A road slope collapse monitoring analysis method for disaster risk identification

Abstract

This invention relates to the field of road slope collapse risk assessment and management technology, specifically a road slope collapse monitoring and analysis method for disaster risk identification. The method includes: based on acquired road collapse events, deploying monitoring and dissemination points along the communication range upstream and downstream of the road segment, and defining the monitoring coverage of each monitoring and dissemination point; using the data delay of road collapse event transmission as a guide, obtaining associated vehicles and vehicle approach indicators for the road collapse event; based on the vehicle approach indicators of associated vehicles, decomposing the driving probability distribution of associated vehicles, and combining this with the regional connectivity of the collapsed road to quantify the risk probability value of associated vehicles; based on the risk probability value of associated vehicles, calculating the intervention time corresponding to each planned route; responding to the intervention time of the planned routes, calculating the average update cycle of all intervention times, and marking the push order of risk warning information transmission.

Description

Technical Field

[0001] This invention relates to the field of road slope collapse risk assessment and management technology, specifically a road slope collapse monitoring and analysis method for disaster risk identification. Background Technology

[0002] During the rainy season and flood season, slope collapses are characterized by their suddenness, short warning window, and rapid spread of impact. They can easily cause traffic accidents on mountain roads, lead to prolonged traffic disruptions, and result in casualties. They are a major early warning scenario for road operation and management.

[0003] In existing technologies, risk warnings are sent by collecting precipitation and temperature information and recording scenarios of road hazards; or by tracking the movement of oncoming vehicles. These existing technologies are fixed-type risk management processes that neglect the coverage area of ​​landslide events and the filtering of related vehicle data. This leads to a mismatch between warning resources and risk scenarios during analysis, resulting in high-risk areas not receiving warnings while low-risk areas are over-covered, reducing the adaptability of risk management scenarios and the accuracy of scope definition. Summary of the Invention

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a road slope collapse monitoring and analysis method for disaster risk identification, including: S1, based on the acquired road collapse events, setting up monitoring and release points along the communication range upstream and downstream of the road segment, and clarifying the monitoring coverage of each monitoring and release point.

[0005] S2 uses the data delay of the road collapse event as a guide to obtain the associated vehicles and vehicle proximity indicators of the road collapse event.

[0006] S3, based on the vehicle proximity index of associated vehicles, decomposes the driving probability distribution of associated vehicles, and combines the regional connectivity of the collapsed road to quantify the risk probability value of associated vehicles.

[0007] S4. Based on the risk probability value of associated vehicles, determine at least one planned route that matches the currently associated vehicles when a road collapse event occurs, and calculate the intervention time corresponding to each planned route.

[0008] S5, in response to the intervention time of the planned route, calculates the average of the update cycles of all intervention times and marks the push order of risk warning information.

[0009] The beneficial effects of this invention are as follows: First, this invention calibrates the data delay corresponding to a landslide event by using the time difference between the first detection of the landslide and the system's event transmission time; this avoids the problem of excessive data screening caused by delayed control. Second, based on the landslide event boundary, upstream and downstream roadside units are retrieved, and the spatial distance between the points and the landslide point is recorded, and the connecting roads complete the initial screening of points. The communication range of the roadside units is divided based on the transmission delay and communication success rate to the vehicle-mounted unit. Finally, the monitoring coverage area is defined by the intersection of the communication range and the event boundary and the minimum bounding box. This ensures that the landslide event boundary and the actual communication scenario of the roadside units are compatible, avoiding coverage blind spots and further improving the accuracy of regional coverage.

[0010] Second, this invention fits the vehicle's road trajectory curve to a coordinate set and calculates the vehicle approach index based on the normalized values ​​of displacement rate and velocity change rate. Simultaneously, it uses the relationship between data latency and the vehicle's minimum braking time as a benchmark to accurately select associated vehicles for different scenarios. This ensures a safety baseline for early warning and further improves the processing accuracy for current scenario adaptation.

[0011] Third, this invention constructs a vehicle driving probability distribution using a Hidden Markov Model, builds a vehicle-road topology based on the merging results of multiple events, and after merging the entry probabilities, corrects the entry probabilities using the remaining available width of the landslide and the normalized ratio of the road length affected, ultimately obtaining the risk probability value. Simultaneously, combined with vehicle size constraints, it completes road drivability determination and regional connectivity labeling. Through the analysis and processing of vehicle driving intentions, it improves the mapping relationship between vehicle driving behavior and landslide events, and provides a decision-making basis for subsequent differentiated management decisions; it also enhances the accuracy of scenario-based risk avoidance settings.

[0012] Fourth, this invention matches planned routes to scenarios based on risk probability, clearly binding intervention time with points along the planned route. It also performs conflict verification on multiple vehicle control commands, outputting the verified intervention time. This avoids conflicting control commands, further improving the efficiency and security of risk control. Furthermore, it calculates the local update cycle and the global update cycle of the same planned route using a weighted average, adjusting the control push order for each planned route based on the ratio of the local to the global update cycles. Ultimately, this achieves the timing and priority control capability of prioritizing routes under the control plan. Attached Figure Description

[0013] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0014] Figure 1 This is a flowchart illustrating a road slope collapse monitoring and analysis method for disaster risk identification.

[0015] Figure 2This is a flowchart illustrating step S2 of a road slope collapse monitoring and analysis method for disaster risk identification.

[0016] Figure 3 This is a flowchart illustrating step S3 of a road slope collapse monitoring and analysis method for disaster risk identification.

[0017] Figure 4 This is a flowchart illustrating step S4 of a road slope collapse monitoring and analysis method for disaster risk identification. Detailed Implementation

[0018] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. Where specific techniques or conditions are not specified in the embodiments, they shall be performed in accordance with the techniques or conditions described in the literature in the art or in accordance with the product manual.

[0019] See Figure 1 A method for monitoring and analyzing road slope collapses for disaster risk identification includes: S1, based on the acquired road collapse events, setting up monitoring and publishing points along the communication range upstream and downstream of the road segment, and clarifying the monitoring coverage of each monitoring and publishing point.

[0020] S2 uses the data delay of the road collapse event as a guide to obtain the associated vehicles and vehicle proximity indicators of the road collapse event.

[0021] S3, based on the vehicle proximity index of associated vehicles, decomposes the driving probability distribution of associated vehicles, and combines the regional connectivity of the collapsed road to quantify the risk probability value of associated vehicles.

[0022] S4. Based on the risk probability value of associated vehicles, determine at least one planned route that matches the currently associated vehicles when a road collapse event occurs, and calculate the intervention time corresponding to each planned route.

[0023] S5, in response to the intervention time of the planned route, calculates the average of the update cycles of all intervention times and marks the push order of risk warning information.

[0024] In one embodiment of the present invention, the road locations of roadbeds, slopes, and bridges are selected as the analysis locations, and debris flows, bridge collapses, bedrock landslides, etc. are regarded as the processing content of road collapse. The normal driving, hazard avoidance, or obstacle avoidance driving scenarios of vehicles are taken as the analysis subjects of disaster response. The roads on which vehicles travel are abstracted into road networks, and the upstream and downstream roads in the road network relative to the collapse road time are taken as the identification data set. The time delay of each information release and the scene blocking effect are determined, thereby determining the detection and early warning effect of the collapse scene.

[0025] Furthermore, when acquiring road collapse events, the implementation method includes: acquiring optical remote sensing images of the road, identifying the settlement and deformation areas of the road according to optical remote sensing images of different time periods, and generating road collapse events.

[0026] The data delay corresponding to the road collapse event is set based on the time difference between the first time the collapse is detected and the time when the system sends the road collapse event. This data delay includes the data time of the road collapse detection delay + processing delay + sending delay, and is used to associate relevant vehicles in the corresponding time period to realize the assessment of road collapse risk for vehicle avoidance.

[0027] The optical remote sensing images capture roads that have subsided, accumulated, and deformed through cameras, drone patrols, and radar data. The data delay and relative location of the collapse are determined in real time according to the time of acquisition to complete the initial generation of the road collapse event.

[0028] In the actual data collection process, a monitoring sensor network is deployed on high-risk slope sections to record the time of road collapse in real time. The road collapse event communicates with upstream and downstream vehicles through roadside units, and monitoring and release points are deployed along the upstream and downstream of the road section according to the communication range, clarifying the monitoring coverage of each monitoring and release point. The monitoring coverage is then linked to the vehicle behavior of related vehicles to clarify the intervention time under different risks, thereby achieving the control and blocking of specific related vehicles.

[0029] Furthermore, the monitoring coverage is achieved by: S11, based on the location of the road collapse incident, data retrieval is performed on the upstream and downstream of the road segment, including determining the event boundary between the roadside unit and the road collapse incident.

[0030] S12, using roadside units as monitoring and reporting points, record the spatial distance and connecting roads between each monitoring and reporting point and the location of the landslide incident, and set the monitoring coverage range of each monitoring and reporting point in combination with the communication range corresponding to the monitoring and reporting point.

[0031] Specifically, the roadside unit represents a statistical unit that tracks vehicle flow and speed, and is used to share data with the vehicle-mounted unit. Its monitoring coverage range indicates the communicable range of the roadside unit in the upstream and downstream roads of the landslide incident, which is usually 300-1000 meters. It is easily affected by antennas and terrain and needs to be adjusted according to the actual scenario; thus, it quantifies the deployment of the monitoring coverage range in the current scenario.

[0032] Furthermore, when obtaining the connection between the monitoring and release points and the actual landslide, there is a direct connection between the monitoring and release points and the path affected by the landslide, and vehicles traveling in the direction of travel of that point will enter the landslide section.

[0033] Furthermore, the event boundary for a road collapse event refers to setting a spatial distance for filtering based on the risk level of the road collapse event. If a road collapse has already occurred or there is an extremely high risk of a road collapse, a red risk response is displayed, targeting upstream distances greater than or equal to 5km and at least 10 times the minimum braking distance of vehicles, and downstream distances greater than or equal to 1km and at least 2 times the minimum braking distance of vehicles. When an orange risk response is displayed, upstream distances greater than or equal to 3km and at least 6 times the minimum braking distance of vehicles are selected, and downstream distances greater than or equal to 500m and at least 1 time the minimum braking distance of vehicles are selected. Here, an event boundary is set for each road collapse event based on the risk level of the road collapse event, and this event boundary is the relative distance of the upstream and downstream selections mentioned above. Then, the length of the event boundary is used to complete the filtering and retrieval of the monitoring coverage area.

[0034] It should be noted that the red and orange alerts are used here only as a schematic illustration of the event boundaries. The specific boundaries need to be set according to the location of the collapsed road, and the boundaries should be drawn for each collapsed road.

[0035] Furthermore, the monitoring coverage can be expanded by filtering the monitoring points based on the spatial distance between each monitoring point and the location of the landslide incident, as well as the connecting roads, to obtain the monitoring points located within the event boundary.

[0036] Import the communication range of each monitoring and reporting point. The communication range is divided and described according to the transmission delay and communication success rate to reach the vehicle communication unit.

[0037] Based on the feature description of the communication range, the intersection of the communication range and the event boundary under different feature descriptions is defined, and when the communication range includes the event boundary, the corresponding communication range is regarded as the output monitoring coverage range.

[0038] If the communication range of the current monitoring and reporting point is smaller than the event boundary, the minimum bounding box containing the location of the road collapse event and the intersection of the ranges is taken as the starting point and considered as the output monitoring coverage range.

[0039] The communication range is divided into different descriptions based on the communication situation in the corresponding area, such as a communication success rate of 90%-99.9%, a transmission delay of 100-500ms, and a large fluctuation in communication quality; or a communication success rate of ≥99.9%, a transmission delay of ≤100ms, and stable communication quality; and the range intersection is divided according to the communication range of the current roadside unit.

[0040] If the current roadside unit can cover the event boundary, it is only necessary to notify vehicles within the communication range of the current roadside unit to prevent the risk of road collapse; if the event boundary is significantly larger than the communication range, the locations of multiple intersections need to be described by the minimum bounding box to form a monitoring coverage area that includes the intersection of the collapsed road and the range.

[0041] In one embodiment of the present invention, based on the data delay of the road collapse event, data is retrieved from the upstream and downstream roads of the collapsed road to determine at least one group of associated vehicles, and the current risk warning information is sent in a specific manner based on the traffic conditions of the roads where the associated vehicles are located.

[0042] Specifically, such as Figure 2 As shown, one implementation of step S2 includes: S21, selecting vehicles from the monitoring coverage area according to the time length corresponding to the data delay, and obtaining at least one group of associated vehicles; specifically, the screened associated vehicles are at least vehicles that have sufficient time to complete safe braking / avoidance diversion after receiving the warning; then, based on the data delay, determining the allowable data delay size of the associated vehicles, and using the data delay size to distinguish data for each group of associated vehicles.

[0043] Furthermore, considering the data delay in road collapse events, it is necessary to consider the specific vehicles that may be affected by this data delay, that is, the situation where vehicles are near the collapse location due to the delay in the identification of the collapse event.

[0044] Specifically, the method for selecting vehicles from the monitoring coverage area in step S21 includes: when the data delay is less than the minimum braking time of the vehicle, the number of vehicles corresponding to the data delay is counted, and the corresponding vehicles are regarded as the associated vehicles in the output.

[0045] When the data delay is greater than the vehicle's minimum braking time, the minimum braking time of the vehicle is used as the starting point for the query, and the corresponding vehicle is regarded as the associated vehicle in the output.

[0046] Specifically, the query range starts with the travel distance corresponding to the data delay and the minimum braking time of the vehicle, respectively, and searches upstream for corresponding vehicles. The retrieved vehicles are then considered as the associated vehicles in the output.

[0047] When the data latency is small, it indicates timely control, and vehicles still have enough time to brake. Data is sent in real time based on the current latency to identify the corresponding vehicles that need to be controlled. When the data latency is large, it indicates that the current control is prone to delays. It is necessary to associate the identified vehicles with the vehicles that need to be controlled based on their minimum braking time and send data to each vehicle. When the data latency is equal to the vehicle's minimum braking time, it does not affect the selection of currently associated vehicles. This ensures that the selection of associated vehicles always starts from the minimum value of the two to retrieve data and form the output list of associated vehicles.

[0048] S22, for the coordinate set of associated vehicles, fit the road direction curve of the associated vehicles.

[0049] S23. For each group of associated vehicles, determine the displacement rate and velocity change rate when fitting the current road direction curve, and calculate the vehicle approach index based on the normalized values ​​of the two. Among them, the displacement rate refers to the speed at which the vehicle moves on the fitted road direction curve per unit time. The higher the value, the faster the vehicle moves towards the risk point and the closer it is to the risk point. The velocity change rate represents the acceleration of the vehicle and is used to distinguish between vehicles that are accelerating and those that are decelerating.

[0050] Secondly, the vehicle approach index calculated by both is represented as a weighted sum. After normalizing the displacement rate and velocity change rate, weights are assigned to the two in sequence, with values ​​of 0.6-0.7 and 0.3-0.4 respectively. Since the relative distance between the vehicle and the risk point is the first consideration, the weight of the displacement rate is strengthened to quantify the specific situation of the vehicle approaching the collapsed road.

[0051] Based on vehicle approach indicators, the dynamic correlation of vehicle movement is measured in real time, and the indicator information that is positively correlated with the current road collapse event is identified, and the vehicle approach indicators are synchronized to the associated vehicles.

[0052] Furthermore, a higher vehicle proximity index value indicates a stronger positive correlation between the vehicle and the landslide event, and a greater impact of the landslide risk on the vehicle. In other words, for the vehicle proximity index, multiple levels of related vehicles are selected for attention, and the related vehicles at each level are updated in real time to determine the current range of related vehicles.

[0053] Specifically, the vehicle approach index classifies associated vehicles according to its value, including dividing vehicles into Level 1 emergency vehicles, Level 2 attention vehicles, and Level 3 alert vehicles using 0.3 and 0.7 as the dividing intervals; for each type of associated vehicle, the index is updated and calculated in real time according to the unit of time to determine the currently output associated vehicle.

[0054] In one embodiment of the present invention, step S3 is used to associate the real-time behavior of vehicles with the current landslide event. The road distance and time margin of the associated vehicle from the landslide point are used as the behavioral dimensions for preliminary analysis. Through regional connectivity analysis, the rationality of the current vehicle warning and blocking is determined. On this basis, combined with the specific driving intentions of vehicles in different road sections, such as going straight / turning / diverting / leaving, these are used as the basis for the selection decision of each vehicle in the overall road network, and finally a driving probability distribution of multi-path selection is formed. This realizes the risk management and handling of landslide roads in various driving scenarios, and then constructs the correlation relationship of vehicle driving.

[0055] Specifically, such as Figure 3 As shown, one implementation of step S3 includes: S31, taking the current associated vehicle's vehicle behavior and vehicle approach index as the observation sequence, and taking the associated vehicle's driving intention as the hidden state, to construct the driving probability distribution of the associated vehicle on multiple roads; wherein, in addition to the vehicle approach index, the observation sequence also includes the road distance from the landslide point and the remaining braking time margin. The remaining braking time margin represents the time amount after subtracting the safety distance from the landslide point from the road distance from the landslide point and dividing by the current speed. At the same time, this time amount needs to be subtracted from the corresponding data delay, thereby constructing multiple sets of discrete observation values.

[0056] The driving probability distribution represents the probability value calculated by the hidden Markov model. Specifically, for each vehicle within the monitoring coverage area, a sliding window is maintained to store the observation values ​​from the most recent multiple moments.

[0057] S32, based on the order of occurrence of different road collapse events, spatially merge multiple road collapse events to clarify the vehicle-road topology relationship between the associated vehicles and the collapsed roads.

[0058] Here, data will be merged according to the type of road collapse. Road collapse events include multiple events such as primary collapse, secondary collapse, road surface subsidence, and traffic control. These events will be spatially merged in the order of occurrence, and the distribution of the road currently being traveled and the road corresponding to the collapsed road in the road network will be recorded in real time to form a vehicle-road topology relationship.

[0059] S33 utilizes the driving probabilities on multiple roads to perform connectivity analysis on the current collapsed road and calculates the risk probability value of related vehicles driving onto the collapsed road.

[0060] The connectivity analysis includes analyzing the traffic conditions of the collapsed road based on the remaining available width and affected road length, and then implementing risk management measures for each related vehicle based on whether it is passable.

[0061] Specifically, the connectivity analysis in step S33 is implemented by: merging all driving probabilities into the collapsed road and outputting the driving probability corresponding to the collapsed road; selecting the remaining available width and the normalized ratio affecting the road length to correct the driving probability, thus obtaining the output risk probability value. The correction methods include directly using the product of the normalized value and the driving probability, and using the conditional probability of the normalized value and the driving probability, treating both values ​​as risk probability values ​​to be considered in the current scenario.

[0062] Secondly, the remaining available width and the affected road length will serve as the basis for verifying whether the road is passable, and the specific form of the landslide will be synchronized to different associated vehicle types.

[0063] Specifically, the size constraints of the current vehicle are obtained. The size constraints represent the safety margin that the vehicle width and length need to be reserved in the current scenario; that is, the safe distance between the vehicle and the landslide body / roadside. If the current remaining available width does not meet the requirements, it means that the road is not passable. If the size constraints do not meet the remaining available width, the current road is considered to be impassable. The subsequent blocking strategy is set to a control prompt to avoid the current road segment.

[0064] If the dimensional constraints satisfy the remaining available width, a risk probability value is introduced to mark the regional connectivity of the collapsed road. Specifically, for the entered risk probability value, a threshold judgment is performed on the current collapsed road. The minimum and maximum values ​​of the risk probability marked as passable in the current scenario are retrieved. If the current vehicle's risk probability is less than or equal to the minimum value, a normal risk warning is issued, explaining the specific parameters of the road collapse. If the risk probability is between the maximum and minimum values, the road is marked as a cautious passage state, and the corresponding vehicle is described as passable or impassable in the risk management communication. If the risk probability is greater than the maximum value, it is marked as impassable, and relevant notification information is added to the warning message.

[0065] In one embodiment of the present invention, given an observation sequence The forward algorithm is used to calculate the probability value of being in each hidden state at the current moment, and to predict the probability value at multiple subsequent moments, thereby determining the output driving probability.

[0066] The computational methods of the forward algorithm include: ;in, The observation sequence up to time t is And the current state is The joint probability; Let represent the forward vector, which is the value corresponding to the j-th hidden state; This represents the hidden state of the vehicle at time t; This represents the j-th hidden state, where j represents the index number of the hidden state. This represents the observation sequence from time 1 to t. The forward algorithm predicts the probability value at subsequent time points through recursion and uses the distribution of the probability value as the current driving probability. At the same time, the output driving probability must conform to the Gaussian distribution form to determine the current cyclical progress process.

[0067] Furthermore, the recursive representation of the forward vector is as follows: .

[0068] .

[0069] in, This represents the initial forward variable, which is the value corresponding to the j-th hidden state; represents the initial probability of a vehicle entering the monitoring coverage area, and represents the initial probability value of it being in the hidden state j; This represents the discrete probability value of hidden state j at the initial time. Let i represent the forward variable of hidden state i at time t-1, where i represents the index of the hidden state. This represents the state transition probability, indicating the probability value of transitioning from state i to state j; Let represent the discrete probability value of hidden state j at time t, and let represent the probability value that each hidden state conforms to the Gaussian mixture model.

[0070] Secondly, based on the recursive forward vector, the probability of driving at multiple subsequent time points is predicted, which is implemented as follows: .

[0071] Where K represents the prediction step size, which represents the prediction time length; This represents the probability of transitioning from state i to state j in K steps; Indicates the state after predicting K steps. The probability of; Represents the posterior probability, used to describe the current state. Probability value. After obtaining multiple probability values, the predicted probability value here is the current output probability of driving.

[0072] In one embodiment of the present invention, after determining the risk probability of the associated vehicles, step S4 is used to push control prompts to the associated vehicles involved according to the value of the risk probability, and adjust the time interval of pushing control prompts according to the value of the risk probability, so as to realize the processing method of blocking vehicles on different routes.

[0073] Specifically, for the relative descriptions of passable and impassable vehicles marked by risk probability in step S3, multiple planned routes including detour routes and alternative routes are generated for each vehicle according to the specific values ​​of vehicle approach index and risk probability; at the same time, the lanes and diversion points corresponding to the routes are determined, thereby obtaining the data format for sending control instructions to each vehicle.

[0074] Secondly, based on the relative time points of the first intervention trigger, continuous push, and intervention termination push in the current scenario, the update cycle of each time point is taken as the main body of data regulation and synchronized to the intervention time of each planned route.

[0075] like Figure 4As shown, one implementation of step S4 includes: S41, classifying associated vehicles into scenarios according to the risk probability of associated vehicles, and determining the planned routes configured in each scenario; the planned routes divided here will directly introduce passable and impassable signs, and record the detour routes and alternative routes that the associated vehicles can currently travel on the road in sequence.

[0076] Secondly, the selected detour route can be calculated based on the above-mentioned forward algorithm, transforming the observed values ​​into a relative scenario of the current risk probability division, and statistically analyzing the probability of entering other roads. That is, after excluding the probability of entering the currently collapsed road, the route with the highest probability of entering is regarded as the detour route, and the two routes with the highest probability of entering among the remaining routes are selected as alternative routes. The recorded planned route is adjusted in real time according to the changes in the entered data.

[0077] S42, based on the location of the planned route on upstream and downstream roads, view the intervention time, which includes the first intervention trigger time, continuous push period, and intervention end push time.

[0078] Among them, the first intervention trigger time represents the time when the control command is first sent to the user after the vehicle enters the monitoring coverage area, according to its risk probability; the intervention end push time represents the time from the first intervention when the user completes obstacle avoidance or safe braking and leaves the landslide-affected section of the road, or when the risk of the landslide section is eliminated and the road section is reopened to traffic; the continuous push cycle is the relative cycle of the control command push according to the risk probability corresponding to the scenario, such as 500ms / time.

[0079] Specifically, the initial intervention trigger time is based on either the remaining time expected for the vehicle to reach the diversion node connecting to the collapsed road on the planned route, or the remaining time expected to reach the average risk probability in historical data. This is subtracted from the time required for vehicle evacuation, data latency, and intervention redundancy time. The corresponding time point is then recorded. The time required for vehicle evacuation is retrieved from the database based on the vehicle's current speed and road segment. The intervention redundancy time represents the additional time set according to the scenario corresponding to the risk probability.

[0080] Secondly, based on the risk probability of vehicle operation, the continuous push cycle of the vehicle is viewed in real time, and the push cycle that needs to be controlled is continuously recorded in a multi-control manner; at the same time, the intervention deadline push time is used to estimate the time window for intervention completion, and the update cycle for changes in intervention time is determined in real time according to the changes in the time point corresponding to the intervention time, thereby synchronously determining the data push of each associated vehicle.

[0081] S43 links the intervention time with the monitoring and reporting points along the planned route, clarifying the control instructions for each round of calculation.

[0082] Specifically, the intervention time and monitoring release points are linked to the route conditions and combined with the description of the collapsed road and related data at that time to form the control instructions to be sent.

[0083] S44 determines whether there are conflicts in the control instructions of multiple related vehicles, and regards the data after the conflict verification as the output intervention time.

[0084] The conflict check is used to verify the resource allocation of each vehicle's planned route, including but not limited to: the same vehicle receiving mutually exclusive instructions from different locations, multiple vehicles in adjacent lanes simultaneously receiving instructions to change lanes to the same lane, vehicles traveling in front and behind simultaneously receiving mutually exclusive instructions, and instructions that would cause vehicle collisions. For these control instructions, the order of instruction sending needs to be adjusted according to their risk probability to reduce the risk to vehicle driving when control instructions are sent.

[0085] In one embodiment of the present invention, by observing the relative period of data updates under the push of control commands in real time, the time for sending the overall control commands is averaged to avoid command conflicts caused by asynchronous decision updates of different vehicles.

[0086] Specifically, one implementation of step S5 includes: calculating the local update cycle of the same planned route and the global update cycle of all planned routes by weighted average based on the update cycle length value of each associated vehicle under the current planned route.

[0087] In the weighted calculation, the length of time from which the intervention time is updated is used as the feature value, the distance and time between the associated vehicle and the collapsed road, and the normalized value of the risk probability are used as weight factors. Local values ​​and global values ​​for all routes are calculated for associated vehicles on the same planned route.

[0088] Specifically, the weights related to distance and time from the collapsed road refer to using a minimum-maximum normalization method within the same planned route to convert the distance and time from the collapsed road into a fixed range of values; or setting the weights related to distance and time in the form of (maximum value - current value + minimum value) / maximum value.

[0089] Based on the ratio of local update cycles to global update cycles, the push order of each planned route is adjusted according to its value, and the part with the smallest ratio value is regarded as the priority push part. The process of pushing related vehicle control instructions on each planned route is determined by sorting in descending order.

[0090] It should be noted that the current update cycle represents the period during which the estimated intervention time for each associated vehicle changes after it enters the monitoring coverage area, and represents the cycle value of the background decision update. If the length of the update cycle on a certain planned route is too small, it means that the probability of vehicles on that route traveling towards the landslide road is higher, the decision refresh frequency is faster, and it is more necessary to adjust the push of vehicle control instructions on that route to achieve real-time transmission of vehicle risk warning information and disaster control.

[0091] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention, which are still covered within the protection scope of the present invention.

Claims

1. A method for monitoring and analyzing road slope collapses for disaster risk identification, characterized in that, include: S1. Based on the acquired road collapse events, deploy monitoring and reporting points along the communication range upstream and downstream of the road segment, and clarify the monitoring coverage of each monitoring and reporting point. S2, guided by the data delay in sending road collapse events, obtain the associated vehicles and vehicle proximity indicators of road collapse events; S3, based on the vehicle proximity index of associated vehicles, decomposes the driving probability distribution of associated vehicles, and combines the regional connectivity of the collapsed road to quantify the risk probability value of associated vehicles. S4. Based on the risk probability value of associated vehicles, determine at least one planned route that matches the currently associated vehicles when the road collapse event occurs, and estimate the intervention time corresponding to each planned route. S5, in response to the intervention time of the planned route, calculates the average of the update cycles of all intervention times and marks the push order of risk warning information.

2. The method for monitoring and analyzing road slope collapse for disaster risk identification according to claim 1, characterized in that, The methods for obtaining road collapse events include: Collect optical remote sensing images of roads, identify areas of road subsidence and deformation based on optical remote sensing images from different time periods, and generate road collapse events; The data delay corresponding to the road collapse event is set based on the time difference between the time when the collapse is first detected and the time when the system sends the road collapse event.

3. The method for monitoring and analyzing road slope collapse for disaster risk identification according to claim 1, characterized in that, The methods for monitoring coverage in step S1 include: S11, based on the location of the road collapse incident, perform data retrieval on the upstream and downstream of the road segment, including determining the event boundary of the roadside unit and the road collapse incident; S12, using roadside units as monitoring and reporting points, record the spatial distance and connecting roads between each monitoring and reporting point and the location of the landslide incident, and set the monitoring coverage range of each monitoring and reporting point in combination with the communication range corresponding to the monitoring and reporting point.

4. The method for monitoring and analyzing road slope collapse for disaster risk identification according to claim 3, characterized in that, The methods for setting the monitoring coverage range of each monitoring and reporting point, based on the communication range corresponding to the monitoring and reporting point, include: Based on the spatial distance and connecting roads between each monitoring and reporting point and the location of the road collapse incident, the monitoring and reporting points are filtered to obtain the monitoring and reporting points located within the event boundary; Import the communication range of each monitoring and reporting point. The communication range is divided and described according to the transmission delay and communication success rate to reach the vehicle communication unit. Based on the feature description of the communication range, the intersection of the communication range and the event boundary under different feature descriptions is defined, and when the communication range includes the event boundary, the corresponding communication range is regarded as the output monitoring coverage range; If the communication range of the current monitoring and reporting point is smaller than the event boundary, the minimum bounding box containing the location of the road collapse event and the intersection of the ranges is taken as the starting point and considered as the output monitoring coverage range.

5. The method for monitoring and analyzing road slope collapse for disaster risk identification according to claim 1, characterized in that, The vehicle approach indicator in step S2 is implemented in the following ways: S21, select vehicles from the monitoring coverage area according to the time length corresponding to the data delay, and obtain at least one set of associated vehicles; S22, for the coordinate set of associated vehicles, fit the road direction curve of the associated vehicles; S23. For each group of associated vehicles, determine the displacement rate and velocity change rate when fitting the current road direction curve, and calculate the vehicle approach index based on the normalized values ​​of the two.

6. The method for monitoring and analyzing road slope collapse for disaster risk identification according to claim 5, characterized in that, The methods for selecting vehicles from the monitoring coverage area in step S21 include: When the data delay is less than the minimum braking time of the vehicle, the number of vehicles corresponding to the data delay is counted, and the corresponding vehicles are regarded as the associated vehicles in the output. When the data delay is greater than the vehicle's minimum braking time, the minimum braking time of the vehicle is used as the starting point for the query, and the corresponding vehicle is regarded as the associated vehicle in the output.

7. The method for monitoring and analyzing road slope collapse for disaster risk identification according to claim 1, characterized in that, The risk probability value in step S3 is implemented in the following ways: S31, using the vehicle behavior and vehicle approach indicators of the currently associated vehicles as the observation sequence and the driving intention of the associated vehicles as the hidden state, construct the driving probability distribution of the associated vehicles on multiple roads. S32, based on the order of occurrence of different road collapse events, spatially merge multiple road collapse events to clarify the vehicle-road topology relationship between the associated vehicles and the collapsed roads; S33 utilizes the driving probabilities on multiple roads to perform connectivity analysis on the current collapsed road and calculates the risk probability value of related vehicles driving onto the collapsed road.

8. The method for monitoring and analyzing road slope collapse for disaster risk identification according to claim 7, characterized in that, The connectivity analysis in step S33 is implemented in the following ways: Combine the driving probabilities of all roads that have collapsed, and output the driving probability corresponding to the collapsed road; select the remaining available width and the normalized ratio of the road length to correct the driving probability, and obtain the output risk probability value. Obtain the current vehicle's size constraints; if the size constraints do not meet the remaining available width, the current road is considered undriveable. If the size constraint satisfies the remaining available width, a risk probability value is introduced to mark the regional connectivity of the collapsed road.

9. A method for monitoring and analyzing road slope collapse for disaster risk identification according to claim 1, characterized in that, The intervention time in step S4 is implemented in the following ways: S41, Based on the risk probability of associated vehicles, classify associated vehicles into scenarios and determine the planned routes configured for each scenario; S42, based on the location of the planned route on upstream and downstream roads, view the intervention time including the first trigger time of the warning, the continuous push period, and the warning end time; S43 links the intervention time with the monitoring and release points along the planned route, and clarifies the early warning instructions for each round of calculation; S44 determines whether there is a conflict between the warning commands of multiple related vehicles, and regards the data after the conflict verification as the output intervention time.

10. A method for monitoring and analyzing road slope collapse for disaster risk identification according to claim 1, characterized in that, The implementation methods for the push order in step S5 include: Based on the update cycle length values ​​corresponding to each associated vehicle under the current planned route, the local update cycle belonging to the same planned route and the global update cycle corresponding to all planned routes are calculated by weighted average. The order in which each planned route is pushed is adjusted based on the ratio of the local update cycle to the global update cycle.

Images