A roadbed and pavement embedded sensing fusion vehicle-road coordination r2x early warning method and system

Abstract

The embodiment of the application provides a kind of roadbed pavement embedded sensing fusion vehicle-road coordination R2X early warning method and system, it is related to road early warning technical field, fuses the response of road structure under the collection of vehicle excitation of multiple buried depth embedded sensing node of roadbed pavement, constructs vehicle-road coupling event, relies on abnormal response inversion road internal risk center and influence area, can identify roadbed, hidden structural disease of base course, make up the defect that traditional early warning cannot locate road internal risk.Combining with the dynamic risk state calculation individual risk when vehicle arrives, R2X early warning strategy is issued as needed, realizes vehicle grading accurate early warning.Target vehicle executes R2X early warning measure and passes through the road structure response after risk area, loop corrects risk range and early warning strategy, continuously optimizes subsequent risk identification and early warning process, improves the reliability of hidden road structure risk identification precision and vehicle-road coordination early warning.

Description

Technical Field

[0001] This application relates to the field of road early warning technology, specifically to a method and system for early warning of vehicles and roads using R2X (Road-to-Road Cooperative) technology with embedded sensors in the roadbed and pavement. Background Technology

[0002] With the development of intelligent transportation, intelligent connected vehicles, and vehicle-road cooperative technologies, road traffic safety control has gradually shifted from a passive management approach relying on driver observation, fixed traffic signs, and manual road patrols to an active early warning approach based on roadside perception, vehicle perception, wireless communication, and traffic control platforms. Existing vehicle-road cooperative systems typically acquire road environment, traffic flow, and vehicle operation information through devices such as cameras, millimeter-wave radar, weather sensors, roadside units (RSUs), and onboard terminals. Based on the location of accidents, construction, congestion, severe weather, or other abnormal events ahead, they send warning information to vehicles approaching the corresponding road area. This type of technology can improve the speed of road traffic information transmission, providing conditions for drivers or autonomous driving systems to take measures such as slowing down or changing lanes in advance.

[0003] In the prior art, patent document CN112735165B discloses a method where, after a vehicle enters the communication range of a roadside unit, the onboard unit receives a message about an abnormal road event ahead. As the vehicle continues to travel, it determines whether the warning output trigger condition is met based on factors such as the dynamic distance between the vehicle and the abnormal event, the effective time of the abnormal event message, and the vehicle's direction of travel. If the trigger condition is met, the corresponding warning information is output. This solution solves the problem of premature broadcasting when the vehicle is still far from the abnormal event, allowing the warning information to be output at a more appropriate location and time. However, in this solution, the abnormal road event is mainly obtained by the cloud platform, MEC device, or roadside unit as already formed event information. Its technical focus is on the transmission of the abnormal event message and the triggering of the warning, without involving multi-depth perception of the road surface layer, base layer, and subgrade internal structural state using vehicle wheel-mounted excitation, nor does it involve inverting hidden risk sources within the road based on cross-layer propagation changes in the road structure response.

[0004] For example, patent document CN113034952B discloses a real-time traffic safety early warning system for road sections based on vehicle-road cooperation. This system includes a roadside system, an on-board system, a data center, a decision-making system, and a display system. It acquires traffic environment information through roadside radar and video equipment, and obtains vehicle position, speed, acceleration, and vehicle-road friction coefficient information through the on-board system. This information is then fused in the data center. The decision-making system assesses the risk of single-vehicle accidents and multi-vehicle accidents separately, thereby obtaining the road section accident risk level and disseminating it to drivers and road managers. While this solution achieves the fusion analysis of roadside and vehicle information, its risk identification targets mainly include vehicle skidding, inter-vehicle traffic conflicts, and comprehensive road section accident risks. The road information used is still primarily focused on the road surface condition and traffic operating environment, making it difficult to directly determine structural risks that are not yet clearly manifested on the road surface, such as subgrade delamination, roadbed softening, and localized internal subsidence, and their specific spatial locations within the road.

[0005] Therefore, although existing vehicle-road cooperative early warning technologies can issue warnings based on abnormal road events, vehicle status, or traffic flow status, they still have difficulty in accurately identifying internal road risks and providing dynamic early warnings for vehicles. Summary of the Invention

[0006] In view of this, this application provides a method and system for early warning of vehicle-road cooperative R2X with embedded sensing fusion in the roadbed and pavement, so as to solve the problem that it is difficult to achieve accurate identification of internal road risks and dynamic early warning for vehicles in the prior art.

[0007] In a first aspect, embodiments of this application provide a method for early warning of vehicle-road cooperative R2X (Road-to-Road) systems with embedded sensing fusion in the roadbed and pavement, including: When a vehicle passes through a target road unit, it acquires the road structure response of sensor nodes at different burial depths within the road unit and associates the responses of the same vehicle to form road unit status information. The vehicle's operating status is obtained, and the vehicle's operating status is correlated with the road unit's status information over time to form a vehicle-road interaction event. Based on the abnormal response, risk events are determined, and the risk center and risk impact area are obtained through inversion. For vehicles approaching the risk impact zone, predict the road risk status upon their arrival and determine individual risks, then generate and issue R2X warning information and strategies. The system collects road structure responses after target vehicles implement R2X warning measures and pass through risk zones, corrects risk centers, risk impact areas, and R2X warning strategies, and uses this data for subsequent risk identification and warning work.

[0008] In one possible implementation, when the vehicle passes through the target road unit, it acquires the road structure response of sensing nodes at different burial depths within the road unit, and associates the responses of the same vehicle to form road unit state information, including: The target road is divided into multiple continuous sensing road units along the direction of travel. Embedded sensing nodes are deployed at different depths in the surface layer, base layer and subgrade of each sensing road unit, and a three-dimensional road space coordinate mapping relationship is established for each embedded sensing node. Once a vehicle enters the current sensing road unit, the continuous road structure responses belonging to the same vehicle are correlated in time according to the order in which multiple sensing nodes arranged along the vehicle's direction of travel are excited by the wheel load, thus forming the road body response fingerprint of the corresponding vehicle. By associating the road response fingerprints formed by the continuous passage of different vehicles within the same perceived road unit, a normal propagation benchmark is established, and the changes in real-time response relative to the normal propagation benchmark are transformed into road unit state information.

[0009] In one possible implementation, the step of acquiring the vehicle's operating status, temporally correlated with the road unit status information to form a vehicle-road interaction event, and determining risk events based on abnormal responses and inverting to obtain the risk center and risk impact area includes: Acquire multiple operational states of the target vehicle throughout the entire process of entering and exiting the current road unit, align them with the structural response time collected by the embedded sensor nodes, and construct a complete chain of continuous vehicle-road interaction events. For the same road location, the inherent road risks, abnormal vehicle loads, and coupled risks are distinguished based on different abnormal responses; To address the inherent risks of the road, spatial information is retrieved based on the response propagation distortion inversion of sensor nodes at different burial depths to determine the risk center and impact zone within the road.

[0010] In one possible implementation, determining the internal risk center and affected area of ​​the road based on the spatial information retrieved by the response propagation distortion inversion of sensor nodes at different burial depths, addressing the inherent risks of the road, includes: For any adjacent sensing nodes and Calculate the first Cross-node propagation distortion under vehicle-excited conditions ; Robust statistics are performed on the cross-node propagation distortions corresponding to the most recent N consecutive vehicle excitations to obtain robust propagation distortions for node pairs. ; Using any candidate spatial point inside the road As the risk source to be inverted, a risk source space inversion function is constructed: in: Indicates that it is located at a candidate point The set of sensor node pairs surrounding and participating in the current inversion; This represents the path consistency term between node pairs and candidate point Q. This represents the spatial propagation direction constraint between the candidate risk source and the corresponding sensor node pair.

[0011] Traverse the candidate spatial locations within the current abnormal road unit and calculate... Candidate positions that will cause the spatial inversion function to reach its maximum value. This has been identified as the current internal risk center of the road. The risk impact zone is determined based on the spatial location around the risk center that meets the preset inversion function decay conditions.

[0012] In one possible implementation, the process of predicting the road risk status upon arrival of vehicles approaching the risk impact zone and determining individualized risks, generating and issuing R2X early warning information and strategies, includes: Based on the internal risk center and risk impact area of ​​the road, the road body response generated when subsequent vehicles pass through is continuously acquired, a road abnormal state function is established, and a road risk state that evolves over time is formed. Predict the arrival time of the target vehicle in the risk impact area, and calculate the individualized real-time traffic risk value by combining the abnormal road conditions, the vehicle's equivalent dynamic action, and the warning margin. The risk value is compared with the classification threshold to generate R2X early warning control messages, and early warning information is issued in layers according to the remaining arrival time of the vehicle.

[0013] In one possible implementation, the predicted arrival time of the target vehicle in the risk impact zone, combined with abnormal road conditions, the vehicle's equivalent dynamic action, and the warning margin, calculates an individualized real-time traffic risk value, including: For a target vehicle about to enter the risk impact zone, calculate the predicted arrival time of the target vehicle in the risk impact zone and the equivalent dynamic effect on the road risk area; Based on the road anomaly status, road risk development speed, equivalent dynamic action, and warning time margin at the expected arrival time of the target vehicle, the spatiotemporal risk input status of the vehicle road is formed. Subsequently, a vehicle-road coupled risk function is constructed using the degree of road anomaly when the target vehicle is expected to arrive at the risk area, the strength of the dynamic load exerted by the target vehicle on the current road risk area, and the time urgency of the target vehicle's approach to the risk area: in: Indicates vehicle For risk areas Individualized real-time traffic risk values, This refers to the road conditions when the target vehicle is expected to reach the risk area. For road change enhancements, This represents the vehicle's equivalent dynamic action. Let r be the allowable equivalent dynamic action of the road region. To accelerate road development, The urgency of the vehicle's arrival.

[0014] In one possible implementation, the comparison of risk values ​​with grading thresholds to generate R2X early warning control messages, and the tiered distribution of early warning information based on the remaining arrival time of the vehicles, includes: Based on vehicle-specific real-time risk values The relationship between the warning level and multiple risk level thresholds determines the warning level; Information is distributed in tiers based on the estimated remaining arrival time of vehicles from the risk impact zone. Trend warnings are sent to vehicles that are far away, behavioral suggestions are sent to vehicles that are approaching, and mandatory emergency alerts are sent to vehicles that are about to enter the risk impact zone.

[0015] In one possible implementation, the data collection of the target vehicle's road structure response after executing R2X early warning measures and passing through the risk zone is used to correct the risk center, risk impact area, and R2X early warning strategy, and is then used for subsequent risk identification and early warning work, including: After the target vehicle implements R2X warning measures, the road structure response generated by its passage after the warning is implemented is collected and compared with the historical structure response of similar vehicles without control. The effectiveness of the early warning strategy is evaluated based on the improvement of the response, and the reliability of the positioning is evaluated based on the stability of the risk center and risk impact area obtained from the inversion of continuous vehicle events. Continuously record the road structure response after early warning is executed under multiple working conditions, construct multi-element mapping relationships, and support the selection of early warning strategies for similar risk areas.

[0016] In one possible implementation, the step of evaluating the effectiveness of the early warning strategy based on response improvement and assessing the reliability of location based on the stability of the risk center and risk impact area obtained from continuous vehicle event inversion includes: If the propagation distortion or abnormal road response in the road risk area is significantly reduced after the target vehicle implements R2X warning measures, the effectiveness evaluation of the current warning strategy will be improved. If the location of the risk center and the risk impact area obtained from the inversion of multiple consecutive vehicle events remain stable, the reliability of the location of the risk center and the risk impact area will be improved. If the road anomaly response does not change according to the predicted trend after the target vehicle implements the warning measures, the risk center and risk impact area will be re-determined based on the road structure response generated by the next vehicle passing by.

[0017] Secondly, embodiments of this application provide a roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning system, including: The acquisition module acquires the road structure response of sensor nodes at different burial depths within the road unit when a vehicle passes through the target road unit, and associates the responses of the same vehicle to form road unit status information. The risk zone identification module is used to obtain the operating status of the vehicle, correlate the operating status of the vehicle with the status information of the road unit in time to form a vehicle-road interaction event, determine the risk event based on the abnormal response, and inversely obtain the risk center and risk impact area. The R2X warning module is used to predict the road risk status when vehicles approach the risk impact area arrive and determine individual risks, and generate and issue R2X warning information and strategies. The verification and optimization module is used to collect the road structure response after the target vehicle executes the R2X warning measures and passes through the risk zone, correct the risk center, risk impact area and R2X warning strategy, and use it for subsequent risk identification and warning work.

[0018] Thirdly, embodiments of this application provide an electronic device, including: processor; Memory; And a computer program, wherein the computer program is stored in the memory, the computer program including instructions that, when executed by the processor, cause the electronic device to perform the method described in any one of the first aspects.

[0019] Fourthly, embodiments of this application provide a computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform the method described in any one of the first aspects.

[0020] In this embodiment, the road structure response under vehicle excitation is collected by integrating multi-depth embedded sensor nodes in the roadbed and pavement. This constructs a vehicle-road coupling event, and relies on anomaly responses to invert the internal risk center and impact zone of the road. This allows for the identification of hidden structural defects in the roadbed and base layer, overcoming the limitation of traditional early warning systems that cannot locate internal road risks. Individualized risks are calculated based on the dynamic risk status at the vehicle's arrival time, and R2X early warning strategies are issued on demand, achieving precise, tiered early warnings for vehicles. Simultaneously, the road structure response after the target vehicle executes R2X early warning measures and passes through the risk zone is collected, allowing for closed-loop correction of the risk range and early warning strategy, continuously optimizing subsequent risk identification and early warning processes. This solution establishes a complete link between road structure perception, risk inversion, vehicle-road cooperative early warning, and strategy iteration, improving the accuracy of hidden road structure risk identification and the reliability of vehicle-road cooperative early warning. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating a roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning method provided in an embodiment of this application; Figure 2 A schematic diagram illustrating the principle of risk source spatial inversion provided in this application embodiment; Figure 3 This application provides a scatter plot for risk event identification in its embodiments. Figure 4 The vehicle individualized risk curve provided in the embodiments of this application; Figure 5 A schematic diagram of the framework of a roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning system provided in this application embodiment; Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application; Detailed Implementation To better understand the technical solution of this application, the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0023] It should be understood that the described embodiments are merely some, not all, of the embodiments in this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0024] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0025] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0026] See Figure 1 The roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning method provided in this embodiment includes: S101, when a vehicle passes through a target road unit, the road structure response of sensor nodes at different burial depths within the road unit is obtained, and the responses of the same vehicle are associated to form road unit status information.

[0027] In this embodiment, the longitudinal length of each sensing road unit can be set according to the road grade, vehicle speed and sensor coverage, for example, 20 to 100m. There can be an overlap of 0 to 10m between adjacent sensing road units to prevent road risks from being missed when they are exactly at the boundary between two sensing road units.

[0028] For each perceived road unit, a three-dimensional road coordinate system is established in the longitudinal, lateral, and depth directions of the road, where: the normal driving direction of the vehicle is set as the x-direction; the direction along the cross section of the road is set as the y-direction; and the direction from the road surface to the interior of the roadbed is set as the z-direction.

[0029] No. i The positions of the embedded sensor nodes are denoted as: in: Let be the longitudinal coordinate of the i-th sensor node on the road. The horizontal coordinates of the road; This refers to the burial depth of the sensor node relative to the road surface.

[0030] like Figure 2As shown, sensor nodes are respectively installed inside the road surface layer, base layer, and subgrade. Each sensing road unit has at least two different burial depths, and each burial depth has at least two sensor nodes arranged at intervals along the longitudinal direction of the road. When a three-dimensional spatial inversion of road internal risks is required, the embedded sensor nodes participating in the inversion are also distributed at least two different lateral positions, so that the sensor nodes participating in the inversion have spatial positional differences in the longitudinal, lateral, and depth directions of the road; when the sensor nodes are only deployed in a single lateral position, a longitudinal-depth two-dimensional spatial inversion of road internal risks is performed. This ensures that subsequent spatial inversion of road internal risk sources can be performed based on the propagation relationship between nodes. The sensor nodes can include one or more of the following, depending on the road structural characteristics: vibration sensors, strain sensors, pressure sensors, temperature sensors, humidity sensors, or water content sensors.

[0031] It should be noted that this invention does not require every node to have all sensing functions simultaneously. For example, vibration sensing nodes can be used as the main dynamic propagation detection nodes, with strain, pressure, or moisture content detection functions added to some of these nodes to help determine whether the anomaly is due to mechanical abnormalities in the road structure or a decrease in load-bearing capacity caused by changes in environmental conditions. Each sensing node pre-stores a unique node number and its corresponding three-dimensional coordinates. After road construction or sensor node installation is completed, the location of each node is calibrated, and the node number, road station number, lane number, lateral position, and burial depth are written into the roadside edge device.

[0032] When a target vehicle enters the sensing road unit, the road surface is excited by the vehicle's wheel load, and the resulting vibration, strain, and pressure responses propagate from the road surface to the base layer and subgrade. Roadside edge devices continuously collect sensing signals from each embedded sensing node according to a unified time reference. Preferably, the sampling frequency of vibration, pressure, and dynamic strain signals is determined based on the dominant frequency of the road structure response; it is only necessary to ensure that the sampling frequency is higher than twice the highest effective frequency of the road response to be analyzed. Slow variables such as temperature, humidity, or water content can use a lower sampling frequency than the dynamic signals.

[0033] Vehicle identity and operational status can be obtained from RSU, vehicle OBU, road camera equipment, millimeter-wave radar, or other vehicle-to-infrastructure (V2I) devices. In cases where a unique vehicle identity cannot be obtained, a temporary vehicle event number can be constructed based on the temporal continuity of the vehicle passing through adjacent nodes, vehicle speed, and lane position; this approach is also feasible.

[0034] For the k-th vehicle or the k-th time it passes through the road sensing unit, establish a vehicle event window: in: This indicates the moment when the target vehicle enters the sensing road unit. Indicates the moment when the target vehicle leaves the sensing road unit; and These represent the extended sampling time before the vehicle enters and after it leaves, respectively.

[0035] In the vehicle event window Within each node, extract the peak response value of the i-th node. Response arrival time Effective response duration and response waveform The arrival time of the response can be determined by the moment when the signal first exceeds the mean of the background noise at the corresponding node plus three times the standard deviation of the background noise, or by the maximum cross-correlation value, the energy mutation point, or other methods that can stably obtain the starting position of the response. The response duration can be defined as the effective duration for which the response signal is continuously higher than the background noise threshold.

[0036] For nodes i and j with longitudinal spacing along the vehicle's direction of travel, when calculating the propagation delay of the structural response between nodes, the time required for the vehicle to travel from the longitudinal position corresponding to node i to the longitudinal position corresponding to node j is first deducted based on the actual speed of the target vehicle to obtain the corrected propagation delay: in, Let be the projected distance between node i and node j along the vehicle's direction of travel. Let be the average speed of the vehicle when it passes through the road segment corresponding to the two nodes in the kth iteration. In the k-th vehicle event, the moment when node i detects the start of the road structure response; For the k-th vehicle event, the time when node j detects the start of the road structure response. For cross-layer node pairs with the same longitudinal position or a longitudinal spacing less than a preset threshold, the measured response arrival time difference can be directly used. Subsequent calculations involving actual propagation delay between nodes, health baseline propagation delay, and equivalent propagation speed all use the propagation delay corrected for vehicle movement time. For the same vehicle event k, the response peak value, response arrival time, response duration, normalized response waveform, and sensor node spatial location of different nodes are associated according to a unified vehicle event number to form the road response fingerprint corresponding to that vehicle.

[0037] To avoid directly comparing raw sensor values ​​from different nodes, this embodiment first establishes a healthy road propagation baseline. This baseline is based on at least [data points] collected during the initial normal operation of the road. The set of vehicle passage events, including those without obvious road abnormalities and with normal vehicle operation, is considered the health baseline event set. Preferably, For any pair of nodes that can form an effective propagation relationship... Based on the set of health baseline events, the baseline propagation delay from node i to node j in a healthy state is obtained: Peak health benchmark for node i: Health duration benchmark for node i: Using the median instead of a single measurement reduces the impact of occasional heavy-duty vehicles and sensor impulse noise on the health baseline. To minimize the impact of vehicle type and operating condition differences on the health propagation baseline, health baseline events are preferably grouped according to vehicle type, vehicle speed range, and available vehicle mass or axle load range, and health propagation baselines that match the current target vehicle's operating conditions are prioritized.

[0038] During long-term road operation, the health baseline can be updated gradually using only vehicle events that have been subsequently verified as normal through closed-loop testing, instead of directly updating the health baseline using data that has been identified as risk events. This avoids abnormal road conditions being incorrectly absorbed into the normal baseline. At this point, each road unit has basic road condition data including load stability, vibration propagation, water content, and local deformation.

[0039] S102, obtain the vehicle's operating status, correlate the vehicle's operating status with the road unit's status information in time to form a vehicle-road interaction event, determine the risk event based on the abnormal response, and inversely obtain the risk center and risk impact area.

[0040] For each vehicle event k obtained in S101, the vehicle's operating state is further obtained before entering the sensing road unit, during the vehicle's passage through the sensing road unit, and after the vehicle leaves the sensing road unit. The vehicle operating state includes at least the vehicle speed. Longitudinal acceleration The system includes lateral acceleration, the vehicle's lane, vehicle type, and available vehicle mass or axle load information. The vehicle's operating state is matched with the obtained road structure response using a unified timestamp, forming a continuous vehicle-road interaction event chain consisting of "vehicle entering state → vehicle wheel load excitation → multi-layered road response → vehicle leaving state".

[0041] For multiple consecutive vehicle events, determine whether the abnormal road response consistently occurs in similar spatial locations. If, when different vehicles pass by consecutively, the anomaly repeatedly occurs in the same or adjacent sensor node areas, and the spatial location of the anomaly does not change with the vehicle's identity, then the anomaly is initially identified as an inherent road risk. If the anomaly mainly accompanies a specific vehicle, and the anomaly excitation location moves forward with that vehicle, but subsequent vehicles do not exhibit similar anomalies when passing the same road location, then the anomaly is initially identified as a vehicle load anomaly. If a fixed road anomaly occurs simultaneously with sudden braking, rapid acceleration, significant lateral deviation, or large lateral acceleration of a vehicle, then it is identified as a vehicle-road coupling risk.

[0042] In this embodiment, inherent road risks are confirmed through repeated events from at least three different vehicles to reduce false alarms caused by sporadic responses from a single vehicle. Inherent road risks, abnormal vehicle loads, and vehicle-road coupling risks are initially distinguished based on the aforementioned abnormal spatial location and vehicle operating status. Figure 3 This diagram illustrates the distribution of different types of risk events in the feature space comprised of the propagation distortion variable and the individualized risk value after subsequent individualized risk calculations. Figure 3 The individualized risk value shown is not a necessary condition for determining the risk event type in stage S102. Figure 3 The scatter plots are used to illustrate the distribution principle of different risk events in the feature space, and do not limit the fixed numerical range of actual roads.

[0043] In this embodiment, to avoid inconsistencies in dimensions caused by the direct combination of different physical quantities, the abnormal characteristics are dimensionless before calculating the propagation distortion.

[0044] For node i, node j, and the kth vehicle event, define the amplitude propagation anomaly: Define duration distortion variable: Define the propagation delay deviation: in: , , These are small positive numbers with the same dimensions as the corresponding physical quantities, used to avoid the denominator being zero; It is a dimensionless small positive number. In actual implementation, Can be taken ~ However, the specific values ​​do not constitute a limitation on the present invention.

[0045] Subsequently, the amplitude of the two node response waveforms is normalized by dividing each waveform by its respective root mean square value, resulting in: in: This is the current analysis window length. To prevent a zero response from resulting in a small positive number divided by zero.

[0046] The normalized waveform of node j is propagated using the propagation delay corrected for vehicle motion time, according to its response propagation delay relative to node i. Perform time shifting and calculate waveform distortion: Based on this, this embodiment adopts the following dimensionless propagation distortion variable: This formula does not linearly add the four anomalies with fixed weights. Instead, it utilizes the multiplicative coupling relationship between amplitude and duration changes to form a basic anomaly, which is then nonlinearly amplified by propagation delay deviation and waveform distortion. Therefore, propagation distortion only occurs when multiple propagation anomalies occur simultaneously between adjacent nodes. Only then will it increase significantly.

[0047] For the most recent N consecutive passing events, take... As a robust propagation distortion variable for the current node pair. Preferably, N is 3 to 10. This can reduce the instantaneous large fluctuations caused by a single abnormal vehicle. Subsequently, a candidate risk source search space is established for the sensing road units with persistent propagation anomalies. The candidate space can be defined as follows: the longitudinal range of the road where the current abnormal node is located extends forward and backward by a certain distance; the lateral range covers the current lane and adjacent areas; and the depth range extends from the road surface to the predetermined roadbed detection depth. The candidate space is divided into multiple candidate points: The search intervals for candidate points in the longitudinal, lateral, and depth directions can be set to 0.1–1 m, 0.1–0.5 m, and 0.05–0.5 m, respectively. The smaller the search interval, the higher the spatial positioning resolution, but the computational load increases accordingly.

[0048] For any candidate point Q, select only the pair of nodes that have a valid propagation association with that candidate point to form a set. Specifically, node i and node j are added when the following conditions are met. First, node i and node j are not located at the exact same spatial coordinates; second, the distance between at least one of node i and node j and candidate point Q is less than the preset influence radius. Third, a stable inter-node reference propagation delay can be obtained in the set of health reference events. .

[0049] For node pairs The equivalent propagation speed is not directly assigned a value based on the theoretical material wave velocity, but is determined using measured data from healthy roads: in: The distance is the three-dimensional Euclidean distance between the two sensing nodes. This represents the equivalent structural response propagation speed from node i to node j under normal road structure conditions. Therefore, it already implies the actual propagation impact caused by differences in pavement surface layer, base layer and subgrade materials, avoiding ambiguity in inversion conditions due to different theoretical wave velocities of different road materials.

[0050] For any candidate point Q, calculate the spatial propagation direction constraint between the candidate risk source and the corresponding sensor node pair: in: Indicates the distance from candidate point Q to the sensing node and The shortest spatial distance between the propagation paths; Indicates a distance-sensitive scale; The minimum directional contribution coefficient, with a value greater than 0 and less than 1. The closer candidate point Q is to the effective propagation path between node i and node j, the better. The larger the value, the greater the inversion contribution of candidate points that are more consistent with the spatial location of the anomaly propagation path.

[0051] Then, the propagation path consistency term between the node pair and the candidate point Q is calculated: in: It is a small positive number with the same dimension as the distance, used to avoid the denominator being too small when the distance between nodes is too small; This represents the propagation delay after vehicle motion time correction for the most recent N consecutive vehicle events. The median, i.e.: Further construct the spatial inversion function after normalizing the number of nodes: Compared with the aforementioned direct product form, this embodiment adds exponential normalization, so that even if different candidate points Q correspond to different numbers of effective node pairs, their spatial inversion functions are still comparable, avoiding the situation where a candidate point has a larger number of nodes participating in the calculation and thus naturally obtains a larger inversion function value.

[0052] After traversing the candidate space, determine ,Will The optimal risk center within the current road is identified, such as... Figure 2 As shown. As a criterion for determining the spatial boundary of the risk-affected zone. Among them, The value can be between 0.5 and 0.8, and in this embodiment, 0.6 is preferred. (This will be compared with...) All candidate points that are spatially connected and meet the above conditions constitute the risk impact zone inside the road, and are projected onto the road surface to obtain the risk road area r in the sense of vehicle passage.

[0053] according to depth coordinates Furthermore, by examining the road structure layers in which they are located, it can be determined whether the risk is primarily located in the pavement surface layer, base layer, or subgrade. This completes the transformation from anomalies at discrete sensor nodes to risk sources within the road's internal space.

[0054] S103, for vehicles approaching the risk impact zone, predicts the road risk status upon their arrival and determines individualized risks, generating and issuing R2X warning information and strategies.

[0055] After obtaining the risk region r, a set is formed for all node pairs whose propagation paths pass through or are adjacent to the risk region r in the most recent N consecutive vehicle events. Calculate the intensity of the propagation anomaly in the current area: Meanwhile, based on the number of events in the most recent N vehicle events where the propagation distortion exceeds the 95th percentile of the corresponding health baseline... Calculate the abnormality duration coefficient: in, The closer it is to 1, the higher the probability that the road anomaly will recur during the passage of multiple vehicles.

[0056] Define the abnormal growth relative to the previous time window: And construct dimensionless road anomaly states: As can be seen from the above form This avoids the problem of inconsistent risk values ​​across different road units. For sporadic single-vehicle incidents that have not yet formed a continuous anomaly, due to... Smaller, corresponding It won't reach a high level directly.

[0057] To obtain the road conditions when the target vehicle is expected to arrive at the risk area, this embodiment uses the most recent M time windows. Calculate the rate of change of road risk. Use the most recent M historical point pairs: Perform robust slope estimation to obtain risk evolution coefficients. .

[0058] Then predict the target vehicle's passage. Road status after the time: When historical data is insufficient for stable estimation season This means using the current road condition as the short-term prediction value.

[0059] For a target vehicle m about to enter the risk area r, obtain the longitudinal distance from its current position to the entrance of the risk area. Current speed Current longitudinal acceleration Traffic density in front of risk areas And the spatial gradient of average vehicle speed along the road direction. To ensure that traffic density and speed gradient in the formula are dimensionless variables, a reference traffic density is first set. Reference length and reference speed .

[0060] in: And define a dimensionless velocity gradient: Therefore, the traffic condition correction coefficient is constructed: The more congested the traffic, Increased speed reduces the vehicle's ability to maintain its current acceleration; when the speed of the vehicle ahead changes significantly along the road direction... Increased, used to reflect the impact of speed fluctuations on arrival time.

[0061] definition: Considering the target vehicle may be decelerating, in order to prevent This results in the square root having no real solutions when When the target vehicle continues to operate according to its current speed and deceleration, and is expected to stop before reaching the risk area, the above arrival time formula is not used. Instead, the target vehicle is marked as an expected stopped vehicle or a queued vehicle. The estimated arrival time is recalculated after the target vehicle's speed or acceleration is updated. The estimated arrival time of the target vehicle in the risk area is calculated using the following formula: When the vehicle is traveling at approximately a constant speed, that is At that time, the above formula naturally approximates: Therefore, the formula can be applied to vehicles moving at a constant speed, accelerating, and decelerating simultaneously. If the vehicle has already stopped and If the speed is lower than the preset minimum operating speed, the above arrival time formula will not be used. Instead, the vehicle will be marked as stationary or in a queue, and the calculation will be repeated after the vehicle starts moving again to prevent the divisor from being too small.

[0062] The equivalent dynamic action of the vehicle on the road is further calculated based on the vehicle's mass, speed, acceleration, and rate of change of acceleration. Let the equivalent mass of the target vehicle m be... When the actual mass can be obtained through the weighing system or the vehicle itself, the actual mass is used directly; when the actual mass cannot be obtained, the pre-stored equivalent mass value for the vehicle type is called according to the vehicle type. Therefore, it is not required that all vehicles be equipped with dynamic weighing devices.

[0063] The rate of change of longitudinal acceleration of a vehicle is denoted as: Set reference time: And define a dimensionless impact correction term: Among them, molecules The dimension of the fraction is acceleration / time, the dimension of the denominator is acceleration, and the dimension of the ratio of the two is the reciprocal of time. Then multiply by... This results in dimensionless parameters, thus solving the problem of directly relating the rate of change of acceleration to acceleration or... The problem of dimensional inconsistency caused by division.

[0064] The equivalent dynamic action of the vehicle is expressed as: in, This is a dimensionless scale adjustment parameter. Because... and Having the same units, they can therefore be combined in the same expression. It is not a direct physical measurement of the actual tire impact energy, but an equivalent index used to compare the strength of road dynamics between different vehicles.

[0065] Establish an allowable equivalent dynamic action quantity for road area r. In the historical data of road health status, the time for normal vehicles to pass is calculated. The distribution, and using, for example, the 95th percentile as a reference dynamic factor for the healthy path. .

[0066] When the road risk status increases, the following settings can be configured: in, This serves as a reference dynamic action for a healthy road condition. This represents the minimum permissible positive dynamic action when road risk is high, and ; This refers to the predicted road conditions when the target vehicle is expected to reach the risk area. Therefore, when hour, As road abnormalities increase, Gradually decrease and approach .

[0067] The road conditions when the target vehicle is expected to reach the risk area are recorded as follows: A dimensionless expression for the speed of road development: in, This is a preset reference time for risk changes.

[0068] Define the urgency of vehicle arrival: in, Indicates according to The minimum advance warning time for the pre-determined abnormality level of the road. In calculating individualized real-time traffic risk values It was confirmed earlier.

[0069] Indicates a time-sensitive scale. When At that time, the vehicle was close to the risk area. Increase; when At that time, the vehicle was still far from the risk area. This reduces the risk. This approach avoids the situation where the risk correction increases with the distance of the warning.

[0070] Further construct road change enhancements: When road anomalies show a decreasing trend, the rate of improvement should not be used to amplify individualized risks. Risk amplification should only be applied when road anomalies show a sustained increasing trend. and Increase the risk value.

[0071] Finally, the individualized real-time risk value of vehicle m for road risk area r is calculated: in: Therefore, the risk values ​​corresponding to different vehicles, roads, and times can be uniformly mapped to the range of 0 to 1. This formula reflects three interdependent factors: first, the degree of road anomaly when the target vehicle is expected to reach the risk area; second, the strength of the dynamic load exerted by the target vehicle on the current road risk area; and third, the time urgency of the target vehicle's approach to the risk area. There is no fixed-weight linear additive relationship among these factors.

[0072] like Figure 4 As shown, for the same risk area with base layer delamination or subgrade softening, due to heavy vehicles... and corresponding Typically larger, their individualized risk values ​​can enter a higher risk level earlier than those of smaller vehicles. Figure 4 This is used to illustrate the trend of change, not to limit the actual road to forming a fixed curve as shown in the diagram.

[0073] In this embodiment, three incremental risk thresholds are set: Risk thresholds can be determined using historical vehicle events on healthy roads and roads identified as having risks. Specifically, the high quantile of the risk value for vehicles in healthy condition can be used as the threshold. Refer to and determine based on the risk value distribution corresponding to confirmed events requiring deceleration and emergency traffic control. and .

[0074] During the initial operation phase of a system lacking historical calibration data, the following settings can be configured as an example: Further adjustments were made gradually.

[0075] when At that time, mandatory traffic warnings will not be implemented; only general road condition information will be provided to vehicles, or no information will be provided at all. At that time, the RSU sends an advance deceleration suggestion and recommended speed to the target vehicle. When this happens, a higher-level warning is issued, and further suggestions such as speed limits, lane changes, or increasing following distance are sent based on the conditions of adjacent lanes. In such cases, an emergency R2X warning is sent, and high-risk event information is simultaneously sent to road variable message signs or regional traffic control platforms. For autonomous vehicles or vehicles with advanced driver assistance capabilities, R2X information can serve as road risk constraint inputs for the vehicle planning and control system; for ordinary connected vehicles, it can be converted into driver-understandable prompts for deceleration, lane changing, and maintaining distance.

[0076] S104 collects the road structure response after the target vehicle executes R2X warning measures and passes through the risk zone, corrects the risk center, risk impact area and R2X warning strategy, and uses it for subsequent risk identification and warning work.

[0077] After receiving an R2X warning, if the target vehicle takes measures such as deceleration, lane changing, or other safe passage measures, the embedded sensor node continues to operate. Once the vehicle actually enters and passes through the risk area r, a corresponding vehicle event is generated again. The peak response value, propagation distortion, abnormal duration, and actual speed and acceleration of the vehicle after the warning are recorded as post-warning response data. A comparison is established with historical events prior to the warning for the same or similar vehicle models, road conditions, and traffic conditions.

[0078] Define the propagation distortion improvement rate as the vehicle passes in front and behind: in: This indicates the propagation distortion of similar vehicles before or during the implementation of the current warning strategy; This represents the propagation distortion variable corresponding to the current vehicle's implementation of warning measures. If This indicates that the overall abnormal dynamic response decreased after traffic control was implemented.

[0079] If different vehicles consistently achieve positive improvement after receiving the same warning, the effectiveness of the current warning strategy is enhanced. If the risk center obtained from the inversion of continuously added vehicle events remains stable within the same spatial neighborhood, and the change in the risk impact area is less than the preset range, the reliability of the current risk center and risk impact area location is improved. If the anomaly propagation still stably appears in the same road location after a vehicle significantly decelerates, it indicates that the anomaly is more likely due to a structural risk in the road itself, rather than being caused by a single vehicle.

[0080] Conversely, if an anomaly corresponds to only a single sensor node, and there are no corresponding propagation anomalies in adjacent nodes, the anomaly cannot be reproduced in at least three subsequent vehicles, and the node's background noise or zero-point drift has significantly increased, then that node is marked as a suspected sensor anomaly node and temporarily removed from [the list of nodes]. After excluding the faulty sensor, the risk source inversion is re-executed. This further prevents false alarms about persistent road risks caused by a single sensor malfunction.

[0081] When a new vehicle event causes a change in the maximum position of the spatial inversion function, the process is re-executed. and redetermine the satisfaction The risk impact zone.

[0082] The traffic control platform stores every complete closed-loop event, including: risk area number, risk center. abnormal road conditions Vehicle type, vehicle equivalent mass, vehicle speed before entry, vehicle equivalent dynamic action. Individualized risk values ​​for vehicles The actual issuance of early warning measures, whether vehicles comply with the warnings, the propagation distortion after vehicles pass, and the risk improvement rate. .

[0083] When similar road conditions and similar vehicle types reappear, priority will be given to calling historically available methods. Early warning strategies that achieve a large positive value without generating new vehicle operation risks.

[0084] For example, before a heavy vehicle enters the target sensing road unit, the RSU obtains the vehicle's current position, speed, vehicle type, and current lane. After the vehicle enters the sensing road unit, multiple sensing nodes P1 to P6 in the pavement surface layer, base layer, and subgrade sequentially generate structural responses. The roadside edge device classifies these responses into a single vehicle event and obtains the node response peak value, response arrival time, duration, and normalized waveform for each.

[0085] As an illustrative example, six sensing nodes are arranged at different lateral, longitudinal, and depth locations within the target sensing road unit. When the road is in a healthy state, small and heavy vehicles pass through, collecting normal vehicle events and establishing a health propagation baseline. Subsequently, a base layer void simulation anomaly centered at (25.0, 1.6, 0.65) m is established within the target sensing road unit, and small and heavy vehicles again pass through, collecting road structure responses and calculating... , Finally obtained =(24.7,1.5,0.68) m, with a positioning error of 0.32 m. Subsequently, the R values ​​for light vehicles and heavy vehicles were calculated to be 0.46 and 0.78, respectively. Corresponding to different warning levels, after heavy vehicles decelerated, the median of the robust propagation distortion of the node pairs corresponding to the risk area decreased from 0.62 to 0.48. According to the aforementioned propagation distortion improvement rate formula, the improvement rate was calculated to be approximately 22.6%.

[0086] If the propagation relationships near nodes P2, P4, and P6 deviate from the healthy propagation baseline for multiple consecutive vehicle events, then the propagation distortion variable for each node is calculated. Subsequently, candidate risk source points Q are established within the three-dimensional structural space of the corresponding road unit, and spatial inversion functions are calculated for all candidate points. If the maximum value is located in the lower part of the base course or the upper part of the roadbed, then the corresponding location will be designated as the risk center. and according to The attenuation range relative to the maximum value forms the risk impact region r.

[0087] Subsequently, the system calculates the abnormal road conditions based on multiple recent vehicle events. And obtain the distance of the target heavy vehicle from the risk area. Current speed acceleration Based on the traffic conditions ahead, calculate the estimated arrival time of the vehicle. Then determine the vehicle's equivalent mass based on the vehicle type. The equivalent dynamic action of the vehicle is calculated by combining velocity, acceleration, and rate of change of acceleration. The road conditions corresponding to the expected arrival time of the target vehicle. Vehicle equivalent dynamic action And by substituting the urgency of the early warning time into the individualized risk function, we obtain... .

[0088] When the obtained risk value reaches the corresponding level, the RSU sends a deceleration or lane change warning to the heavy vehicle. As the vehicle executes the warning and passes through the risk area at a reduced speed, the road-embedded sensor nodes again collect the road structure response after the vehicle has executed the warning and passed through the risk area. If the road structure response after the warning is executed is significantly reduced compared to similar historical heavy vehicles, but the propagation anomaly at a fixed location still exists, the system confirms that the location belongs to an inherent road structural risk, and simultaneously confirms that reducing speed can reduce the vehicle-road coupling risk.

[0089] The platform therefore writes the corresponding early warning strategies into the strategy library. For small vehicles entering the same road area later, because their equivalent dynamic impact is lower than that of the aforementioned heavy vehicles, even if the road conditions are the same, they may only generate a lower level of individualized risk value, thus sending different levels of R2X warnings. This achieves differentiated traffic control for different vehicles under the same road risk conditions.

[0090] Corresponding to the above embodiments, this application also provides a roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning system.

[0091] See Figure 5The roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning system 20 provided in this embodiment includes: The acquisition module 201 acquires the road structure response of sensor nodes at different burial depths within the road unit when a vehicle passes through the target road unit, and associates the responses of the same vehicle to form road unit status information.

[0092] The risk zone identification module 202 is used to obtain the operating status of the vehicle, correlate the operating status of the vehicle with the status information of the road unit in time to form a vehicle-road interaction event, determine the risk event based on the abnormal response, and inversely obtain the risk center and risk impact area.

[0093] The R2X warning module 203 is used to predict the road risk status when a vehicle approaches a risk impact area and determine the individualized risk, and generate and issue R2X warning information and strategies.

[0094] The verification and optimization module 204 is used to collect the road structure response after the target vehicle executes the R2X warning measures and passes through the risk zone, correct the risk center, risk impact area and R2X warning strategy, and use it for subsequent risk identification and warning work.

[0095] It should be noted that the specific content involved in the embodiments of this application can be found in the description of the above method embodiments, and will not be repeated here for the sake of brevity.

[0096] Corresponding to the above embodiments, this application also provides an electronic device.

[0097] See Figure 6 The electronic device 30 provided in this application embodiment includes a processor 301, a memory 302, and a communication unit 303. These components communicate via one or more buses, as those skilled in the art will understand. Figure 6 The electronic device structure shown does not constitute a limitation on the embodiments of this application. It can be a bus topology or a star topology, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0098] The communication unit 303 is used to establish a communication channel, thereby enabling the electronic device to communicate with other devices.

[0099] The processor 301 serves as the control center of the electronic device, connecting various parts of the device via various interfaces and lines. It executes software programs and / or modules stored in the memory 302, and calls data stored in the memory to perform various functions and / or process data. The processor can be composed of integrated circuits (ICs), such as a single packaged IC or multiple packaged ICs with the same or different functions connected together. For example, the processor 301 may consist only of a central processing unit (CPU). In this embodiment, the CPU may have a single processing core or include multiple processing cores.

[0100] Memory 302 is used to store the execution instructions of processor 301. Memory 302 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0101] When the execution instructions in memory 302 are executed by processor 301, the electronic device 30 is able to perform some or all of the steps in the above method embodiments.

[0102] Corresponding to the above embodiments, this application also provides a computer-readable storage medium, wherein the computer-readable storage medium may store a program, wherein when the program runs, it can control the device where the computer-readable storage medium is located to execute some or all of the steps in the above method embodiments. In specific implementation, the computer-readable storage medium may be a magnetic disk, an optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0103] Corresponding to the above embodiments, this application also provides a computer program product containing executable instructions that, when executed on a computer, cause the computer to perform some or all of the steps in the above method embodiments.

[0104] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, the simultaneous existence of A and B, or the existence of B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects have an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, and c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0105] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of electronic hardware and software. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0106] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0107] In the several embodiments provided in this application, any function, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0108] The above description is merely a specific embodiment of this application. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application. The protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A method for early warning of roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X, characterized in that, include: When a vehicle passes through a target road unit, it acquires the road structure response of sensor nodes at different burial depths within the road unit and associates the responses of the same vehicle to form road unit status information. The vehicle's operating status is obtained, and the vehicle's operating status is correlated with the road unit's status information over time to form a vehicle-road interaction event. Based on the abnormal response, risk events are determined, and the risk center and risk impact area are obtained through inversion. For vehicles approaching the risk impact zone, predict the road risk status upon their arrival and determine individual risks, then generate and issue R2X early warning information and strategies. The system collects road structure responses after target vehicles implement R2X warning measures and pass through risk zones, corrects risk centers, risk impact areas, and R2X warning strategies, and uses this data for subsequent risk identification and warning work.

2. The roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning method according to claim 1, characterized in that, When a vehicle passes through a target road unit, the road structure response of sensor nodes at different burial depths within the road unit is acquired, and the responses of the same vehicle are correlated to form road unit state information, including: The target road is divided into multiple continuous sensing road units along the direction of travel. Embedded sensing nodes are deployed at different depths in the surface layer, base layer and subgrade of each sensing road unit, and a three-dimensional road space coordinate mapping relationship is established for each embedded sensing node. When a vehicle enters the current sensing road unit, the continuous road structure responses belonging to the same vehicle are temporally correlated according to the order in which multiple sensing nodes arranged along the vehicle's direction of travel are excited by the wheel load, thus forming the road body response fingerprint of the corresponding vehicle. By associating the road response fingerprints formed by the continuous passage of different vehicles within the same perceived road unit, a normal propagation benchmark is established, and the changes in real-time response relative to the normal propagation benchmark are transformed into road unit state information.

3. The roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning method according to claim 1, characterized in that, The process of acquiring the vehicle's operating status, temporally correlating the vehicle's operating status with the road unit status information to form a vehicle-road interaction event, determining risk events based on abnormal responses, and inverting to obtain the risk center and risk impact area includes: Acquire multiple operational states of the target vehicle throughout the entire process of entering and exiting the current road unit, align them with the structural response time collected by the embedded sensor nodes, and construct a complete chain of continuous vehicle-road interaction events. For the same road location, the inherent road risks, abnormal vehicle loads, and coupled risks are distinguished based on different abnormal responses; To address the inherent risks of the road, spatial information is retrieved based on the response propagation distortion inversion of sensor nodes at different burial depths to determine the risk center and impact zone within the road.

4. The roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning method according to claim 3, characterized in that, To address the inherent risks of the road, based on the spatial information retrieved from the propagation distortion of the response of sensor nodes at different burial depths, the risk center and affected area within the road are determined, including: For any adjacent sensing nodes and Calculate the first Cross-node propagation distortion under vehicle-excited conditions ; Robust statistics are performed on the cross-node propagation distortions corresponding to the most recent N consecutive vehicle excitations to obtain robust propagation distortions for node pairs. ; Using any candidate spatial point inside the road As the risk source to be inverted, a risk source space inversion function is constructed: in: Indicates that it is located at a candidate point The set of sensor node pairs surrounding and participating in the current inversion; This represents the path consistency term between node pairs and candidate point Q. This represents the spatial propagation direction constraint between candidate risk sources and their corresponding sensor node pairs; Traverse the candidate spatial locations within the current abnormal road unit and calculate... Candidate positions that will cause the spatial inversion function to reach its maximum value. This has been identified as a current internal risk center of the road. The risk impact zone is determined based on the spatial location around the risk center that meets the preset inversion function decay conditions.

5. The R2X early warning method for roadbed and pavement embedded sensor fusion vehicle-road cooperative system according to claim 1, characterized in that, For vehicles approaching the risk impact zone, the system predicts the road risk status upon their arrival and determines individualized risks, generating and issuing R2X early warning information and strategies, including: Based on the internal risk center and risk impact area of ​​the road, the road body response generated when subsequent vehicles pass through is continuously acquired, a road abnormal state function is established, and a road risk state that evolves over time is formed. Predict the arrival time of the target vehicle in the risk impact area, and calculate the individualized real-time traffic risk value by combining the abnormal road conditions, the vehicle's equivalent dynamic action, and the warning margin. The risk value is compared with the classification threshold to generate R2X early warning control messages, and early warning information is issued in layers according to the remaining arrival time of the vehicle.

6. The roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning method according to claim 5, characterized in that, The predicted arrival time of the target vehicle in the risk impact area is used to calculate an individualized real-time traffic risk value, taking into account abnormal road conditions, the vehicle's equivalent dynamic action, and the warning margin. This includes: For a target vehicle about to enter the risk impact zone, calculate the predicted arrival time of the target vehicle in the risk impact zone and the equivalent dynamic effect on the road risk area; Based on the road anomaly status, road risk development speed, equivalent dynamic action, and warning time margin at the expected arrival time of the target vehicle, the spatiotemporal risk input status of the vehicle road is formed. Subsequently, a vehicle-road coupled risk function is constructed using the degree of road anomaly when the target vehicle is expected to arrive at the risk area, the strength of the dynamic load exerted by the target vehicle on the current road risk area, and the time urgency of the target vehicle's approach to the risk area: in: Indicates vehicle For risk areas Individualized real-time traffic risk values, This refers to the road conditions when the target vehicle is expected to reach the risk area. For road change enhancements, This represents the vehicle's equivalent dynamic action. Let r be the allowable equivalent dynamic action of the road region. To accelerate road development, The urgency of the vehicle's arrival.

7. The roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning method according to claim 6, characterized in that, The comparison of risk values ​​with grading thresholds generates R2X early warning control messages, and early warning information is issued in tiers based on the remaining arrival time of the vehicles, including: Based on vehicle-specific real-time risk values The relationship between the warning level and multiple risk level thresholds determines the warning level; Information is distributed in tiers based on the estimated remaining arrival time of vehicles from the risk impact zone. Trend warnings are sent to vehicles that are far away, behavioral suggestions are sent to vehicles that are approaching, and mandatory emergency alerts are sent to vehicles that are about to enter the risk impact zone.

8. The roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning method according to claim 1, characterized in that, The data collection process involves assessing the road structure response after the target vehicle executes R2X warning measures and passes through the risk zone, correcting the risk center, risk impact area, and R2X warning strategy, and using this data for subsequent risk identification and warning work. After the target vehicle implements R2X warning measures, the road structure response generated by its passage after the warning is implemented is collected and compared with the historical structure response of similar vehicles without control. The effectiveness of the early warning strategy is evaluated based on the improvement of the response, and the reliability of the positioning is evaluated based on the stability of the risk center and risk impact area obtained from the inversion of continuous vehicle events. Continuously record the road structure response after early warning is executed under multiple working conditions, construct multi-element mapping relationships, and support the selection of early warning strategies for similar risk areas.

9. The roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning method according to claim 8, characterized in that, The evaluation of the effectiveness of the early warning strategy based on the response improvement, and the assessment of the reliability of the location based on the stability of the risk center and risk impact area obtained from the inversion of continuous vehicle events, include: If the propagation distortion or abnormal road response in the road risk area is significantly reduced after the target vehicle implements R2X warning measures, the effectiveness evaluation of the current warning strategy will be improved. If the location of the risk center and the risk impact area obtained from the inversion of multiple consecutive vehicle events remain stable, the reliability of the location of the risk center and the risk impact area will be improved. If the road anomaly response does not change according to the predicted trend after the target vehicle implements the warning measures, the risk center and risk impact area will be re-determined based on the road structure response generated by the next vehicle passing by.

10. A roadbed and pavement embedded sensor fusion vehicle-road cooperative R2X early warning system, characterized in that, include: The acquisition module acquires the road structure response of sensor nodes at different burial depths within the road unit when a vehicle passes through the target road unit, and associates the responses of the same vehicle to form road unit status information. The risk zone identification module is used to obtain the operating status of the vehicle, correlate the operating status of the vehicle with the status information of the road unit in time to form a vehicle-road interaction event, determine the risk event based on the abnormal response, and inversely obtain the risk center and risk impact area. The R2X warning module is used to predict the road risk status when vehicles approach the risk impact area arrive and determine individual risks, and generate and issue R2X warning information and strategies. The verification and optimization module is used to collect the road structure response after the target vehicle executes the R2X warning measures and passes through the risk zone, correct the risk center, risk impact area and R2X warning strategy, and use it for subsequent risk identification and warning work.

Images