Road disease risk intelligent early warning method based on Internet of Things

Abstract

The invention relates to the technical field of road disease early warning, in particular to a road disease risk intelligent early warning method based on the Internet of Things, and the method comprises the steps: collecting minute-by-minute road surface strain and vibration acceleration frequency spectrum data in real time through a sensing array disposed in a road structure layer, recognizing a damage accumulation acceleration section, and carrying out the early warning of the road disease risk. Constructing a strain-vibration collaborative damage section set; by calculating the damage rate difference of adjacent nodes and the offset at the starting moment, screening node clusters with synchronous responses, and generating a multi-source coupling damage mutation node cluster; combining the axis direction, the three-dimensional coordinates and the damage time sequence to construct a disease space expansion trajectory chain, and forming a road surface damage propulsion path set; and extracting a structure safety threshold based on a track end node, measuring the Euclidean distance from the structure safety threshold to a failure critical point, comparing the Euclidean distance with a mechanical extension critical distance, and screening to generate a high-risk penetration path list. According to the method, a closed loop from multi-source sensing to risk early warning is realized, and the disease identification precision and the early warning capability are improved.

Description

Technical Field

[0001] This invention relates to the field of road defect early warning technology, and in particular to an intelligent early warning method for road defect risks based on the Internet of Things. Background Technology

[0002] With the acceleration of urbanization and the continuous growth of traffic load, road infrastructure is subjected to repeated vehicle loads over long periods, leading to fatigue damage, crack propagation, rutting deformation, and other defects in the pavement structure. Traditional road maintenance relies heavily on manual inspections or post-incident repairs, which suffers from slow response times, low detection efficiency, and inaccurate defect identification, making it difficult to achieve early warning and risk prediction for hidden structural damage. In recent years, the development of Internet of Things (IoT) technology has provided new technical means for real-time monitoring of road conditions. Some studies have attempted to embed sensors inside the road structure to collect single physical parameters such as temperature, humidity, and strain to assess pavement health.

[0003] However, existing monitoring systems generally lack in-depth fusion of multi-source sensor data and systematic modeling of damage evolution mechanisms. They often focus only on static threshold alarms, failing to effectively identify the dynamic accumulation process of structural damage and its spatial expansion trend. Especially in complex traffic environments, single-parameter monitoring struggles to distinguish between normal load responses and early damage signals, leading to high false alarm rates and poor early warning reliability. Furthermore, current technology lacks a complete analytical chain from local damage identification to overall structural risk assessment, failing to achieve closed-loop management from data acquisition to risk decision-making.

[0004] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main objective of this invention is to provide an intelligent early warning method for road damage risks based on the Internet of Things, which aims to solve the technical problems of existing road damage monitoring technologies, such as the difficulty in achieving deep fusion of multi-source sensor data and dynamic tracking of damage evolution processes, resulting in inaccurate early damage identification, delayed risk warning, and lack of quantitative prediction of structural penetration risk.

[0006] To achieve the above objectives, the present invention provides an intelligent early warning method for road defect risks based on the Internet of Things, the method comprising:

[0007] Based on the IoT sensor array deployed in the road structure layer, the minute-by-minute pavement strain amplitude and vibration acceleration spectrum data of the paving location during continuous traffic load cycles are collected in real time to identify the structural damage accumulation acceleration segment and generate a set of strain-vibration synergistic damage segments.

[0008] Based on the information of each segment in the strain-vibration coordinated damage segment set, the difference in damage accumulation rate and the offset of the disease initiation time between adjacent sensing nodes are calculated, node clusters that meet the structural response synchronization conditions are screened, and multi-source coupled damage mutation node clusters are generated.

[0009] Based on each node cluster in the multi-source coupled damage mutation node cluster, the road axis direction, three-dimensional spatial coordinates of the sensing node and damage evolution time series are extracted to construct the spatial expansion trajectory chain of the disease and generate a road surface damage advancement path information set.

[0010] Based on the trajectory end node information in the road surface damage advancement path information set, the corresponding road structure safety threshold benchmark set is extracted, the spatial Euclidean distance from the path end node to the structural failure critical point is measured and compared with the mechanical expansion critical distance, the path set that meets the disease penetration condition is screened, and a list of high-risk structural penetration paths is generated.

[0011] Optionally, the strain-vibration co-damage segment set includes the cumulative strain increment, vibration energy attenuation rate, and damage acceleration duration within the damage segment; the multi-source coupled damage mutation node cluster includes the sensor node number that satisfies the structural response synchronization condition, the dynamic modal consistency index between nodes, and the road spatial continuity identifier; the road surface damage propagation path information set includes the spatial topological distance value, damage propagation time delay value, and axial propagation direction sequence of each node in the trajectory chain; the high-risk structural penetration path list includes the structural layer coordinate information of the path terminal node, the safety threshold distance measurement value, and the corresponding road functional level code.

[0012] Optionally, the steps for obtaining the strain-vibration synergistic damage segment set are as follows:

[0013] Based on the piezoelectric strain sensor array deployed at the interface between the asphalt surface layer and the base layer, the minute-by-minute pavement strain amplitude sequence of each node in the continuous traffic load cycle is collected, and dynamic sliding window integration calculation is performed on each set of strain amplitude sequences to determine the direction of the slope of the integral value change, identify the continuous positive offset segment of the slope, and generate a set of pavement strain damage accumulation segments.

[0014] Based on each segment in the set of accumulated road strain damage segments, the corresponding wheel vibration acceleration spectrum data is extracted, the frequency band energy density is integrated for each spectrum segment, the energy attenuation rate of high frequency components is calculated, key damage segments are screened by setting an attenuation rate threshold, and a vibration energy damage feature set is generated.

[0015] Based on the set of pavement strain damage accumulation segments and the set of vibration energy damage features, the damaged segments are spatiotemporally correlated. By judging the coupling relationship between strain accumulation rate and vibration energy decay rate, segments that meet the fatigue damage acceleration conditions are screened, and a set of strain-vibration synergistic damage segments is generated.

[0016] Optionally, the step of obtaining the multi-source coupled damage mutation node cluster specifically includes:

[0017] Based on the information of each segment in the strain-vibration coordinated damage segment set, the corresponding sensor node identifier, road station coordinates and the start and end times of the damage acceleration segment are extracted. The duration of the damage acceleration segment within the node is calculated. According to the spatial topological connection relationship of the sensor nodes in the road structure layer, a topological combination list between any adjacent nodes is established. The damage acceleration duration value of the nodes in the combination is called in sequence, the time difference is calculated, and a sequence of damage acceleration time difference values ​​of adjacent nodes is generated.

[0018] Based on the start time of the damage acceleration segment of neighboring nodes, the offset of the damage evolution start time between each pair of nodes is obtained, and a time offset sequence is constructed. The damage acceleration time difference sequence of neighboring nodes and the damage evolution start time offset sequence are called and compared with the set damage duration tolerance threshold and the start synchronization tolerance threshold, respectively. Node combinations that are both less than the two tolerance thresholds are selected to obtain the structural response synchronization satisfied combination index set.

[0019] The system calls the number identifier of each node combination in the combined index set to meet the structural response synchronization, extracts the road structure layer code corresponding to the node in the original damage segment information, constructs a sensor node topology cluster that meets the structural response synchronization condition, calculates and obtains the synchronization cluster strength value, and filters according to whether the cluster strength value falls within the safe strength range of the pavement structure, obtains the node combination that passes the filter, and establishes a multi-source coupled damage mutation node cluster.

[0020] Optionally, the steps for obtaining the road surface damage advancement path information set are as follows:

[0021] Based on each node cluster in the multi-source coupled damage mutation node cluster, the orientation vector, three-dimensional spatial coordinate value and damage initiation time of the sensing node on the road axis are extracted, the pairwise combination between nodes is constructed and the road axis arrangement order is determined. The spatial Euclidean distance between adjacent nodes in the combination and the time difference of damage evolution initiation are calculated to generate the axial damage expansion distance-time difference value set.

[0022] The node combination in the axial damage expansion distance-time difference set is called to determine whether the angle between the road axis direction difference vector and the Euclidean line direction is less than the spatial path consistency angle threshold. At the same time, it is determined whether the damage evolution time strictly meets the axial increasing trend. The node sequence combination that meets the dual constraints is selected to obtain the axial increasing expansion combination index set.

[0023] Based on the node combination index in the axially incremental extended combination index set, an axially connected list of damage trajectory segments is established, a road surface damage propagation path chain is constructed, and the cumulative structural layer propagation distance and response time domain window width of the trajectory segments are recorded. The maximum structural layer propagation length and the minimum response time domain window value are extracted to obtain the road surface damage propagation path information set.

[0024] Optionally, the steps for obtaining the list of high-risk structural penetration paths are as follows:

[0025] Based on the trajectory end node information in the road surface damage advancement path information set, the corresponding pavement structure layer code and end node spatial coordinates are extracted. The structural safety threshold benchmark set of the road level to which the structure layer code belongs is retrieved according to the structure layer code. The coordinate values ​​are called to calculate the spatial Euclidean distance between the end node and all critical points in the structural safety threshold benchmark set, and the distance information from the end of the path to the safety threshold point is generated.

[0026] Based on the distance information from the end of the path to the safety threshold point, the minimum distance value corresponding to each end node of the path is extracted, a safety distance sequence of end nodes is constructed, and the mechanical extension critical distance benchmark value is called to compare each value in the distance sequence. The path combination index that is less than or equal to the benchmark value is filtered to obtain the structural penetration risk path index set.

[0027] Based on the path number identified in the structural penetration risk path index set, the corresponding trajectory segment information is extracted from the road surface damage advancement path information set, the structural risk penetration path list is reconstructed and the shortest penetration distance value between the end of the path and the safety threshold point is marked to obtain the high-risk structural penetration path list.

[0028] Optionally, after obtaining the list of high-risk structural penetration paths, the method further includes:

[0029] For each path in the list of high-risk structural penetration paths, the maximum strain change amplitude within the damaged segment of the starting node of the path, the total time span of the path evolution, and the number of damage acceleration nodes in the path are extracted. The corresponding structural risk index is calculated, and the maintenance priority is marked according to the functional level of the road to which it belongs, generating a multi-dimensional road disease risk labeling result.

[0030] The multi-dimensional road defect risk assessment results include risk level labels, road function identifiers corresponding to the levels, and structural risk index codes.

[0031] Optionally, the steps for obtaining the multi-dimensional road defect risk assessment results are as follows:

[0032] Based on each path in the list of high-risk structural penetration paths, the strain change sequence of the damaged segment corresponding to the starting node of the path is extracted, the difference between the maximum value in the sequence and the minimum value of the historical baseline is identified, and normalization is performed based on the road structure layer modulus to obtain the normalized structural damage amplitude of each path and generate a path damage amplitude sequence.

[0033] The path information corresponding to the path damage magnitude sequence is called, the start and end times of each path are extracted and the time domain span is calculated, the number of sensor nodes marked as damage acceleration state in the path is counted, and a multi-parameter risk assessment set is constructed by combining three indicators and weighting them. The structural risk index value is calculated and the structural risk index value set is generated.

[0034] Based on the index corresponding to each path in the structural risk index value set, the road function level identifier of the path is retrieved. According to the preset structural safety risk classification benchmark value range within the road function level, the structural risk index is classified and judged, the corresponding maintenance priority level is marked, and the multi-dimensional road defect risk labeling results are obtained.

[0035] Furthermore, to achieve the above objectives, the present invention also provides an Internet of Things (IoT)-based intelligent early warning device for road damage risks. The device includes: a memory, a processor, and an IoT-based intelligent early warning program for road damage risks stored in the memory and executable on the processor. The IoT-based intelligent early warning program for road damage risks is configured to implement the steps of the IoT-based intelligent early warning method for road damage risks as described above.

[0036] In addition, to achieve the above objectives, the present invention also provides a medium storing an Internet of Things (IoT)-based intelligent early warning program for road damage risks, wherein when the IoT-based intelligent early warning program for road damage risks is executed by a processor, it implements the steps of the IoT-based intelligent early warning method for road damage risks as described above.

[0037] This invention provides an intelligent early warning method for road damage risks based on the Internet of Things (IoT). The method utilizes an IoT sensor array deployed within the road structure layer to simultaneously collect pavement strain amplitude and vibration acceleration spectrum data, achieving multi-dimensional, high-temporal-resolution perception of the internal damage evolution process of the road structure. By identifying strain-vibration co-damage segments and constructing a coupled analysis model of the damage accumulation rate difference and the offset of the initial time, the method effectively improves the identification accuracy of early damage signals and reduces the risk of misjudgment caused by single-parameter monitoring. Furthermore, by generating multi-source coupled damage mutation node clusters, the method enhances the sensitivity to sudden structural response anomalies, achieving precise capture from local damage to regional coordinated response. By combining three-dimensional spatial coordinates and temporal evolution sequences to construct a spatial expansion trajectory chain for damage, the method achieves visualized tracking of the path of damage development from the interior to the road surface. Finally, by introducing a dual-criteria comparison mechanism of structural safety threshold benchmark and mechanical expansion critical distance, the method quantitatively assesses the penetration risk of damage and accurately generates a list of high-risk structural penetration paths. Overall, this method breaks through the limitations of traditional monitoring methods that emphasize data collection but neglect mechanism analysis. It forms a closed loop of perception, identification, tracking, and prediction, significantly improving the intelligence level, spatial positioning accuracy, and early warning capability of road disease risk early warning, and providing scientific, real-time, and quantifiable technical support for road maintenance decision-making. Attached Figure Description

[0038] Figure 1 This is a flowchart illustrating an embodiment of the intelligent early warning method for road damage risks based on the Internet of Things according to the present invention.

[0039] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0040] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0041] Reference Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the intelligent early warning method for road damage risks based on the Internet of Things (IoT) of the present invention.

[0042] In one embodiment, the IoT-based intelligent early warning method for road defect risks includes:

[0043] Step S100: Based on the IoT sensor array deployed in the road structure layer, collect the minute-by-minute pavement strain amplitude and vibration acceleration spectrum data of the paving location in the continuous traffic load cycle, identify the structural damage accumulation acceleration segment, and generate a set of strain-vibration synergistic damage segments.

[0044] The strain-vibration synergistic damage segment set can be a collection of structural damage acceleration time periods identified from synchronously acquired pavement strain amplitude and vibration acceleration spectrum data, where the two exhibit synergistic change characteristics. This set can be used to characterize periods within the road structure where strain anomalies and vibration response distortions occur simultaneously under traffic loads, improving the robustness of early-stage damage signal identification. In this embodiment, the strain-vibration synergistic damage segment set can be obtained by detecting continuous time periods in the time or frequency domain where both strain and vibration data sequences synchronously deviate from the normal response pattern. For example, the strain-vibration synergistic damage segment set may include, but is not limited to, one or more of the following: low-frequency resonance synergistic segments, transient impact synergistic segments, and cyclic fatigue synergistic segments.

[0045] Identifying the accelerated accumulation phase of structural damage and generating a set of strain-vibration co-damage segments can be achieved by performing joint time-series analysis on synchronously acquired strain amplitude and vibration acceleration spectrum data to identify time periods where both deviate from the baseline synchronously and the rate of change continues to increase. Furthermore, this operation can be achieved by using a sliding window to calculate the cross-correlation coefficient between strain and vibration energy, marking segments exceeding a threshold and continuously increasing as co-damage segments, or by using wavelet packet decomposition to extract common high-frequency components of strain and vibration signals to identify synchronous energy surge intervals. This overcomes the problem of single parameters being easily affected by normal traffic loads and improves the signal-to-noise ratio of early damage signals.

[0046] Step S200: Based on the information of each segment in the strain-vibration co-damage segment set, calculate the difference in damage accumulation rate between adjacent sensing nodes and the offset of the disease initiation time, screen the node clusters that meet the structural response synchronization conditions, and generate a multi-source coupled damage mutation node cluster.

[0047] The multi-source coupled damage mutation node cluster can be a spatial node group that satisfies the structural response synchronization condition, consists of adjacent sensing nodes, and whose damage accumulation rate difference and initial time offset are both within a specific coupling relationship range. It can be used to achieve transitional identification from single-point anomalies to regional coordinated damage, enhancing sensitivity to sudden structural response anomalies. In an exemplary embodiment, the multi-source coupled damage mutation node cluster is generated based on a strain-vibration coordinated damage segment set, serving as the input basis for constructing the spatial expansion trajectory chain of the defect. Furthermore, the multi-source coupled damage mutation node cluster can include, but is not limited to, transverse shear-dominated node clusters, longitudinal compression-dominated node clusters, and composite stress-dominated node clusters.

[0048] The difference in damage accumulation rate between adjacent sensing nodes and the offset of the disease initiation time are calculated. Node clusters that meet the structural response synchronization conditions are screened to generate multi-source coupled damage mutation node clusters. This can be achieved by calculating the time derivative difference of the damage accumulation rate for each node within the co-damaged segment and comparing it with the time offset of the disease initiation time, retaining node combinations that meet the synchronization constraints. In a specific embodiment, this operation can be achieved by setting node pairs whose absolute value of the rate difference is less than a certain percentage threshold and whose initiation offset is less than an integer multiple of the sampling period as synchronized, or by constructing a graph neural network model, using the rate difference and time offset between nodes as edge weights to cluster into high-cohesion node clusters. This can eliminate non-cooperative local noise interference and focus on the damage response of physically related regions.

[0049] Step S300: Based on each node cluster in the multi-source coupled damage mutation node cluster, extract the road axis direction, three-dimensional spatial coordinates of the sensing node, and damage evolution time series, construct the spatial expansion trajectory chain of the disease, and generate a road surface damage advancement path information set.

[0050] The spatial expansion trajectory chain of a disease can be a path sequence formed by connecting nodes in a multi-source coupled damage mutation node cluster according to their temporal evolution order and three-dimensional spatial coordinates. It reflects the dynamic process of damage development from the interior to the road surface and can be used to achieve visualized tracking and spatial positioning of damage evolution paths, supporting quantitative analysis of the disease's advancement direction and speed. In this embodiment, the spatial expansion trajectory chain of a disease can be obtained by sorting the timestamps, axis positions, and three-dimensional coordinates of each node cluster by time and fitting them into a continuous trajectory. For example, the spatial expansion trajectory chain of a disease can include, but is not limited to, vertically upward expansion chains, oblique lateral expansion chains, and horizontal inter-layer expansion chains.

[0051] By extracting the road axis direction, the three-dimensional spatial coordinates of the sensing nodes, and the damage evolution time series, a spatial expansion trajectory chain of the damage can be constructed. This can be achieved by associating the spatial location (including depth) of each node cluster with its first appearance time and continuous evolution time, connecting them in chronological order to form a trajectory. Furthermore, this operation can be achieved by using cubic spline interpolation to smoothly fit the spatiotemporal points of discrete nodes to generate continuous trajectories, or by predicting the damage location at the next moment based on Kalman filtering and then correcting the trajectory direction using measured points. This allows for visualization and dynamic tracking of the path of damage development from within the road surface.

[0052] Step S400: Based on the trajectory end node information in the road surface damage advancement path information set, extract the corresponding road structure safety threshold benchmark set, measure the spatial Euclidean distance from the path end node to the structural failure critical point, compare it with the mechanical expansion critical distance, screen the path set that meets the disease penetration conditions, and generate a list of high-risk structural penetration paths.

[0053] The structural safety threshold benchmark set can be a pre-set set of upper limits for strain or damage indicators used to determine whether a structure is in a safe state, based on different road structural layer materials and design parameters. It can be used as one of the static criteria for assessing the risk of damage penetration and provide a reference boundary for structural failure. For example, the structural safety threshold benchmark set may include, but is not limited to, the safety threshold for asphalt surface layer, the modulus attenuation threshold for base course, and the shear strain threshold for subbase course.

[0054] The mechanical propagation critical distance, calculated based on a material fracture mechanics model, is the minimum spatial distance required for structural failure from the current damage location. It can be used as a dynamic criterion, compared with the actual Euclidean distance to determine whether the defect has the potential to penetrate. In one specific embodiment, the mechanical propagation critical distance is derived based on material toughness, interlaminar bond strength, and the current stress field distribution. Furthermore, the mechanical propagation critical distance may include, but is not limited to, the critical distance for Type I crack propagation, the critical distance for Type II shear slip, and the critical distance for mixed-mode fracture.

[0055] The spatial Euclidean distance from the end node of a path to the structural failure threshold is measured and compared with the mechanical expansion threshold distance. A set of paths that meet the disease penetration criteria is then selected. This can be achieved by calculating the straight-line distance from the end node of the trajectory chain to the nearest structural failure threshold (such as the road surface or inter-layer interface). If the distance is less than or equal to the mechanical expansion threshold distance, the path is considered a high-risk path. In an exemplary embodiment, this operation can be achieved by modeling the structure in layers, calculating the Euclidean distance for each layer and comparing it with the corresponding mechanical threshold distance, or by introducing Monte Carlo simulation to consider material parameter uncertainties and generating a probabilistic penetration risk ranking. This allows for a quantitative assessment of the disease penetration probability and the accurate generation of a list of high-risk paths requiring priority intervention.

[0056] For example, in the scenario of monitoring heavy-load lanes on urban main roads, the IoT-based intelligent early warning method for road damage risks in this embodiment can be to deploy an IoT sensor array at the junction of the asphalt surface layer and the base layer, and the system collects strain and vibration data in real time. When a 20-meter section is detected to have a sudden increase in strain amplitude and a shift in vibration spectrum energy to a lower frequency during morning and evening rush hours for several consecutive days, it is identified as a strain-vibration synergistic damage section. Further analysis reveals that the damage initiation time shift of 5 adjacent nodes in this section does not exceed 3 minutes and the cumulative rate difference is less than 15%, forming a cluster of multi-source coupled damage mutation nodes. Combining its burial depth of 0.15 meters, lateral offset of 0.8 meters, and time series, a damage expansion trajectory chain pointing obliquely upwards towards the road surface is constructed. Finally, the Euclidean distance from the end of the trajectory to the road surface is calculated to be 0.08 meters, which is less than the mechanical expansion critical distance of 0.1 meters corresponding to the asphalt mixture in this section. The system includes it in the list of high-risk structural penetration paths and pushes it to the maintenance department.

[0057] In one embodiment, the strain-vibration co-damage segment set includes the cumulative strain increment, vibration energy decay rate, and damage acceleration duration within the damage segment.

[0058] The strain-vibration co-damage segment set can be a multidimensional parameter set used to characterize the early-stage damage evolution of road structures, capturing the coupling behavior of strain and vibration response during the damage acceleration phase. In an exemplary embodiment, the cumulative strain increment reflects the degree of irreversible deformation of the material under repeated loading; the vibration energy attenuation rate characterizes the energy dissipation changes caused by structural stiffness degradation; and the duration of damage acceleration quantifies the time window for the damage to enter the rapid expansion phase. Furthermore, by fusing the dynamic correlation characteristics of the two types of physical responses, this set enables the system to effectively filter out non-damage signals caused by normal traffic disturbances.

[0059] The cluster of multi-source coupled damage mutation nodes includes sensor node numbers that meet the structural response synchronization conditions, dynamic modal consistency indices between nodes, and road spatial continuity indicators.

[0060] The multi-source coupled damage mutation node cluster can be a spatiotemporal clustering result used to identify regional collaborative anomalies, and can be used to achieve consistent discrimination of abnormal responses from multiple adjacent sensor nodes. For example, the sensor node number can uniquely identify the IoT sensing unit deployed in the road structure; the dynamic modal consistency index between nodes can reflect the similarity of each node's response mode in the frequency or time domain; and the road spatial continuity identifier can characterize the adjacency relationship and topological connectivity of nodes in three-dimensional space. In a specific embodiment, this cluster is generated by setting a synchronization criterion that the offset of the damage initiation time does not exceed a preset threshold and the cumulative rate difference is less than a limited proportion, thereby achieving the technical effect of highly sensitive identification of regional collaborative damage.

[0061] The road surface damage propagation path information set includes the spatial topological distance value of each node in the trajectory chain, the damage propagation time delay value, and the axial propagation direction sequence.

[0062] The road surface damage propagation path information set can be a data structure used to describe the dynamic trajectory of internal defects evolving towards the road surface, and can be used to map latent damage into a traceable spatial expansion path. In this embodiment, the spatial topological distance value can characterize the Euclidean distance between adjacent damaged nodes in a three-dimensional coordinate system; the damage propagation time delay value can reflect the time interval required for the defect to propagate from one node to the next; and the axial propagation direction sequence can indicate the dominant expansion trend of the defect along depth, lateral, or longitudinal directions. Furthermore, this information set, by combining three-dimensional coordinates and time series to construct a defect spatial expansion trajectory chain, makes the internal damage evolution process visualized and predictable.

[0063] The list of high-risk structural penetration paths includes the structural layer coordinates of the path's terminal nodes, safety threshold distance measurements, and corresponding road function level codes.

[0064] The high-risk structural penetration path list can be used to quantify the penetration risk of defects and support maintenance decisions. It can be used to accurately locate defects that are about to breach critical structural layers. In an exemplary embodiment, structural layer coordinate information can identify the specific location of the path endpoint in the road cross-section; the safety threshold distance measurement value can represent the actual distance of the endpoint from the road surface and be compared with the mechanical expansion critical distance; the road function level code can reflect the importance level of the road segment in the road network. Furthermore, this list is generated through a dual-criteria mechanism of structural safety threshold benchmark set and mechanical expansion critical distance, thereby achieving the technical effect of quantifying penetration risk and triggering early warning.

[0065] Taking the monitoring of heavy-load lanes on urban main roads as an example, the intelligent early warning method for road defects based on the Internet of Things in this embodiment can be to deploy an Internet of Things sensor array at the junction of the asphalt surface layer and the base layer, and the system collects strain and vibration data in real time. When a 20-meter section is detected to have a sudden increase in strain amplitude and a shift of vibration spectrum energy to low frequency during morning and evening peak hours for several consecutive days, it is identified as a strain-vibration synergistic damage section. Further analysis reveals that the damage initiation time shift of 5 adjacent nodes in this section does not exceed 3 minutes and the cumulative rate difference is less than 15%, forming a cluster of multi-source coupled damage mutation nodes. Combining its burial depth of 0.15 meters, lateral offset of 0.8 meters and time series, a defect expansion trajectory chain pointing obliquely upwards towards the road surface is constructed. Finally, the Euclidean distance from the end of the trajectory to the road surface is calculated to be 0.08 meters, which is less than the mechanical expansion critical distance of 0.1 meters corresponding to the asphalt mixture in this section. The system includes it in the list of high-risk structural penetration paths and pushes it to the maintenance department.

[0066] In one embodiment, the steps for obtaining the strain-vibration synergistic damage segment set are as follows:

[0067] Based on the piezoelectric strain sensor array deployed at the interface between the asphalt surface layer and the base layer, the minute-by-minute pavement strain amplitude sequence of each node in the continuous traffic load cycle is collected, and dynamic sliding window integration calculation is performed on each set of strain amplitude sequences to determine the direction of the slope of the integral value change, identify the continuous positive offset segment of the slope, and generate a set of pavement strain damage accumulation segments.

[0068] Dynamic sliding window integration calculation can be a signal processing method that locally accumulates and integrates time-series strain amplitudes and dynamically adjusts the window length to capture trend changes. It can be used to identify damage accumulation segments with a continuously increasing trend from minute-by-minute strain data, highlighting the progressive characteristics of fatigue damage. In this embodiment, dynamic sliding window integration calculation can set a variable-length sliding window on the strain amplitude sequence, integrate the data within the window, and calculate the slope of the change in integral values ​​of adjacent windows. When the slope is continuously positive, it is marked as a potential damage acceleration segment. For example, dynamic sliding window integration calculation can include, but is not limited to, one or more of the following: fixed-step sliding integration, adaptive window length integration, and weighted sliding integration.

[0069] For each strain amplitude sequence, a dynamic sliding window integral is performed to determine the direction of the slope of the integral value change, identify continuous positive slope offset segments, and generate a set of pavement strain damage accumulation segments. This can be achieved by applying a sliding window integral to the strain time series, calculating the first-order difference slope of the integral value sequence, and retaining intervals with continuously positive slopes as candidate segments for damage accumulation. Furthermore, this operation can be achieved by smoothing the integral values ​​using an exponentially weighted moving average window before calculating the slope, or by setting a minimum duration of continuous positive slope and a minimum slope amplitude threshold for joint determination. This effectively extracts strain responses with a continuous accumulation trend, eliminating interference from instantaneous shocks or random fluctuations.

[0070] Based on each segment in the set of cumulative road strain damage segments, the corresponding wheel vibration acceleration spectrum data is extracted, the frequency band energy density is integrated for each spectrum segment, the energy attenuation rate of high frequency components is calculated, and key damage segments are screened by setting an attenuation rate threshold to generate a vibration energy damage feature set.

[0071] The high-frequency component energy attenuation rate can be the rate at which the energy in the high-frequency band (e.g., above 500Hz) of the vibration acceleration spectrum caused by wheel excitation decreases with time or the number of load cycles. It can be used as a sensitive indicator of material stiffness degradation, reflecting the increase in structural damping and the weakening of high-frequency response caused by latent defects such as early cracks and interlaminar voids. In an exemplary embodiment, the high-frequency component energy attenuation rate can be obtained by performing spectral analysis on each segment of the vibration acceleration signal, selecting a preset high-frequency sub-band for energy density integration, and calculating the energy ratio or derivative of adjacent periods or time intervals. In a specific embodiment, the high-frequency component energy attenuation rate may include, but is not limited to, narrowband high-frequency attenuation rate, broadband high-order mode attenuation rate, and impact response high-frequency attenuation rate.

[0072] For each frequency spectrum segment, the energy density of the frequency band is integrated to calculate the energy attenuation rate of the high-frequency components. Key damage segments are screened by setting an attenuation rate threshold, generating a vibration energy damage feature set. This can be achieved by integrating the energy density within a preset high-frequency range of the vibration spectrum aligned with the strain segment, calculating its attenuation ratio or time derivative relative to the reference period, and marking segments exceeding the threshold as damage-sensitive segments. Furthermore, this operation can be implemented by extracting specific high-frequency sub-band energy using wavelet packet decomposition and then calculating its attenuation rate, or by deriving the high-frequency energy attenuation rate from model coefficients based on AR model fitting of the spectral envelope. This transforms the vibration response into a quantifiable and comparable damage-sensitive indicator, enhancing the ability to perceive material stiffness degradation.

[0073] Based on the set of pavement strain damage accumulation segments and the set of vibration energy damage characteristics, the damaged segments are spatiotemporally correlated. By judging the coupling relationship between strain accumulation rate and vibration energy decay rate, segments that meet the fatigue damage acceleration condition are screened, and a set of strain-vibration synergistic damage segments is generated.

[0074] The fatigue damage acceleration condition can be a coupled criterion of a continuously increasing strain accumulation rate and a synchronous and significant decay of high-frequency vibration energy, which can be used to define actual structural damage rather than normal load response. In this embodiment, the fatigue damage acceleration condition can rely on the synchronous anomaly judgment of the strain accumulation rate and the high-frequency component energy decay rate calculated by dynamic sliding window integration. For example, the fatigue damage acceleration condition can include, but is not limited to, one or more of the following: monotonically increasing acceleration condition, abrupt jump acceleration condition, and periodic resonance acceleration condition.

[0075] By performing spatiotemporal correlation on damaged segments and determining the coupling relationship between strain accumulation rate and vibration energy decay rate, segments that meet the conditions for accelerated fatigue damage are screened, generating a set of strain-vibration co-damaged segments. This can be achieved by aligning strain damage accumulation segments and vibration energy damage feature sets by timestamp and spatial location, retaining only segments where both show abnormal trends within the same spatiotemporal interval. Furthermore, this operation can be achieved by constructing a two-dimensional coupling matrix and setting a joint threshold region for screening, or by introducing mutual information or Granger causality tests to quantify the coupling strength between the two. This enables physical mechanism-driven fusion of multi-source heterogeneous sensor data, significantly reducing false alarm rates and improving the reliability of early-stage defect identification.

[0076] Taking the monitoring of heavy-load lanes on highways as an example, the intelligent early warning method for road damage risks based on the Internet of Things in this embodiment can be achieved by deploying a piezoelectric strain and acceleration sensor array at the interface between the asphalt surface layer and the base layer. The system collects minute-by-minute strain data of a certain node for 72 consecutive hours. Through dynamic sliding window integration, it is found that there is a continuous positive shift in the integration slope for 15 minutes after the morning rush hour each day, which is marked as the strain damage accumulation segment. Simultaneously, the vibration spectrum generated by the passing of wheels during this period is extracted, and the energy density of the 500-2000Hz frequency band is calculated. It is found that its attenuation rate is 22% compared with the same period of the previous day, exceeding the 15% threshold, and is included in the vibration energy damage feature set. After spatiotemporal alignment, it is confirmed that the strain segment and the vibration attenuation segment completely overlap, and the change trends of the two are strongly negatively correlated, which meets the fatigue damage acceleration condition. Finally, the strain-vibration synergistic damage segment is generated and the subsequent regional cluster analysis process is triggered.

[0077] In one embodiment, the step of obtaining the multi-source coupled damage mutation node cluster is as follows:

[0078] Based on the information of each segment in the strain-vibration co-damage segment set, the corresponding sensor node identifier, road station coordinates and the start and end times of the damage acceleration segment are extracted. The duration of the damage acceleration segment within the node is calculated. Based on the spatial topological connection relationship of the sensor nodes in the road structure layer, a topological combination list between any adjacent nodes is established. The damage acceleration duration value of the nodes in the combination is called in turn, the time difference is calculated, and a sequence of damage acceleration time difference values ​​of adjacent nodes is generated.

[0079] The damage acceleration time difference sequence can be an ordered set of values ​​consisting of the differences in the duration of damage acceleration in the respective strain-vibration coordinated damage segments of neighboring sensing node pairs. This sequence can be used to quantify the synchronization of the damage processes of adjacent nodes in the time dimension, and to eliminate non-coordinated responses caused by local transient disturbances. In this embodiment, the damage acceleration time difference sequence can be calculated by traversing all neighboring node pairs based on topological connectivity, calculating the difference in the duration of their damage acceleration segments, and arranging them in combination order. For example, the damage acceleration time difference sequence can include, but is not limited to, one or more of the following: lateral adjacent node time difference sequence, longitudinal inter-layer node time difference sequence, and diagonal oblique node time difference sequence.

[0080] Calculating the damage acceleration time difference between neighboring nodes and generating a sequence of neighboring node damage acceleration time differences can be achieved by traversing all neighboring node pairs based on the spatial topological connections of the sensing nodes, obtaining the duration of their damage acceleration segments, calculating the differences, and forming a sequence in combination order. Furthermore, this operation can be implemented by using Manhattan distance to measure the duration difference to suit asymmetric damage scenarios, or by introducing normalization processing to divide the time difference by the average duration of the two nodes to eliminate the influence of dimensions. This allows for a quantitative assessment of the consistency of the damage process across time scales and suppresses erroneous correlations caused by local transient interference.

[0081] Based on the start time of the damage acceleration segment of neighboring nodes, obtain the damage evolution start time offset between each pair of nodes, construct a time-series offset sequence, call the neighboring node damage acceleration time difference sequence and the damage evolution start time offset sequence, and compare them with the set damage duration tolerance threshold and the start synchronization tolerance threshold respectively, and filter the node combinations that are both less than the two tolerance thresholds to obtain the structural response synchronization satisfied combination index set.

[0082] The damage evolution initiation time offset sequence can be an ordered set of values ​​consisting of time intervals between the initiation times of their respective damage acceleration segments for neighboring sensing node pairs. This sequence can reflect the phase consistency of damage propagation or triggering in space and identify whether it originates from the same mechanical cause. In an exemplary embodiment, the damage evolution initiation time offset sequence can extract the initiation timestamps of the collaborative damage segments of each pair of neighboring nodes, calculate their absolute differences, and organize them according to topological combination order. Furthermore, the damage evolution initiation time offset sequence can include, but is not limited to, same-layer front-to-back wheel track offset sequences, cross-layer vertical transmission offset sequences, and intersecting lane lateral offset sequences.

[0083] Obtaining the offset of the damage evolution start time between each pair of nodes and constructing a temporal offset sequence can be achieved by extracting the start time of the collaborative damage segment of each pair of neighboring nodes, calculating the absolute value of their time difference, and organizing them into a sequence according to the topological combination order. In a specific embodiment, this operation can be achieved by considering traffic flow direction and using directional offsets to distinguish between upstream triggering and downstream response, or by combining GPS timestamp calibration to eliminate the influence of sensor clock drift. This enables the capture of the triggering phase relationship of damage in spatial propagation and identification of whether it originates from a shared load event or a continuous expansion process.

[0084] The time difference sequence and the initial time offset sequence are compared with tolerance thresholds respectively. Node combinations that simultaneously satisfy both types of tolerance conditions are selected to obtain a set of structural response synchronization satisfaction combination indexes. This can be achieved by setting a damage duration tolerance threshold ΔT and an initial synchronization tolerance threshold Δt0, retaining node pairs that satisfy |ΔTᵢⱼ| ≤ ΔT and |Δt0ᵢⱼ| ≤ Δt0. For example, this operation can be achieved by using an adaptive threshold to adjust the tolerance range based on traffic flow or temperature, or by introducing a fuzzy membership function to replace the hard threshold to output a synchronization probability score. Thus, through dual spatiotemporal synchronization constraints, the physical rationality and noise resistance of collaborative damage identification can be significantly improved.

[0085] The structural response synchronization satisfaction combination index set can be a set of neighboring node combination identifiers that simultaneously satisfy the damage duration tolerance threshold and the initial synchronization tolerance threshold constraints. This set can be used as an intermediate result to filter out nodes with genuine collaborative damage behavior from the original node pairs, supporting the construction of high-confidence clusters. In this embodiment, the structural response synchronization satisfaction combination index set can be generated by comparing the damage acceleration time difference sequence and the temporal offset sequence using a dual-threshold comparison, which is a prerequisite for constructing a multi-source coupled damage mutation node cluster. Furthermore, the structural response synchronization satisfaction combination index set can include, but is not limited to, strictly synchronized combination sets, loosely synchronized combination sets, and hierarchical synchronized combination sets.

[0086] The system calls the number identifier of each node combination in the combined index set to meet the structural response synchronization requirements. It extracts the road structure layer code corresponding to the node in the original damage segment information, constructs a sensor node topology cluster that meets the structural response synchronization conditions, calculates and obtains the synchronization cluster strength value, and filters the cluster strength value based on whether it falls within the safe strength range of the pavement structure. It then obtains the node combinations that pass the filtering and establishes a multi-source coupled damage mutation node cluster.

[0087] The synchronization cluster strength value can be a quantitative indicator characterizing the overall damage coordination degree of a sensing node topology cluster that meets the structural response synchronization condition. It can be used to determine whether the cluster represents genuine structural degradation rather than random noise aggregation, supporting subsequent validity screening. In a specific embodiment, the synchronization cluster strength value can be calculated based on a weighted average of parameters such as the number of nodes within the cluster, the average damage acceleration magnitude, and the time synchronization accuracy. For example, the synchronization cluster strength value may include, but is not limited to, geometric density strength value, temporal consistency strength value, and multi-source coupling energy strength value.

[0088] The node numbers and structural layer codes of synchronized combinations are extracted to construct sensor node topology clusters. Synchronous cluster strength values ​​are calculated and filtered based on whether they fall within the pavement structure's safe strength range. Multi-source coupled damage abrupt change node clusters are established. This can be achieved by clustering nodes whose structural response synchronously satisfies the combination index set into topology clusters based on spatial connectivity, calculating cluster strength values ​​based on their structural layer codes, and retaining only clusters whose strength values ​​fall outside the preset safe strength range. Furthermore, this operation can be achieved by automatically merging highly connected synchronous nodes into clusters using graph clustering algorithms, or by multi-dimensionally weighting the synchronous cluster strength values ​​using strain amplitude, vibration attenuation rate, and spatial density. This ensures that the identified node clusters reflect real structural degradation rather than accidental synchronization, improving the engineering reliability of abrupt change node clusters.

[0089] Taking the monitoring of curved sections of urban expressways as an example, the IoT-based intelligent early warning method for road defect risks in this embodiment can be implemented by deploying a sensor array on the base layer to identify five adjacent nodes with strain-vibration coordinated damage segments. The calculated duration difference between nodes A and B is 1.8 minutes, with an initial offset of 0.9 minutes; the difference between nodes B and C is 2.1 minutes, with an offset of 1.2 minutes; all other combinations exceed the threshold. With tolerance thresholds set at 2 minutes and 1.5 minutes, the combination of nodes A, B, and C is included in the structural response synchronization satisfaction combination index set. Considering that they are all located in the asphalt base layer and are spatially continuous, they are clustered into a single topological cluster. The calculated synchronization cluster strength value is 0.78, exceeding the upper limit of the pavement structure safety strength range of 0.7. Ultimately, this cluster is confirmed as a multi-source coupled damage mutation node cluster, triggering the trajectory chain construction process.

[0090] In one embodiment, the steps for obtaining the road surface damage advancement path information set are as follows:

[0091] Based on each node cluster in the multi-source coupled damage mutation node cluster, the orientation vector, three-dimensional spatial coordinate value and damage initiation time of the sensing node on the road axis are extracted. Pairwise combinations between nodes are constructed and the road axis arrangement order is determined. The spatial Euclidean distance between adjacent nodes in the combination and the time difference of damage evolution initiation are calculated to generate a set of axial damage expansion distance-time difference values.

[0092] The axial damage propagation distance-time difference set can be a coupled data set consisting of the spatial Euclidean distance between adjacent node pairs in a multi-source coupled damage mutation node cluster and the time difference between their damage initiation. This set can be used to establish the spatiotemporal correlation of damage propagation along the road axis, supporting subsequent directional and causal verification. In this embodiment, the axial damage propagation distance-time difference set can be used to combine nodes sorted by road axis pairwise, calculating their three-dimensional spatial distance and time difference to form a (distance, time difference) tuple sequence. For example, the axial damage propagation distance-time difference set can include, but is not limited to, one or more of the following: lateral offset distance-time difference set, vertical inter-layer distance-time difference set, and oblique composite distance-time difference set.

[0093] Calculating the spatial Euclidean distance between adjacent nodes and the time difference of damage evolution initiation generates a set of axial damage propagation distance-time difference values. This can be achieved by pairing nodes according to the road axis orientation vector and calculating the Euclidean distance between their three-dimensional coordinates and the difference in damage initiation time, forming a structured dataset. Furthermore, this operation can be adapted to long-distance road scenarios by using WGS84 distance calculation in a geodetic coordinate system, or by introducing depth-weighted Euclidean distance to give higher weight to vertical differences to highlight inter-layer propagation characteristics. This allows for quantitative coupling of the damage propagation process in both spatial and temporal dimensions, providing fundamental data for path rationality assessment.

[0094] Call the node combination in the axial damage expansion distance-time difference set, determine whether the angle between the road axis direction difference vector and the Euclidean line direction is less than the spatial path consistency angle threshold, and at the same time determine whether the damage evolution time strictly meets the axial increasing trend. Filter the node sequence combination that meets the dual constraints and obtain the axial increasing expansion combination index set.

[0095] The spatial path consistency angle threshold can be a preset maximum allowable angle used to determine the consistency between the Euclidean connection direction between nodes and the road axis direction. It can be used to constrain the geometric rationality of damage propagation paths and exclude non-physical propagation directions caused by sensor noise or local disturbances. In an exemplary embodiment, the spatial path consistency angle threshold can be set based on road design specifications and typical damage propagation patterns, typically taking an empirical value within the range of 15°–30°. For example, the spatial path consistency angle threshold may include, but is not limited to, a consistency angle threshold for straight sections, a dynamic angle threshold for curved sections, and a corrected angle threshold for slope sections.

[0096] The axially increasing expansion combination index set can be a set of node sequence combination identifiers that simultaneously satisfy the conditions of spatial direction consistency and strict axial increasing in time. It can be used to ensure that the selected node sequence conforms to the physical causal logic of real damage manifestation from upstream to downstream and from deep to road surface. Furthermore, the axially increasing expansion combination index set can be generated by filtering the axial damage expansion distance-time difference set through dual constraints, and serves as a direct input for constructing the road surface damage propagation path chain. In a specific embodiment, the axially increasing expansion combination index set can include, but is not limited to, one or more of the following: a unidirectional continuous expansion index set, a bifurcation and merging expansion index set, and a multi-lane intersection expansion index set.

[0097] The algorithm determines whether the angle between the road axis direction difference vector and the Euclidean line direction is less than a spatial path consistency angle threshold, and whether the damage evolution time strictly satisfies the axial increasing trend. It then filters node sequence combinations that satisfy these dual constraints, obtaining an axially increasing extended combination index set. This can be achieved by calculating the angle between the connecting line direction vector and the road axis direction vector for each node combination; if the angle is less than a preset angle threshold and the damage start time of the later node is later than that of the previous node, the combination is retained. Furthermore, this operation can be implemented by dynamically adjusting the angle threshold using a sliding window to adapt to changes in curve curvature, or by introducing a time tolerance band to handle minor synchronization errors while maintaining the overall increasing trend. This effectively eliminates false associations and ensures that the trajectory chain conforms to the true disease expansion mechanism.

[0098] Based on the node combination index in the axially incremental extended combination index set, a list of axially connected damage trajectory segments is established, a road surface damage propagation path chain is constructed, and the cumulative structural layer propagation distance and response time domain window width of the trajectory segments are recorded. The maximum structural layer propagation length and the minimum response time domain window value are extracted to obtain the road surface damage propagation path information set.

[0099] A list of axially concatenated damage trajectory segments is established based on an axially incrementally expanding combination index set. A road surface damage propagation path chain is constructed, and the maximum structural layer propagation length and the minimum response time-domain window value are extracted to obtain a road surface damage propagation path information set. This can be achieved by connecting the selected node sequences in chronological order into a continuous path chain, summing the spatial distances of each segment to obtain the total propagation length, calculating the time difference between the first and last nodes to obtain the response time-domain window, and recording the extreme values. Furthermore, this operation can be achieved by fitting discrete nodes with Bézier curves to generate a smooth trajectory chain, or by performing Dijkstra's shortest time expansion path optimization on multiple candidate paths and then extracting extreme value parameters. This reconstructs discrete damage points into quantifiable and traceable continuous evolution paths, enabling visualization and parameterization of the internal damage propagation process towards the road surface.

[0100] Taking the monitoring of long straight sections of highways as an example, the IoT-based intelligent early warning method for road damage risks in this embodiment can identify a cluster of six sensor nodes at a certain level. After sorting them by station number, the Euclidean distance and time difference between adjacent nodes are calculated to form a set of axial damage propagation distance-time difference values. Upon verification, the angles between the line connecting nodes 2→3→5 and the road axis are 12° and 18°, respectively, both less than the 20° spatial path consistency angle threshold. Furthermore, the damage initiation times are 08:15, 08:18, and 08:22, strictly increasing. This sequence is included in the axially increasing expansion combination index set. The system strings these together into a damage trajectory chain from K12+300 to K12+345, with a cumulative structural layer propagation length of 45.2 meters and a response time window of 7 minutes. Finally, the maximum propagation length and minimum window value are written into the road surface damage propagation path information set for subsequent penetration risk assessment.

[0101] This embodiment provides an intelligent early warning method for road damage risks based on the Internet of Things. It generates a set of axial damage expansion distance-time difference values ​​based on a cluster of multi-source coupled damage mutation nodes. By introducing a spatial path consistency angle threshold to constrain the rationality of the geometric direction and requiring the damage evolution time to strictly meet the axial increasing trend, it filters the axially increasing expansion combination index set. By constructing a road surface damage advancement path chain based on the index set and extracting extreme value parameters, it can achieve the technical effects of reconstructing discrete internal damage points into continuous spatiotemporal evolution trajectories, making the damage advancement process measurable and visualized, and providing high-confidence spatiotemporal input for penetration risk prediction.

[0102] In one embodiment, the steps for obtaining the list of high-risk structural penetration paths are as follows:

[0103] Based on the trajectory end node information in the road surface damage advancement path information set, the corresponding pavement structure layer code and end node spatial coordinates are extracted. The structural safety threshold benchmark set of the road level to which the structure layer code belongs is retrieved according to the structure layer code. The coordinate values ​​are called to calculate the spatial Euclidean distance between the end node and all critical points in the structural safety threshold benchmark set, and the distance information from the end of the path to the safety threshold point is generated.

[0104] The end-node safety distance sequence can be an ordered set of Euclidean distances from the end node of each damage path to the nearest critical point in the corresponding structural safety threshold benchmark set. This sequence can be used to quantify the actual spatial margin between the front end of each damage path and the structural failure boundary, serving as a direct input for penetration risk assessment. In an exemplary embodiment, the end-node safety distance sequence can be formed by calculating the Euclidean distances between the coordinates of each path's end node and all safety threshold critical points of its corresponding structural layer, and then taking the minimum value. Furthermore, the end-node safety distance sequence can include, but is not limited to, surface layer penetration margin sequences, base layer shear failure margin sequences, and subbase layer settlement boundary margin sequences.

[0105] Based on the structural layer code, the structural safety threshold benchmark set of the corresponding road layer is retrieved. The spatial Euclidean distance between the end node and the critical point is calculated, generating distance information from the end of the path to the safety threshold point. This can be achieved by matching the structural layer code of the end node of the trajectory to the corresponding structural safety threshold benchmark set, traversing all the critical point coordinates, and calculating the three-dimensional Euclidean distance with the end node. For example, this operation can accelerate the nearest neighbor search using KD-trees to improve computational efficiency in multi-critical-point scenarios, or by grouping critical points by functional partitions and calculating the distance separately to support differentiated assessment, thereby establishing a spatial correlation between the damage location and the structural failure boundary, providing a geometric measurement basis for penetration risk.

[0106] Based on the distance information from the end of the path to the safety threshold point, the minimum distance value corresponding to each end node of the path is extracted, and a safety distance sequence of end nodes is constructed. The mechanical extension critical distance benchmark value is called to compare each value in the distance sequence, and the path combination index that is less than or equal to the benchmark value is filtered to obtain the structural penetration risk path index set.

[0107] The mechanical propagation critical distance benchmark value can be derived from the material's fracture toughness, interlaminar bond strength, and current stress state. It represents the minimum theoretical spatial distance required for damage to propagate from its current location to structural failure. This value can be used as a dynamic physical criterion, compared with the measured safe distance to determine whether there is penetration potential. In a specific embodiment, the mechanical propagation critical distance benchmark value can be calculated based on linear elastic fracture mechanics or elastoplastic damage models, combined with road structure layer material parameters and load history. Furthermore, the mechanical propagation critical distance benchmark value can include, but is not limited to, the critical distance benchmark value for Type I open cracks, the critical distance benchmark value for Type II slip cracks, and the critical distance benchmark value for mixed-mode fractures.

[0108] The minimum distance value corresponding to the terminal node of each path is extracted to construct a safe distance sequence for the terminal nodes. This sequence is then compared with the mechanical expansion critical distance benchmark value. Path combinations with distances less than or equal to the benchmark value are filtered to obtain a structural penetration risk path index set. This can be achieved by taking the minimum distance from each path to the safe threshold point to form a sequence, comparing each path with the preset mechanical expansion critical distance benchmark value, and retaining the path indexes that satisfy the condition of a distance less than or equal to the benchmark value. In an exemplary embodiment, this operation can conservatively correct the mechanical expansion critical distance by introducing a safety factor to adapt to the uncertainty of material aging, or by using a probability distribution to represent the mechanical critical distance and output the penetration risk probability instead of a binary judgment. This allows for the accurate identification of paths with real penetration risk by integrating the dual criteria of geometric distance and mechanical mechanism, avoiding over-warning or under-warning caused by static thresholds.

[0109] Based on the path number identified in the structural penetration risk path index set, the corresponding trajectory segment information is extracted from the road surface damage advancement path information set, and the structural risk penetration path list is reconstructed and the shortest penetration distance value between the end of the path and the safety threshold point is marked to obtain the high-risk structural penetration path list.

[0110] Based on the structural penetration risk path index set, corresponding trajectory segment information is extracted, and a structural risk penetration path list is constructed by reorganizing the data and marking the shortest penetration distance value. This process yields a high-risk structural penetration path list, which can be obtained by retrieving complete trajectory data corresponding to the risk index from the road surface damage advancement path information set and appending the shortest penetration distance value to form a structured list. For example, this operation can embed path topology maps and 3D coordinate visualization links into the list for easy on-site positioning, or automatically raise the warning level for high-traffic road sections based on road function level codes. This outputs a set of high-confidence risk paths that can be directly used for maintenance decisions, supporting sorting by penetration urgency and prioritizing resource allocation.

[0111] For example, in the scenario of monitoring the approach lanes at intersections of urban arterial roads, the IoT-based intelligent early warning method for road damage risks in this embodiment can identify a damage trajectory chain whose end is located at the base layer at K5+210, with the structural layer coded as "cement-stabilized crushed stone base layer". Based on this, the system retrieves the structural safety threshold benchmark set for this layer, which includes a shear failure critical surface 0.25 meters from the road surface. The Euclidean distance from the end node (buried depth 0.18 meters) to this critical surface is calculated to be 0.07 meters, constituting one item in the end node safety distance sequence. The mechanical extension critical distance benchmark value corresponding to this road segment's base layer is 0.08 meters. Since 0.07 ≤ 0.08, this path is included in the structural penetration risk path index set. The system extracts its complete trajectory information, marks the shortest penetration distance as 0.07 meters, and, considering that this road segment is a dedicated bus lane (high functional level), includes it in the high-risk structural penetration path list and pushes it to the maintenance platform.

[0112] In one embodiment, after obtaining the list of high-risk structural penetration paths, the method further includes:

[0113] For each path in the list of high-risk structural penetration paths, the maximum strain change amplitude within the damaged segment of the starting node of the path, the total time span of the path evolution, and the number of nodes in the path where damage accelerates are extracted. The corresponding structural risk index is calculated, and the maintenance priority is marked according to the functional level of the road to generate multi-dimensional road disease risk assessment results.

[0114] The multi-dimensional road defect risk assessment results include risk level labels, road function identifiers corresponding to the levels, and structural risk index codes.

[0115] The structural risk index is a quantitative indicator calculated based on the maximum strain change amplitude at the path starting point, the total time span of path evolution, and the number of damage acceleration nodes. It is used to characterize the overall severity of degradation of a single high-risk penetration path. In this embodiment, the structural risk index can fuse multidimensional damage characteristics into a unified and comparable risk measure, supporting priority ranking among different paths. Furthermore, the structural risk index can be a weighted synthesis of three original indicators after normalization, with weights set based on historical disease development data or expert experience. For example, the structural risk index may include, but is not limited to, one or more of the following: intensity-dominated risk index, velocity-dominated risk index, and breadth-dominated risk index.

[0116] Road function levels can be administrative or technical classifications based on a road's capacity, service targets, and importance within the urban transportation network. These classifications serve as external constraints for prioritizing maintenance, ensuring that technical risk assessments match operational value. In one exemplary embodiment, road function levels may include, but are not limited to, urban expressways, arterial roads, secondary arterial roads, and local roads. Multi-dimensional road defect risk assessment results can be structured output information sets containing risk level labels, corresponding road function identifiers, and structural risk index codes. These sets provide directly analyzable technical-management joint criteria for maintenance decisions, enabling risk visualization and resource scheduling linkage. In a specific embodiment, multi-dimensional road defect risk assessment results can be generated jointly from the structural risk index and road function level, serving as the final output of the early warning system. For example, multi-dimensional road defect risk assessment results can take the form of high-risk assessment results for primary arterial roads, medium-risk assessment results for secondary arterial roads, and low-risk assessment results for local roads.

[0117] The structural risk index is calculated by extracting the maximum strain change amplitude within the damaged segment at the starting node of the path, the total temporal span of the path evolution, and the number of nodes in the path where damage accelerates. This can be achieved by obtaining the peak strain change from the damaged segment at the starting point of a high-risk path, calculating the time difference between the first and last nodes of the path as the total temporal span, and counting the total number of nodes marked as having accelerated damage. These three factors are then standardized and weighted to generate the structural risk index. Furthermore, this operation can automatically determine the optimal weight combination of the three indicators using principal component analysis, or introduce a nonlinear function to correct the risk attenuation of the temporal span, highlighting rapidly evolving paths. This integrates the three dimensions of local intensity, evolution speed, and regional breadth to form a comprehensive risk measure with physical interpretability.

[0118] Maintenance priorities are assigned based on the functional level of the roads they belong to, generating multi-dimensional road defect risk assessment results. This can be achieved by mapping the road functional level of each high-risk path segment to its structural risk index, and then mapping this index to preset maintenance priority rules, generating a structured result containing risk level labels, functional identifiers, and index codes. Furthermore, this operation can establish a priority matrix, with the horizontal axis representing the risk index range and the vertical axis representing the functional level. Maintenance levels can be determined by looking up tables, or priority rules can be dynamically adjusted. For example, the level of specific road segments can be temporarily upgraded based on seasonal traffic flow or major events. This allows for coordinated decision-making regarding engineering and technical risks and traffic maintenance value, avoiding excessive intervention in high-risk, low-value paths or delayed response in low-risk, high-value paths.

[0119] Taking the comparative assessment of main roads and peripheral roads in the core urban area as an example, the IoT-based intelligent early warning method for road defect risks in this embodiment can identify two high-risk penetration paths: Path A is located on the main urban road, with a maximum strain change amplitude of 180 microstrains at the starting point, a total time span of 5 days, including 7 acceleration nodes, and a structural risk index of 0.82; Path B is located on a peripheral road, with an amplitude of 150 microstrains, a span of 12 days, including 4 nodes, and an index of 0.61. Based on the road functional level, the main road corresponds to the "emergency" maintenance priority, and the peripheral road corresponds to "routine". The system generates multi-dimensional road defect risk labeling results: Path A is labeled as "high risk - main road - 0.82", and Path B is labeled as "medium risk - peripheral road - 0.61". Based on this, the maintenance platform prioritizes the scheduling of milling and repaving resources to the road section where Path A is located.

[0120] In one embodiment, the steps for obtaining the multi-dimensional road defect risk assessment results are as follows:

[0121] Based on each path in the list of high-risk structural penetration paths, the strain change sequence of the damaged segment corresponding to the starting node of the path is extracted, the difference between the maximum value in the sequence and the minimum value of the historical baseline is identified, and the normalization is performed based on the road structure layer modulus to obtain the normalized structural damage amplitude of each path and generate a path damage amplitude sequence.

[0122] The normalized structural damage amplitude can be the difference between the maximum strain change amplitude in the damaged segment at the starting node of the path and the minimum historical baseline value. This dimensionless index, normalized by the road structural layer modulus, can be used to eliminate the influence of different material stiffnesses on the original strain amplitude, making the damage intensity comparable across different structural layers or road segments. In this embodiment, the normalized structural damage amplitude can be obtained by calculating the difference between the peak value of the strain sequence and the minimum long-term operating baseline value, and then dividing by the dynamic modulus or reference elastic modulus of the corresponding structural layer. For example, the normalized structural damage amplitude can include, but is not limited to, one or more of the following: asphalt surface layer normalized amplitude, base layer modulus normalized amplitude, and composite layer equivalent normalized amplitude.

[0123] The strain variation sequence of the damaged segment at the starting node of the path is extracted, the difference between the maximum value and the historical baseline minimum value is identified, and normalized based on the road structural layer modulus to obtain the normalized structural damage amplitude. This can be achieved by extracting the strain time series from the damaged segment at the starting node, calculating the difference between the maximum value and the long-term operating baseline minimum value, and then dividing by the modulus parameter of the structural layer where the node is located. Furthermore, this operation can be achieved by using a dynamic modulus instead of a static modulus to reflect temperature and frequency dependence, or by using a sliding window historical baseline instead of a global minimum value to adapt to seasonal changes. This transforms the original strain amplitude into a physically consistent damage intensity index, supporting cross-structural layer comparisons.

[0124] The path information corresponding to the path damage magnitude sequence is called, the start and end times of each path are extracted and the time domain span is calculated, the number of sensor nodes in the path that are identified as damage acceleration state is counted, and a multi-parameter risk assessment set is constructed by combining three indicators and weighting, and the structural risk index value is calculated to generate a set of structural risk index values.

[0125] The multi-parameter risk assessment set can be a set of structured assessment vectors formed by combining three indicators—normalized structural damage amplitude, path temporal span, and number of damage acceleration nodes—with preset weights. This set can be used as an intermediate data structure for calculating the structural risk index, integrating damage intensity, evolution velocity, and spatial breadth. In an exemplary embodiment, the multi-parameter risk assessment set can be input to the structural risk index calculation module, and the output can be a set of structural risk index values. Exemplary examples include, but are not limited to, intensity-velocity coupled assessment sets, breadth-intensity dominated assessment sets, and balanced multi-parameter assessment sets.

[0126] The structural risk index set can be a collection of structural risk index values ​​corresponding to each high-risk path, reflecting the overall deterioration degree of each path. It can provide a quantifiable risk basis that can be ranked and thresholded, supporting subsequent classification and priority labeling. In a specific embodiment, the structural risk index set can be generated by performing a weighted summation or nonlinear mapping operation on each vector in the multi-parameter risk assessment set. For example, the structural risk index set may include, but is not limited to, a linearly weighted index set, an exponentially decaying index set, or a piecewise threshold mapping index set.

[0127] A multi-parameter risk assessment set is constructed by weighting three indicators: normalized structural damage amplitude, temporal span, and number of damage acceleration nodes. The structural risk index value is then calculated. This can be achieved by standardizing the three indicators and summing them according to a preset weight vector, or by inputting them into a pre-trained regression model and outputting a single structural risk index value. Furthermore, this operation can automatically determine the objective weights of the three indicators using the entropy weight method, or introduce a sigmoid function to nonlinearly compress the index, enhancing the discriminative power of high-risk intervals. This allows for the fusion and quantification of multi-dimensional heterogeneous damage characteristics, forming a unified risk measure.

[0128] Based on the index corresponding to each path in the structural risk index value set, the road function level identifier of the path is retrieved. According to the preset structural safety risk classification benchmark value range within the road function level, the structural risk index is classified and judged, the corresponding maintenance priority level is marked, and the multi-dimensional road defect risk labeling results are obtained.

[0129] The preset structural safety risk grading benchmark value range within the road functional level can be a range of structural risk index grading thresholds set separately for different road functional levels (such as arterial roads, secondary arterial roads, etc.). This can be used to achieve differentiated management of risk tolerance, enabling high traffic value road sections to adopt stricter safety standards. In this embodiment, the preset structural safety risk grading benchmark value range within the road functional level can be determined based on historical maintenance records, traffic load, and statistical analysis of structural failure cases, and can be configured independently according to the functional level. For example, the preset structural safety risk grading benchmark value range within the road functional level can include, but is not limited to, the Level 3 risk range for arterial roads (0.7–1.0 is high), the Level 3 risk range for secondary arterial roads (0.8–1.0 is high), and the Level 3 risk range for local roads (0.85–1.0 is high).

[0130] Based on the preset structural safety risk grading benchmark value range within the road functional level, the structural risk index is graded and judged, maintenance priority levels are marked, and multi-dimensional road defect risk labeling results are obtained. This can be achieved by calling the corresponding risk grading benchmark range according to the road functional level to which the path belongs, mapping the structural risk index value to "high / medium / low" levels, and assigning maintenance priority labels. Furthermore, this operation can be achieved by using fuzzy membership degrees instead of hard interval division to output risk probability distribution, or by dynamically adjusting the grading threshold in conjunction with real-time traffic flow to improve emergency response capabilities. This allows for precise alignment of technical risks and operational value, avoiding one-size-fits-all decisions.

[0131] Taking the risk classification comparison between urban arterial roads and secondary arterial roads as an example, the IoT-based intelligent early warning method for road defect risks in this embodiment can process two high-risk paths: Path X is located on an arterial road, with a normalized structural damage amplitude of 0.65, a time span of 4 days, 6 acceleration nodes, and a structural risk index of 0.78; Path Y is located on a secondary arterial road, with an amplitude of 0.70, a span of 6 days, 5 nodes, and an index of 0.76. The risk classification benchmark for arterial roads is [0.7, 1.0] as high risk, and for secondary arterial roads it is [0.8, 1.0] as high risk. Therefore, Path X is marked as "high risk - emergency maintenance", and Path Y is marked as "medium risk - planned maintenance". The system generates multi-dimensional road defect risk labeling results, and the maintenance department prioritizes the handling of Path X accordingly.

[0132] Furthermore, to achieve the above objectives, the present invention also provides an Internet of Things (IoT)-based intelligent early warning device for road damage risks. The device includes: a memory, a processor, and an IoT-based intelligent early warning program for road damage risks stored in the memory and executable on the processor. The IoT-based intelligent early warning program for road damage risks is configured to implement the steps of the IoT-based intelligent early warning method for road damage risks as described above.

[0133] In addition, to achieve the above objectives, the present invention also provides a medium storing an Internet of Things (IoT)-based intelligent early warning program for road damage risks, wherein when the IoT-based intelligent early warning program for road damage risks is executed by a processor, it implements the steps of the IoT-based intelligent early warning method for road damage risks as described above.

[0134] Other embodiments or specific implementations of the Internet of Things-based intelligent early warning device for road damage risks described in this invention can be found in the above-described method embodiments, and will not be repeated here.

[0135] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for intelligent early warning of road damage risks based on the Internet of Things, characterized in that, The method includes: Based on the IoT sensor array deployed in the road structure layer, the minute-by-minute pavement strain amplitude and vibration acceleration spectrum data of the paving location during continuous traffic load cycles are collected in real time to identify the structural damage accumulation acceleration segment and generate a set of strain-vibration synergistic damage segments. Based on the information of each segment in the strain-vibration coordinated damage segment set, the difference in damage accumulation rate and the offset of the disease initiation time between adjacent sensing nodes are calculated, node clusters that meet the structural response synchronization conditions are screened, and multi-source coupled damage mutation node clusters are generated. Based on each node cluster in the multi-source coupled damage mutation node cluster, the road axis direction, three-dimensional spatial coordinates of the sensing node and damage evolution time series are extracted to construct the spatial expansion trajectory chain of the disease and generate a road surface damage advancement path information set. Based on the trajectory end node information in the road surface damage advancement path information set, the corresponding road structure safety threshold benchmark set is extracted, the spatial Euclidean distance from the path end node to the structural failure critical point is measured and compared with the mechanical expansion critical distance, the path set that meets the disease penetration condition is screened, and a list of high-risk structural penetration paths is generated.

2. The intelligent early warning method for road defect risks based on the Internet of Things as described in claim 1, characterized in that, The strain-vibration co-damage segment set includes the cumulative strain increment, vibration energy attenuation rate, and damage acceleration duration within the damage segment; the multi-source coupled damage mutation node cluster includes the sensor node number that satisfies the structural response synchronization condition, the dynamic modal consistency index between nodes, and the road spatial continuity identifier; the road surface damage propagation path information set includes the spatial topological distance value, damage propagation time delay value, and axial propagation direction sequence of each node in the trajectory chain; the high-risk structural penetration path list includes the structural layer coordinate information of the path terminal node, the safety threshold distance measurement value, and the corresponding road functional level code.

3. The intelligent early warning method for road defect risks based on the Internet of Things as described in claim 2, characterized in that, The specific steps for obtaining the strain-vibration synergistic damage segment set are as follows: Based on the piezoelectric strain sensor array deployed at the interface between the asphalt surface layer and the base layer, the minute-by-minute pavement strain amplitude sequence of each node in the continuous traffic load cycle is collected, and dynamic sliding window integration calculation is performed on each set of strain amplitude sequences to determine the direction of the slope of the integral value change, identify the continuous positive offset segment of the slope, and generate a set of pavement strain damage accumulation segments. Based on each segment in the set of accumulated road strain damage segments, the corresponding wheel vibration acceleration spectrum data is extracted, the frequency band energy density is integrated for each spectrum segment, the energy attenuation rate of high frequency components is calculated, key damage segments are screened by setting an attenuation rate threshold, and a vibration energy damage feature set is generated. Based on the set of pavement strain damage accumulation segments and the set of vibration energy damage features, the damaged segments are spatiotemporally correlated. By judging the coupling relationship between strain accumulation rate and vibration energy decay rate, segments that meet the fatigue damage acceleration conditions are screened, and a set of strain-vibration synergistic damage segments is generated.

4. The intelligent early warning method for road defect risks based on the Internet of Things as described in claim 3, characterized in that, The specific steps for obtaining the multi-source coupled damage mutation node cluster are as follows: Based on the information of each segment in the strain-vibration coordinated damage segment set, the corresponding sensor node identifier, road station coordinates and the start and end times of the damage acceleration segment are extracted. The duration of the damage acceleration segment within the node is calculated. According to the spatial topological connection relationship of the sensor nodes in the road structure layer, a topological combination list between any adjacent nodes is established. The damage acceleration duration value of the nodes in the combination is called in sequence, the time difference is calculated, and a sequence of damage acceleration time difference values ​​of adjacent nodes is generated. Based on the start time of the damage acceleration segment of neighboring nodes, the offset of the damage evolution start time between each pair of nodes is obtained, and a time offset sequence is constructed. The damage acceleration time difference sequence of neighboring nodes and the damage evolution start time offset sequence are called and compared with the set damage duration tolerance threshold and the start synchronization tolerance threshold, respectively. Node combinations that are both less than the two tolerance thresholds are selected to obtain the structural response synchronization satisfied combination index set. The system calls the number identifier of each node combination in the combined index set to meet the structural response synchronization, extracts the road structure layer code corresponding to the node in the original damage segment information, constructs a sensor node topology cluster that meets the structural response synchronization condition, calculates and obtains the synchronization cluster strength value, and filters according to whether the cluster strength value falls within the safe strength range of the pavement structure, obtains the node combination that passes the filter, and establishes a multi-source coupled damage mutation node cluster.

5. The intelligent early warning method for road defect risks based on the Internet of Things as described in claim 4, characterized in that, The specific steps for obtaining the road surface damage advancement path information set are as follows: Based on each node cluster in the multi-source coupled damage mutation node cluster, the orientation vector, three-dimensional spatial coordinate value and damage initiation time of the sensing node on the road axis are extracted, the pairwise combination between nodes is constructed and the road axis arrangement order is determined. The spatial Euclidean distance between adjacent nodes in the combination and the time difference of damage evolution initiation are calculated to generate the axial damage expansion distance-time difference value set. The node combination in the axial damage expansion distance-time difference set is called to determine whether the angle between the road axis direction difference vector and the Euclidean line direction is less than the spatial path consistency angle threshold. At the same time, it is determined whether the damage evolution time strictly meets the axial increasing trend. The node sequence combination that meets the dual constraints is selected to obtain the axial increasing expansion combination index set. Based on the node combination index in the axially incremental extended combination index set, an axially connected list of damage trajectory segments is established, a road surface damage propagation path chain is constructed, and the cumulative structural layer propagation distance and response time domain window width of the trajectory segments are recorded. The maximum structural layer propagation length and the minimum response time domain window value are extracted to obtain the road surface damage propagation path information set.

6. The intelligent early warning method for road defect risks based on the Internet of Things as described in claim 5, characterized in that, The specific steps for obtaining the list of high-risk structural penetration paths are as follows: Based on the trajectory end node information in the road surface damage advancement path information set, the corresponding pavement structure layer code and end node spatial coordinates are extracted. The structural safety threshold benchmark set of the road level to which the structure layer code belongs is retrieved according to the structure layer code. The coordinate values ​​are called to calculate the spatial Euclidean distance between the end node and all critical points in the structural safety threshold benchmark set, and the distance information from the end of the path to the safety threshold point is generated. Based on the distance information from the end of the path to the safety threshold point, the minimum distance value corresponding to each end node of the path is extracted, a safety distance sequence of end nodes is constructed, and the mechanical extension critical distance benchmark value is called to compare each value in the distance sequence. The path combination index that is less than or equal to the benchmark value is filtered to obtain the structural penetration risk path index set. Based on the path number identified in the structural penetration risk path index set, the corresponding trajectory segment information is extracted from the road surface damage advancement path information set, the structural risk penetration path list is reconstructed and the shortest penetration distance value between the end of the path and the safety threshold point is marked to obtain the high-risk structural penetration path list.

7. The intelligent early warning method for road defect risks based on the Internet of Things as described in claim 6, characterized in that, After obtaining the list of high-risk structural penetration paths, the process also includes: For each path in the list of high-risk structural penetration paths, the maximum strain change amplitude within the damaged segment of the starting node of the path, the total time span of the path evolution, and the number of damage acceleration nodes in the path are extracted. The corresponding structural risk index is calculated, and the maintenance priority is marked according to the functional level of the road to which it belongs, generating a multi-dimensional road disease risk labeling result. The multi-dimensional road defect risk assessment results include risk level labels, road function identifiers corresponding to the levels, and structural risk index codes.

8. The intelligent early warning method for road defect risks based on the Internet of Things as described in claim 7, characterized in that, The specific steps for obtaining the multi-dimensional road defect risk assessment results are as follows: Based on each path in the list of high-risk structural penetration paths, the strain change sequence of the damaged segment corresponding to the starting node of the path is extracted, the difference between the maximum value in the sequence and the minimum value of the historical baseline is identified, and normalization is performed based on the road structure layer modulus to obtain the normalized structural damage amplitude of each path and generate a path damage amplitude sequence. The path information corresponding to the path damage magnitude sequence is called, the start and end times of each path are extracted and the time domain span is calculated, the number of sensor nodes marked as damage acceleration state in the path is counted, and a multi-parameter risk assessment set is constructed by combining three indicators and weighting them. The structural risk index value is calculated and obtained, and a structural risk index value set is generated. Based on the index corresponding to each path in the structural risk index value set, the road function level identifier of the path is retrieved. According to the preset structural safety risk classification benchmark value range within the road function level, the structural risk index is classified and judged, the corresponding maintenance priority level is marked, and the multi-dimensional road defect risk labeling results are obtained.

9. A smart early warning device for road damage risks based on the Internet of Things, characterized in that, The device includes: a memory, a processor, and an Internet of Things (IoT)-based intelligent early warning program for road damage risks stored in the memory and executable on the processor, wherein the IoT-based intelligent early warning program for road damage risks is configured to implement the steps of the IoT-based intelligent early warning method for road damage risks as described in any one of claims 1 to 8.

10. A medium, characterized in that, The medium stores an Internet of Things (IoT)-based intelligent early warning program for road damage risks. When the IoT-based intelligent early warning program for road damage risks is executed by a processor, it implements the steps of the IoT-based intelligent early warning method for road damage risks as described in any one of claims 1 to 8.

Images