in the form of software called by processing elements; or they can all be implemented in the form of hardware; or some modules can be implemented in the form of processing elements calling software, and some modules can be implemented in the form of hardware. For example, the receiving module can be a separately established processing element, or it can be integrated into a chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a processing element of the above device to perform the functions of the above signal processing module. The implementation of other modules is similar. In addition, these modules can all or partly be integrated together, or they can be implemented independently. The processing element here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each of the above modules can be completed by the hardware integrated logic circuit in the processor element or by instructions in the form of software.

[0177] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application-specific integrated circuits (ASICs), one or more digital signal processors (DSPs), or one or more field programmable gate arrays (FPGAs). For another example, when a module is implemented by scheduling program code through a processing element, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processor that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0178] A terminal device includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The memory stores the computer program capable of running on the processor. When the processor loads and executes the computer program, a road network abnormal event identification and risk assessment and early warning method based on vibration perception is adopted.

[0179] It should be noted that the terminal device can be a computer device such as a desktop computer, a laptop computer or a cloud server, and the terminal device includes but is not limited to a processor and a memory. For example, the terminal device can also include input and output devices, network access devices and buses, etc.

[0180] Specifically, the processor can be a central processing unit (CPU). Of course, depending on the actual usage, other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. can also be used. The general-purpose processor can be a microprocessor or any conventional processor, etc., and this application does not impose any restrictions on this.

[0181] A storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to execute the vibration-sensing-based road network abnormal event identification and risk assessment early warning method.

[0182] Among them, the computer program can be stored in a computer-readable medium, the computer program includes computer program code, the computer program code can be in the form of source code, object code, executable file or certain middleware, etc. The computer-readable medium includes any entity or device that can carry computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that computer-readable medium includes but is not limited to the above-mentioned components.

[0183] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0184] For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. When an element is referred to as being "assembled on", "installed on", "fixed on" or "set on" another element, it can be directly on the other element or there can be a central element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there can be a central element at the same time. The terms "vertical", "horizontal", "up", "down", "left", "right" and similar expressions used herein are for illustrative purposes only and are not intended to be the only embodiment.

[0185] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

[0186] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present disclosure. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

Claims

1. A road network abnormal event identification and risk assessment early warning method based on vibration perception, characterized by: The method comprises: Collect road vibration data of the road network through sensors; Preprocessing of road vibration data through Kalman filtering and multi-dimensional feature extraction; Using a differentiation algorithm to identify abnormal events on the processed road vibration data, the abnormal events include vehicle overloading, illegal construction, and road structure damage; Abnormal events are scored for risk through dynamic weight allocation, and risk scores are divided and responded to in a graded manner according to risk levels.

2. The vibration-based road network abnormal event identification and risk assessment early warning method according to claim 1 is characterized in that: The specific implementation process of collecting road vibration data of the road network includes: The vibration signals of the road network are monitored in real time by sensors, and the vibration signals of the road network are synchronized with GPS through the PTP protocol to generate road vibration data. The sensors include acceleration sensors, seismic wave sensors and temperature sensors, among which: the temperature sensor is used to collect ambient temperature and correct the output signal of the sensor through an adaptive baseline correction algorithm.

3. The vibration-based road network abnormal event identification and risk assessment early warning method according to claim 2 is characterized in that: The specific implementation process of preprocessing the road vibration data includes: The road vibration data collected by the acceleration sensor is preprocessed using Kalman filtering, and features of the road vibration data are extracted from the time domain, frequency domain, and time-frequency domain. Dimensionality reduction is performed using principal component analysis, and the first three principal components are extracted as fusion features. Among them, the Kalman filter process includes: Construct the state vector: x k =[a k ,a' k ] T Among them, x k represents the state vector, a k Represents the real-time measurement value of the acceleration sensor, a' k Indicates the rate of change of acceleration; According to the state equation x k|k-1 =Fx k-1 Predict the current state, where x k|k-1 represents the predicted state at the current moment k, x k-1 represents the state of the previous moment k-1, and F represents the state x of the previous moment k-1. k-1 Mapped to the predicted state x at the current time k k|k-1 The state transition matrix of State transition matrix Where Δt represents the sampling interval; Predict the covariance at the current moment: P k|k-1 =FP k-1 F T +Q, where P k|k-1 represents the predicted covariance of the current moment k, P k-1 represents the covariance of the previous moment k-1, Q represents the process noise covariance matrix used to characterize the dynamic uncertainty of the system, Q = diag(0.01,0.1); Combined observations z k , calculate the Kalman gain D k =P k|k-1 H T (HP k|k-1 H T +R k ) -1 , where D k represents the Kalman gain at the current moment k, H represents the observation matrix, R k represents the observation noise covariance; The observation noise covariance R k For dynamic adjustment, the formula is: R k =αR k-1 +(1-α)(z k -Hx k|k-1 ) 2 Among them, α represents the smoothing factor, R k-1 represents the observation noise covariance of the previous moment k-1; z k represents the observation value at the current moment k; Update the state estimate x k =x k|k-1 +D k (z k -Hx k|k-1 ) and the covariance matrix P * k =(ID k H)P k|k-1 , output smoothed acceleration and acceleration change rate, where x k represents the updated state estimate, P * k represents the updated covariance, and I represents the identity matrix.

4. The vibration-based road network abnormal event identification and risk assessment early warning method according to claim 3 is characterized in that: The specific implementation process of the multi-dimensional feature extraction includes: Extract time domain features on the time axis and construct a time domain feature set. The time domain features include statistical features and waveform indicators, where statistical features include mean, peak, variance, root mean square, kurtosis and skewness, and waveform indicators include peak factor C for quantifying signal impact strength. r and the impulse factor L used to identify transient abnormal events r ; The formula for crest factor is as follows: Among them, C r represents the peak factor, Pvl represents the peak value, and Rms represents the root mean square. The formula for the pulse factor is as follows: Among them, L r represents the impulse factor, Me represents the mean; Convert the road vibration data to the frequency domain, obtain the spectrum through Fourier transform, and extract the frequency domain features: The 1 / 3 octave energy distribution is calculated by Fourier transform. The center frequency of the band covers 10-500Hz. The energy E of each band is m for: Among them, E m Indicates the energy of the mth frequency band, Band m represents the frequency band center frequency coverage of the mth frequency band, and X(f) represents the power spectrum density value of the fth frequency point of the mth frequency band; Extract the power spectrum entropy PSE of the low-frequency band that represents the signal spectrum complexity f : Among them, PSE f Represents the power spectrum entropy of the low frequency band, M represents the total number of frequency points in the low frequency band, p i represents the normalized probability distribution of the low-frequency power spectrum, E i Represents the energy of the i-th frequency point in the low frequency band, E h Represents the energy of the hth frequency point in the low frequency band; Combine time and frequency information to extract time-frequency domain features: The wavelet basis function is used to decompose the wavelet into 5 layers, and the energy proportion of the detail coefficients of each layer is extracted: Among them, RA j Indicates the ratio of the energy of the detail coefficient of the jth layer to the energy of the total detail coefficient, d j (t) represents the detail coefficient of the jth layer, s j (t) represents the energy of detail coefficients at layer j, and t represents the index of wavelet coefficients; The time-frequency energy matrix is constructed, the dimension is reduced through principal component analysis, and the first three principal components are extracted as fusion features.

5. The vibration-based road network abnormal event identification and risk assessment early warning method according to claim 4 is characterized in that: The specific implementation process of identifying abnormal events in processed road vibration data through differentiation algorithms includes: Abnormal events include overloaded vehicles, illegal construction, and damaged road structures; Based on the differences in vibration characteristics of different abnormal events, a differentiated algorithm was designed, using an optimized support vector machine to detect overloaded vehicles; illegal construction was detected through K-Medoids clustering combined with dynamic time warping; and road structure damage was detected by constructing a Bi-LSTM+Attention network. Optimized support vector machine is used to detect overloaded vehicles; Feature selection is performed on the processed road vibration data, and the root mean square value in the time domain, the energy proportion of 50-200Hz in the frequency domain, and the energy of the detail coefficient of the third layer of the wavelet are extracted as input features; The features are mapped to a high-dimensional feature space for processing through a hybrid kernel function. The specific formula of the hybrid kernel function is: Among them, xp q ,xp j represents the normalized eigenvector, xp q T xp j represents vector dot product, ||xp q -xp j || 2 represents the square of the Euclidean distance, γ poly and γ rbf They represent the weights of the polynomial kernel and the radial basis kernel respectively, and σ represents the parameter of the radial basis kernel; Find the optimal hyperplane, the specific formula is: yp im (w T xl im +b)≥1-ξ im ,ξ im ≥0,im=1,...,n Among them, w represents the normal vector of the hyperplane, b represents the intercept, C represents the penalty parameter, ξ i represents the slack variable, xl im and yp im Represent the feature vector and category label of the im-th sample respectively; im represents the total number of hyperplane samples; Deploy an online support vector machine, incrementally train the model based on newly collected road vibration data in each cycle, and update the model parameters; Detect illegal construction through K-Medoids clustering combined with dynamic time warping: Resample the low-frequency seismic wave signal and add Gaussian noise to expand the data set; Divide the data set into K clusters, select the actual points in the data set as cluster centers, use Manhattan distance as the similarity metric, and optimize the selection of cluster centers through continuous iteration to minimize the objective function; Define a construction feature template, use the dynamic time warping algorithm to obtain the optimal alignment path between the two time series, and obtain the similarity between the signal to be detected and the construction feature template; When the similarity between the signal to be detected and the construction feature template exceeds the similarity threshold, it is determined to be an illegal construction event; By building a Bi-LSTM+Attention network, we can detect road structure damage: The input layer receives the root mean square value of the time series vibration within a unit time; a two-layer bidirectional LSTM is built, with each layer containing 64 LSTM units. Weights are assigned to different time steps through the self-attention mechanism, and the output layer outputs the vibration prediction value of the future time node; Huber is used as the loss function, and its formula is: Among them, y represents the true value, represents the predicted value, δ represents the threshold parameter, δ = 0.1; Pre-training is performed based on a public bridge vibration dataset to obtain common vibration features. During the fine-tuning phase, the parameters of the two-layer bidirectional LSTM layer are frozen, and only the Attention layer and the fully connected layer are trained. Road structure damage data is used to output road structure damage results.

6. The vibration-based road network abnormal event identification and risk assessment early warning method according to claim 5 is characterized in that: The specific implementation process of assigning risk scores to abnormal events through dynamic weight allocation, dividing the risk scores according to risk levels, and responding in a graded manner includes: Based on the identification results of abnormal events, the objective weight of the indicator is calculated using the entropy weight method: Among them, pp uo Indicates the proportion of the u-th sample under the o-th indicator, g uo represents the index of the u-th sample after standardization under the o-th index, g Ψo represents the index of the Ψth sample after standardization under the oth index, zi represents the number of samples; e o represents the entropy value of the oth indicator; pp τo Indicates the proportion of the τth sample under the oth index, HL entropy,o represents the objective weight of the oth indicator, Oi represents the number of indicators, e η represents the entropy value of the ηth indicator; Combined with the hierarchical analysis method to determine the subjective weight, construct the judgment matrix, compare the indicators pairwise, and obtain the relative weight HL AHP,o ; The final weight is calculated by integrating subjective and objective weights: LG o =βHL entropy,o +(1-β)HL AHP,o , among which LG o represents the final weight of the oth indicator, and β represents the balance coefficient; Obtain risk score based on final weight; The risk score is divided into responses according to the risk level, which includes low risk, medium risk, high risk and very high risk. When the risk level is low risk, the relevant information is generated into a log and recorded in the database; When the risk level is medium, the risk information will be sent to the maintenance personnel in the jurisdiction; When the risk level is high, the sound and light alarm is triggered; When the risk level is extremely high, an audible and visual alarm will be triggered, the traffic signal system will be linked, and the information will be reported to the emergency management platform.

7. A road network abnormal event identification and risk assessment warning system based on vibration sensing, using the road network abnormal event identification and risk assessment warning method based on vibration sensing according to any one of claims 1 to 6, characterized in that: The system includes a sensor deployment and data acquisition module, a data preprocessing module, an abnormal event module, and a risk assessment and early warning module: The sensor deployment and data acquisition module is used to collect road vibration data. Accelerometers, seismic wave sensors, and temperature sensors are deployed to monitor the vibration signals of the road network in real time. Temperature sensors sense the ambient temperature and use an adaptive baseline correction algorithm to correct the sensor output signals. The module synchronizes the vibration signals of the road network with GPS using the PTP protocol to generate road vibration data. A data preprocessing module is used to preprocess the road vibration data by improving the Kalman filter parameter design and multi-dimensional feature extraction; the road vibration data collected by the acceleration sensor is preprocessed by the Kalman filter, and the features of the road vibration data are extracted from the time domain, frequency domain, and time-frequency domain by multi-dimensional feature extraction; The abnormal event recognition module is used to identify abnormal events in processed road vibration data using a differentiation algorithm; detect overloaded vehicles using an optimized support vector machine; detect illegal construction using K-Medoids clustering combined with dynamic time warping; and detect road structure damage using a Bi-LSTM+Attention network. The risk assessment and early warning module is used to assign risk scores to abnormal events through dynamic weight allocation, divide the risk scores according to the risk level and respond in a graded manner.

8. A terminal device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: The memory stores a computer program that can be run on the processor. When the processor loads and executes the computer program, the road network abnormal event identification and risk assessment and early warning method based on vibration perception according to any one of claims 1 to 6 is adopted.

9. A storage medium containing computer-executable instructions, characterized in that: When executed by a computer processor, the computer executable instructions are used to execute the road network abnormal event identification and risk assessment and early warning method based on vibration perception as described in any one of claims 1 to 6.

Images