Vehicle driving early warning method and system for sudden collapse on highway under severe conditions

Abstract

The invention discloses an early warning method and system for driving of vehicles in sudden collapse of an expressway under severe conditions, and relates to the field of road traffic data analysis, the early warning method for driving of vehicles in sudden collapse of the expressway under severe conditions mainly comprises the following steps: obtaining vehicle information, historical expressway data and meteorological environment data of a target expressway section; obtaining a risk level by using the historical high-speed data, the meteorological environment data, the real-time high-speed data and the prediction model; according to the surrounding vehicle information, performing abnormity judgment and event single occasional judgment to obtain reminding information; according to the reminding information, vehicle stopping information and vehicle stopping sensor permission are obtained; and obtaining a danger range by using an information fusion technology according to the stopping vehicle sensor permission. According to the early warning method and the early warning system for the vehicle running in the sudden collapse of the expressway under the severe condition, the range and the advance of road safety early warning can be improved.

Description

Technical Field

[0001] The present invention relates to the field of road traffic data analysis, and more specifically to a method and system for early warning of vehicles traveling on highways in the event of sudden landslides under adverse conditions. Background Art

[0002] In recent years, the expressway network has been continuously expanded. The characteristics of expressways such as high speed, large traffic volume and road closures have led to high accident rates and the loss of life and property after accidents. Among the many causes of road traffic accidents, landslides on expressways or collapse of expressway bridges, especially at night, can cause serious accidents.

[0003] At present, the traditional traffic incident detection system mainly provides simple traffic information and detects basic traffic data for urban roads, mountain roads, and expressways, thereby realizing the release of macro-static information. As for real-time safety warning, roadside information intelligence is usually used to carry out safety information warning in accident-prone areas, which has the problems of small safety warning range, poor effect, and inability to carry out early warning and prevention. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for early warning of vehicles traveling on highways under severe conditions, which can improve the scope and advance time of road safety early warning.

[0005] The present invention provides a vehicle driving warning method for sudden landslide on a highway under adverse conditions, comprising the following steps: S100: acquiring vehicle information, historical highway data and meteorological environment data of a target highway section; S200: constructing a prediction model, and obtaining a risk level by using the historical highway data, meteorological environment data, real-time highway data and the prediction model; S300: acquiring surrounding vehicle information, performing abnormality determination and single contingency determination of an event according to the surrounding vehicle information, and obtaining reminder information; S400: obtaining stopped vehicle information and stopped vehicle sensor authority according to the reminder information; and obtaining a danger range according to the stopped vehicle sensor authority by using information fusion technology.

[0006] Furthermore, step S200 of the vehicle driving warning method for sudden landslides on highways under the above-mentioned severe conditions specifically includes: S201: constructing a prediction model, and training the prediction model using historical highway data to obtain a trained prediction model; S202: acquiring real-time highway data, and obtaining the probability of landslides based on the real-time highway data and meteorological environment data using the trained prediction model; S203: conducting risk assessment and grading of the probability of landslides to obtain a risk level.

[0007] Furthermore, step S201 of the above-mentioned method for early warning of sudden landslides on highways under severe conditions specifically includes: constructing a prediction model, training the prediction model using historical highway data, and obtaining a trained prediction model; the prediction model is a logistic regression model, such as the formula: , in, is the probability of landslide, β 0 is the intercept of the model, β 1 , β 2 , …, β n is the regression coefficient of each feature.

[0008] Furthermore, step S300 of the vehicle driving warning method for sudden landslides on highways under severe conditions specifically includes: S301: using wireless communication technology to obtain surrounding vehicle information, and the surrounding vehicle information is used to determine whether there are abnormal conditions in the surrounding vehicles; S302: making an abnormality determination based on the surrounding vehicle information to obtain the status of the surrounding vehicles; based on the surrounding vehicle status, obtaining the vehicle information around the abnormal vehicle, and making a single and contingent event determination based on the vehicle information around the abnormal vehicle to obtain the single and contingent event status; S303: obtaining reminder information based on the single and contingent event status.

[0009] Furthermore, step S400 of the vehicle driving warning method for sudden landslides on highways under the above-mentioned severe conditions specifically includes: S401: obtaining stopped vehicle information based on the reminder information, and obtaining stopped vehicle sensor authority based on the stopped vehicle information; S402: obtaining stopped vehicle sensor data based on the stopped vehicle sensor authority, and fusing the stopped vehicle sensor data using wireless communication technology to obtain second data; obtaining the landslide range and landslide location based on the second data; S403: obtaining the danger range based on the landslide range and landslide location.

[0010] The present invention also provides a vehicle driving warning system for sudden landslides on highways under adverse conditions, comprising the following modules: a data acquisition module, configured to acquire vehicle information, historical highway data and meteorological environment data of a target highway section; a risk module, configured to construct a prediction model, and obtain risk levels using historical highway data, meteorological environment data, real-time highway data and the prediction model; an abnormality module, configured to acquire surrounding vehicle information, and perform abnormality determination and single and sporadic determination of events based on the surrounding vehicle information to obtain reminder information; a range module, configured to acquire stopped vehicle information and stopped vehicle sensor authority based on the reminder information; and obtain a dangerous range based on the stopped vehicle sensor authority using information fusion technology.

[0011] Furthermore, the risk module of the vehicle driving warning system for sudden landslides on highways under the above-mentioned severe conditions includes: a model unit, configured to: construct a prediction model, train the prediction model using historical highway data, and obtain a trained prediction model; a calculation unit, configured to: obtain real-time highway data, and obtain the probability of landslide occurrence based on the real-time highway data and meteorological environment data using the trained prediction model; an evaluation unit, configured to: perform risk evaluation and grade classification on the probability of landslide occurrence to obtain a risk level.

[0012] Furthermore, the model unit of the vehicle driving warning system for sudden landslides on highways under the above-mentioned severe conditions is specifically configured as follows: constructing a prediction model, training the prediction model using historical highway data, and obtaining a trained prediction model; the prediction model is a logistic regression model, such as the formula: , in, is the probability of landslide, β 0 is the intercept of the model, β 1 , β 2 , …, β n is the regression coefficient of each feature.

[0013] Furthermore, the abnormal module of the vehicle driving warning system for sudden landslides on highways under the above-mentioned severe conditions includes: an abnormal unit, configured to: use wireless communication technology to obtain surrounding vehicle information, and the surrounding vehicle information is used to determine whether there is an abnormal situation with the surrounding vehicles; a trajectory unit, configured to: make an abnormality determination based on the surrounding vehicle information to obtain the surrounding vehicle status; obtain the vehicle information around the abnormal vehicle based on the surrounding vehicle status, and make a single and contingent event determination based on the vehicle information around the abnormal vehicle to obtain the single and contingent event status; a determination unit, configured to: obtain reminder information based on the single and contingent event status.

[0014] Furthermore, the range module of the vehicle driving warning system for sudden landslides on highways under the above-mentioned severe conditions includes: an authority unit, configured to: obtain stopped vehicle information according to the reminder information, and obtain stopped vehicle sensor authority according to the stopped vehicle information; a fusion unit, configured to: obtain stopped vehicle sensor data according to the stopped vehicle sensor authority, and use wireless communication technology to fuse the stopped vehicle sensor data to obtain second data; obtain the landslide range and landslide location according to the second data; a processing unit, configured to: obtain the danger range according to the landslide range and landslide location.

[0015] The method and system for early warning of sudden landslides on highways under severe conditions provided by the present invention have the following beneficial effects: The present invention can reduce the amount of backend calculation by evaluating the road according to the probability of road collapse in different areas, build a prediction model according to historical high-speed data, input the current meteorological, geological and road structure data into the logistic regression model in real time, calculate the probability of landslide, obtain real-time high-speed data, input the real-time high-speed data into the logistic regression model, calculate the probability of landslide, perform risk assessment according to the probability result, divide the risk level, obtain the surrounding vehicle information by wireless communication technology, adopt vehicle-to-vehicle communication, determine whether an abnormal situation occurs, if it is determined to be an abnormal situation, obtain the vehicle information around the abnormal vehicle, obtain the vehicle driving track around the abnormal vehicle, determine whether it is a single accidental event, if it is determined to be a single accidental event, send a reminder to the rear vehicle terminal, if it is determined to be a non-single accidental event, combine the risk level to determine it as a landslide, obtain the vehicle information of the vehicle not affected by the landslide, fuse multiple groups of laser radar data and visual data by wireless communication technology, obtain second data, determine the abnormal range according to the second data, identify the range and position of the landslide, solve the technical problems of small safety warning range, poor effect and inability to perform early warning and prevention in the prior art, and improve the range and advance amount of road safety warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which: Figure 1 It is a flow chart of a method for early warning of sudden landslides on highways under adverse conditions provided by the present invention; Figure 2 A flowchart of the steps of building a prediction model based on historical high-speed data, acquiring real-time high-speed data, and dividing risk levels provided by the present invention; Figure 3 A flowchart of the steps of obtaining surrounding vehicle information through wireless communication technology and determining whether an abnormal situation occurs provided by the present invention; Figure 4 A flowchart of the steps of obtaining vehicle information of a stopped vehicle and determining an abnormal range based on laser radar data and visual data provided by the present invention; Figure 5 This is a structural diagram of a vehicle driving warning system for sudden landslides on highways under severe conditions provided by the present invention; Figure 6 The architecture diagram of the risk module provided by the present invention; Figure 7 The architecture diagram of the abnormal module provided by the present invention; Figure 8 An architectural diagram of the scope module provided by the present invention; Fig. 9 A schematic diagram of lane-level navigation provided by the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0019] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0020] Figure 1 A schematic diagram of the method for early warning of sudden landslides on highways under severe conditions according to the present embodiment is shown.

[0021] In this embodiment, the method for early warning of sudden landslide on a highway under severe conditions includes the following steps: S100: Acquire vehicle information, historical highway data, and meteorological environment data of the target highway section; Specifically, vehicle information includes vehicle location and vehicle speed, historical highway data includes geological environment data and highway structure data, and meteorological environment data includes historical meteorological data, meteorological real-time monitoring data, and meteorological forecast data; Specifically, the global positioning system (GPS) is used to obtain the accurate position of each vehicle on the highway in real time, that is, the vehicle position, including latitude and longitude, lane information, and relative distance; the vehicle speed sensor and GPS are used to obtain real-time speed information, that is, the vehicle speed, so as to obtain the vehicle information of the target highway section, which is used to help analyze the driving status of the vehicle and determine whether it is necessary to slow down or detour; Specifically, the relative distance is the distance between the target vehicle and the host vehicle; Specifically, geological environment data include soil types along the highway, including clay, sand, and rock; highway structure data include highway slope; historical meteorological data include historical temperature data and historical humidity data; meteorological real-time monitoring data include current rainfall; Specifically, geological environment data also includes topography, water system distribution, vegetation and land use type; highway structure data also includes road surface type and roadbed type; meteorological real-time monitoring data also includes real-time temperature data and real-time humidity data; meteorological forecast data includes future rainfall, future temperature data and future humidity data; Specifically, understand the soil types along the highway. Different soil types react differently to rainfall and geological activities, affecting the risk of landslides. The slope of the highway, especially steep slopes, is prone to landslides after heavy rain. Real-time monitoring of current rainfall, especially in heavy rain and continuous rainfall, and historical temperature and humidity data can help determine under what conditions soil saturation increases, leading to increased landslide risk. By acquiring and integrating highway vehicle information, historical highway data, geological environment data, and meteorological environment data, a comprehensive risk assessment can be conducted before a landslide occurs. S200: Build a prediction model, using historical high-speed data, meteorological environment data, real-time high-speed data and prediction models to obtain risk levels; Specifically, the parameters of the prediction model include topography, water system distribution, vegetation and land use type, highway structure and meteorological environment data; Specifically, Figure 2 The flowchart is shown as follows: Steps of building a prediction model based on historical high-speed data, acquiring real-time high-speed data, and dividing risk levels. Step S200 includes: S201: construct a prediction model, and train the prediction model using historical high-speed data to obtain a trained prediction model; Specifically, step S201 includes: constructing a prediction model, training the prediction model using historical high-speed data, and obtaining a trained prediction model; the prediction model is a logistic regression model, such as the formula: , in, is the probability of landslide, β 0 is the intercept of the model, β 1 , β 2 , …, β n is the regression coefficient of each feature; Specifically, the prediction model uses a logistic regression model, which is suitable for predicting the probability of occurrence of binary events and judging whether a landslide will occur. The logistic regression model is trained using historical highway data and is used to predict the possibility of a landslide on the highway. For example, the cumulative rainfall in the past 24 hours or 72 hours, the current moisture content in the soil, and the steeper slopes may increase the risk of landslides. A sudden increase in the groundwater level may cause geological instability, and the aging or damage of the slope protection may increase the risk of landslides. S202: acquiring real-time high-speed data, and obtaining the probability of landslide occurrence by using a trained prediction model according to the real-time high-speed data and meteorological environment data; Specifically, real-time highway data is obtained, and the collected data is processed using the trained prediction model. Whether a landslide has occurred is marked as a label variable, y=1 indicates that a landslide has occurred, and y=0 indicates that a landslide has not occurred. For example, the cumulative rainfall in the past 72 hours is: 150 mm, the current soil moisture is: 85%, the road slope is: 35 degrees, and the slope protection structure status is: there are signs of aging, but it has not been completely damaged. At this time, the probability of a landslide occurring in this highway section is input into the model, which is 65%; Specifically, real-time highway data includes road slope and slope protection structure status; S203: Perform risk assessment and classification on the probability of landslide occurrence to obtain a risk level; Specifically, risk assessment is conducted through probability results, and risk levels are divided into low risk, medium risk, high risk and extremely high risk. Low risk means that the probability of landslide is low and no special measures are required. Medium risk means that the probability of landslide or the impact is medium, and it is recommended to strengthen monitoring and make preventive preparations. High risk means that the probability of landslide is high or the impact is serious, and preventive measures need to be taken, such as restricting traffic or strengthening structural inspections. Extremely high risk means that the probability of landslide is extremely high and may cause major losses, and emergency measures must be taken immediately, such as closing roads and evacuating people. In this embodiment, the intercept β 0 = -5, rainfall β 1 = 0.02, soil moisture β 2 = 0.05, groundwater level change β 3 = 0.03, slope β 4 = 0.04, soil type β 5 = 1.2, slope protection structure state β 6 = 1.5, historical landslide record β 7 = 2.0; Substituting into the formula: ; The probability of landslide is close to 100%; Specifically, the risk levels include low risk, medium risk, high risk, and very high risk; Specifically, when the probability of landslide is less than 20%, the risk level is low risk; when the probability of landslide is not less than 20% and less than 50%, the risk level is medium risk; when the probability of landslide is not less than 50% and less than 75%, the risk level is high risk; when the probability of landslide is not less than 75%, the risk level is ultra-high risk; For example, in the past 72 hours, the cumulative rainfall on a highway section is 80 mm, the soil moisture is 60%, and the road slope is 20 degrees. The model calculates that the probability of landslide is 10%. In this case, the risk level is low risk; for example, the cumulative rainfall on a highway section in the past 72 hours is 120 mm, the soil moisture is 75%, and the slope is 30 degrees. The model calculates that the probability of landslide is 35%. In this case, the risk level is low risk; for example, the cumulative rainfall on a highway section in the past 72 hours is 180 mm, the soil moisture is 85%, and the slope is 35 degrees. The model calculates that the probability of landslide is 60%; for example, the cumulative rainfall on a highway section in the past 72 hours reached 220 mm, the soil moisture is 90%, and the slope is 40 degrees. The model calculates that the probability of landslide is 85%; Specifically, step S203 further includes: sending the risk level to the user terminal; S300: Acquire surrounding vehicle information, perform abnormality determination and single incident determination based on the surrounding vehicle information, and obtain reminder information; Specifically, Figure 3 The flowchart is shown as the steps of obtaining surrounding vehicle information through wireless communication technology and determining whether an abnormal situation occurs; step S300 includes: S301: Using wireless communication technology to obtain surrounding vehicle information, the surrounding vehicle information is used to determine whether there is an abnormal situation in the surrounding vehicles; Specifically, the above wireless communication technology is vehicle-to-vehicle communication V2V technology, which is a wireless communication technology that enables vehicles to exchange information with each other, such as speed, location, direction, etc. This technology can help monitor the status and behavior of surrounding vehicles in real time, and can transmit the vehicle's location information, speed, driving trajectory and acceleration in real time; Specifically, the wireless communication technology can also adopt mobile phone to vehicle communication V2P technology, so that the smart phone can communicate with the vehicle, usually through Bluetooth or wireless network, and when the vehicle detects potential danger or abnormality, it sends an alert to the driver through the mobile phone, allowing the user to share the current location and synchronize with other vehicles or traffic management systems; Specifically, using abnormal behaviors detected by sensors, such as sudden braking, sharp turns, abnormal driving trajectories, and long periods of stagnation, the vehicle analyzes these data through the onboard computer to determine whether there is an abnormality; S302: Perform abnormality determination based on surrounding vehicle information to obtain surrounding vehicle status; obtain vehicle information around the abnormal vehicle based on the surrounding vehicle status, perform single contingency determination of the event based on the vehicle information around the abnormal vehicle, and obtain a single contingency status of the event; Specifically, an abnormality determination is performed based on the surrounding vehicle information to obtain the state of the surrounding vehicles, wherein the state of the surrounding vehicles includes a normal state and an abnormal state; if the state of the surrounding vehicles is determined to be an abnormal state, for example, the vehicle in front brakes suddenly in a short period of time, the system will mark it as a potential problem, if the vehicle is stagnant in a non-parking area for a long time, the system will warn of a possible fault, if the vehicle in front frequently deviates from the lane, that is, the trajectory deviation may also be a sign of a problem, then the state of the vehicle is determined to be abnormal; after determining it to be an abnormal state, the vehicle information around the abnormal vehicle is obtained, that is, the driving trajectory of the vehicles around the abnormal vehicle is obtained through the mobile phone and vehicle communication V2P technology, when the driving trajectory of one of the vehicles changes, the driving trajectory of the remaining vehicles is obtained, and it is determined whether the driving trajectory of the remaining vehicles has changed, if the remaining vehicles have not changed their trajectories, it is determined to be a single accidental event, if the remaining vehicles have changed their trajectories, it is determined to be a non-single accidental event; the single accidental state of the above event includes a single accidental event and a non-single accidental event; S303: Obtain reminder information according to the single sporadic state of the event; Specifically, if the single contingency state of the event is determined to be a single contingency event, the reminder information is a non-landscape risk, and a reminder is sent to the rear vehicle terminal; if the single contingency state of the event is determined to be a non-single contingency event, the reminder information is a landslide risk in combination with the risk level, and an alarm signal is sent to the rear vehicle terminal; Specifically, if it is determined to be a single accidental event, it means that a certain vehicle has an accident. If it is determined not to be a landslide risk, a reminder will be sent to the rear vehicle terminal to remind that there is an accident vehicle ahead and to pay attention to driving safety; if it is determined to be a non-single accidental event, the driving trajectories of multiple groups of vehicles have changed, and the vehicle signals have disappeared. Since cars now have lane-level navigation, the vehicle will not sink to the bottom immediately after falling under the highway, and the navigation will not disappear immediately. The threshold is set to 20 seconds before the mobile phone signal or navigation is lost. If the vehicle signal disappears for more than 20 seconds, and the navigation software detects that the vehicle's driving trajectory has also undergone a large abnormality, combined with the predicted risk of landslide, it is determined that a landslide has occurred, and an alarm signal is sent to the rear vehicle terminal to stop the rear vehicle; Specifically, step S303 also includes: sending the reminder information to the user terminal; S400: according to the reminder information, obtaining the vehicle stopping information and the vehicle stopping sensor authority; according to the vehicle stopping sensor authority, using the information fusion technology, obtaining the danger range; Specifically, Figure 4 The flowchart is shown as the steps of obtaining vehicle information of a stopped vehicle and determining an abnormal range according to laser radar data and visual data; step S400 includes: S401: acquiring the stopped vehicle information according to the reminder information, and obtaining the stopped vehicle sensor authority according to the stopped vehicle information; Specifically, according to the reminder information, the vehicle information of the stopped vehicle is obtained, that is, the stopped vehicle information, the stopped vehicle is a vehicle not affected by the landslide, and according to the stopped vehicle information, the stopped vehicle sensor authority is obtained, the stopped vehicle sensor authority includes the vehicle-mounted laser radar authority and the visual authority; Specifically, the vehicle information of the stopped vehicle is obtained. When a landslide has been determined, the vehicle-mounted laser radar authority and visual authority of the vehicle stopped outside the landslide range are obtained to obtain the range of the landslide through the vehicle's laser radar and optical radar; S402: obtaining stopped vehicle sensor data according to the stopped vehicle sensor authority, fusing the stopped vehicle sensor data using wireless communication technology to obtain second data; obtaining the landslide range and landslide location according to the second data; Specifically, according to the stopped vehicle sensor authority, the stopped vehicle sensor data is obtained, and the stopped vehicle sensor data is integrated using wireless communication technology to obtain second data; according to the second data, the landslide range and the landslide position are obtained; Specifically, wireless communication technology is used to fuse multiple sets of LiDAR data and visual data to obtain the second data. After the landslide, there will be a black hole under the illumination of the car lights, and the road ahead will be different in the LiDAR, with low reflection and abnormality. When the vehicle approaches the landslide area, the on-board LiDAR system begins to scan the road ahead. LiDAR will emit laser pulses, measure the light returning from different angles, generate three-dimensional point cloud data of the landslide area, determine the abnormal range based on the second data, and identify the range and location of the landslide. Multiple vehicles approaching the landslide area can collect data at the same time, covering different angles and ranges, to ensure full coverage of the landslide area and form a more complete landslide area model. S403: Obtaining a danger range according to the landslide range and the landslide location; Specifically, the danger range is sent to the rear vehicle terminal, the landslide area model is preprocessed, noise and irrelevant data points such as other vehicles or non-landslide related objects are removed, and a three-dimensional topographic map of the landslide area is generated by processing the point cloud data; these maps can show the undulations of the surface and the specific form of the landslide; the data can be transmitted to the traffic management center or other vehicles in real time through vehicle-to-vehicle communication V2V technology or vehicle-to-infrastructure V2I communication technology, the danger range is uploaded to the terminal, the processing results are obtained, the length of the road section that needs to be closed and the detour route are determined, and the emergency rescue team is notified to rush to the scene to clear the landslide; Specifically, the vehicle information of the stopped vehicle is obtained. When the driving trajectories of multiple vehicles change and multiple groups of vehicles are stationary, the on-board laser radar authority and visual authority of the stopped vehicle are obtained, and the distance between the object and the sensor is accurately measured by emitting laser pulses and optical radar and measuring their return time; Vehicles approaching the landslide area through multiple groups can collect data simultaneously, covering different angles and ranges to ensure full coverage of the landslide area, identify the scope and location of the landslide, and send the danger range to the rear vehicle terminal to remind the rear vehicle of the distance from the landslide range; Specifically, step S403 also includes: sending the danger range to the user terminal.

[0022] This embodiment provides a vehicle driving warning system for sudden landslides on highways under severe conditions, such as Figure 5 The figure shows the architecture diagram of the vehicle driving warning system for sudden landslides on highways under adverse conditions; the system includes the following modules: a data acquisition module, configured to obtain vehicle information, historical highway data and meteorological environment data of the target highway section; a risk module, configured to build a prediction model, and obtain the risk level by using historical highway data, meteorological environment data, real-time highway data and the prediction model; an abnormality module, configured to obtain surrounding vehicle information, perform abnormality judgment and single event sporadic judgment based on the surrounding vehicle information, and obtain reminder information; a range module, configured to obtain stopped vehicle information and stopped vehicle sensor authority based on the reminder information; and obtain the danger range based on the stopped vehicle sensor authority using information fusion technology.

[0023] Specifically, Figure 6 Shown is an architecture diagram of the risk module; the risk module of the vehicle driving warning system for sudden landslides on highways under the above-mentioned severe conditions includes: a model unit, configured to: construct a prediction model, train the prediction model using historical highway data, and obtain a trained prediction model; a calculation unit, configured to: obtain real-time highway data, and obtain the probability of landslides based on the real-time highway data and meteorological environment data using the trained prediction model; an evaluation unit, configured to: perform risk evaluation and grade classification on the probability of landslides to obtain a risk level.

[0024] Specifically, the model unit of the vehicle driving warning system for sudden landslides on highways under the above-mentioned severe conditions is specifically configured as follows: constructing a prediction model, training the prediction model using historical highway data, and obtaining a trained prediction model; the prediction model is a logistic regression model, such as the formula: , in, is the probability of landslide, β 0 is the intercept of the model, β 1 , β2 , …, β n is the regression coefficient of each feature.

[0025] Specifically, Figure 7 The figure shows the architecture diagram of the abnormal module; the abnormal module of the vehicle driving warning system for sudden landslides on highways under the above-mentioned severe conditions includes: an abnormal unit, configured to: use wireless communication technology to obtain surrounding vehicle information, and the surrounding vehicle information is used to determine whether there is an abnormal situation with the surrounding vehicles; a trajectory unit, configured to: make an abnormality determination based on the surrounding vehicle information to obtain the surrounding vehicle status; based on the surrounding vehicle status, obtain the vehicle information around the abnormal vehicle, and make a single accidental determination of the event based on the vehicle information around the abnormal vehicle to obtain the single accidental status of the event; a determination unit, configured to: obtain reminder information based on the single accidental status of the event.

[0026] Specifically, Figure 8 The figure shows the architecture diagram of the range module; the range module of the vehicle driving warning system for sudden landslides on highways under the above-mentioned severe conditions includes: an authority unit, configured to: obtain stopped vehicle information according to the reminder information, and obtain stopped vehicle sensor authority according to the stopped vehicle information; a fusion unit, configured to: obtain stopped vehicle sensor data according to the stopped vehicle sensor authority, and use wireless communication technology to fuse the stopped vehicle sensor data to obtain second data; obtain the landslide range and landslide location according to the second data; a processing unit, configured to: obtain the danger range according to the landslide range and landslide location.

[0027] Embodiment 1: This embodiment provides a method for early warning of sudden landslides on highways under severe conditions, comprising the following steps: Step 100, obtaining high-speed vehicle information, the vehicle information includes vehicle position and vehicle speed, obtaining historical high-speed data and meteorological environment data, the historical high-speed data includes geological environment data and highway structure data, the meteorological environment data includes historical meteorological data, meteorological real-time monitoring data and meteorological forecast data; In this step, high-speed vehicle information is obtained. The global positioning system (GPS) technology is used to obtain the accurate position of each vehicle on the highway in real time, including latitude and longitude, lane information, and relative distance. The real-time speed information is obtained through the vehicle's speed sensor and GPS; this can help analyze the vehicle's driving status and determine whether it is necessary to slow down or take a detour; Understand the soil types along the highway, such as clay, sand, rock, etc. Different soil types react differently to rainfall and geological activities, affecting the risk of landslides. The slope of the highway, especially the steep slope, is prone to landslides after heavy rain. Real-time monitoring of current rainfall, especially in heavy rain and continuous rainfall, historical temperature and humidity data can help determine under what conditions soil saturation increases, leading to increased landslide risk. By acquiring and integrating highway vehicle information, historical highway data, geological environment data, and meteorological environment data, a comprehensive risk assessment can be conducted before a landslide occurs. Step 200, constructing a prediction model based on historical highway data, the parameters of the prediction model include topography, water system distribution, vegetation and land use type, highway structure and meteorological environment data, obtaining real-time highway data, calculating the current road collapse risk through the model, conducting risk assessment, and classifying the risk level; In this step, a prediction model is constructed based on historical highway data. The model can adopt a logistic regression model. The logistic regression model is suitable for predicting the probability of occurrence of binary events and judging whether a landslide will occur. The logistic regression model is trained by historical highway data. The logistic regression model is used to predict the possibility of landslides on highways. The logistic regression model P(y=1|X)= ,The parameters of the prediction model include topography and geomorphology, water system distribution, vegetation and land use type, highway structure form and meteorological environment data; The high-speed data acquired in real time is input into the logistic regression model to calculate the probability of a problem with the vehicle ahead and conduct risk assessment. Based on the calculated probability value, the system determines whether it is necessary to send an alarm to the vehicle behind and classify the risk level. Step 300, obtaining surrounding vehicle information through wireless communication technology, determining whether an abnormal situation occurs, the abnormal situation includes sudden deceleration, stopping, track deviation and signal disappearance, and sending an abnormal reminder to the rear vehicle terminal; In this step, the surrounding vehicle information is obtained through wireless communication technology. For example, vehicle-to-vehicle communication (V2V) is a wireless communication technology that enables vehicles to exchange information with each other, such as speed, location, direction, etc. This technology can help monitor the status and behavior of surrounding vehicles in real time, and can transmit the vehicle's location information, speed, driving trajectory and acceleration in real time; Abnormal behaviors detected by sensors, such as sudden braking, sharp turns, abnormal driving trajectories, and long periods of stagnation; the vehicle analyzes these data through the on-board computer to determine whether there are any abnormalities; If the vehicle ahead brakes suddenly within a short period of time, the system will flag it as a potential problem. If the vehicle is stationary in a non-parking area for a long time, the system will warn of a possible fault. Frequent lane deviations, i.e., trajectory deviations, may also be signs of problems. Step 400, obtaining vehicle information of the stopped vehicle, obtaining the on-board laser radar authority and visual authority of the stopped vehicle, determining the abnormal range according to the laser radar data and the visual data, and sending the dangerous range to the rear vehicle terminal; In this step, the vehicle information of the stopped vehicle is obtained. When the driving trajectories of multiple vehicles change and multiple groups of vehicles stop, the on-board laser radar authority and visual authority of the stopped vehicle are obtained. The distance between the object and the sensor is accurately measured by emitting laser pulses and optical radar and measuring their return time. Vehicles approaching the landslide area through multiple groups can collect data simultaneously, covering different angles and ranges to ensure full coverage of the landslide area, identify the scope and location of the landslide, and send the danger range to the rear vehicle terminal to remind the rear vehicle of the distance from the landslide range; like Figure 2 As shown, as a preferred embodiment of the present invention, a prediction model is constructed based on historical highway data, and the parameters of the prediction model include topography, water system distribution, vegetation and land use type, highway structure form and meteorological environment data. Real-time highway data is obtained, and the current road collapse risk is calculated through the model, and risk assessment is performed. The steps of dividing the risk level specifically include: Step 201, train a logistic regression model using historical highway data. The logistic regression model is used to predict the possibility of landslides on highways. The logistic regression model P(y=1|X)= ; In this step, the logistic regression model is trained by historical highway data to predict the possibility of landslides on the highway. The logistic regression model P(y=1|X)= ; Where P(y=1|X) is the probability of landslide, β 0 is the intercept of the model, β 1， β 2 , …, β n is the regression coefficient of each feature, such as topography, water system distribution, vegetation and land use type, highway structure and meteorological environment data, For example, you can check the accumulated rainfall in the past 24 hours or 72 hours, the current moisture content in the soil, steeper slopes may increase the risk of landslides, sudden increases in groundwater levels may cause geological instability, and aging or damage to slope protection may increase the risk of landslides; Step 202, obtaining real-time high-speed data, inputting the real-time high-speed data into a logistic regression model, calculating the probability of landslide occurrence, and obtaining a probability result; In this step, real-time highway data is obtained, the collected data is processed, and whether a landslide has occurred is marked as a label variable, y=1 means a landslide has occurred, and y=0 means no landslide has occurred. For example, the cumulative rainfall in the past 72 hours: 150 mm, the current soil moisture: 85%, the road slope: 35 degrees, the slope protection structure status: there are signs of aging, but not completely damaged; at this time, the probability of landslide occurring in this highway section is input into the model, which is 65%; Step 203, perform risk assessment based on the probability results, classify the risk levels, and send the risk levels to the user terminal; In this step, risk assessment is performed based on the probability results, and the risk levels are divided into low risk, medium risk, high risk and very high risk. Low risk means that the probability of landslide is low and no special measures are required. Medium risk means that the probability of landslide or the impact is medium, and it is recommended to strengthen monitoring and make preventive preparations. High risk means that the probability of landslide is high or the impact is serious, and preventive measures need to be taken, such as restricting traffic or strengthening structural inspections. Very high risk means that the probability of landslide is extremely high and may cause heavy losses, and emergency measures must be taken immediately, such as closing roads and evacuating people. For example, β0 (intercept) = -5, β 1 (Rainfall) = 0.02, β 2 (soil moisture) = 0.05, β 3 (groundwater level change) = 0.03, β 4 (slope) = 0.04, β 5 (soil type) = 1.2, β 6 (slope protection structure status) = 1.5, β 7 (historical landslide record) = 2.0; Substituting into the formula: z = −5 + (0.02 × 180) + (0.05 × 85) + (0.03 × 20) + (0.04 × 35) + (1.2 × 1) + (1.5 × 1) + (2.0 × 1) = 9.55; Then P(y=1|X)= ≈1, the probability of landslide is close to 100%; The low risk probability is less than 20%. For example, in the past 72 hours, the cumulative rainfall on a highway section is 80 mm, the soil moisture is 60%, and the road slope is 20 degrees; the model calculates that the probability of landslide is 10%; The probability of medium risk is 20%-50%. The cumulative rainfall of a road section in the past 72 hours is 120 mm, the soil moisture is 75%, and the slope is 30 degrees. The model calculates that the probability of landslide is 35%; The probability of high risk is 50%-75%. The cumulative rainfall in a certain road section in the past 72 hours is 180 mm, the soil moisture is 85%, and the slope is 35 degrees. The model calculates that the probability of landslide is 60%; The probability of ultra-high risk is greater than 75%. The cumulative rainfall in a certain road section in the past 72 hours reached 220 mm, the soil moisture was 90%, and the slope was 40 degrees. The model calculated that the probability of landslide was 85%; like Figure 3 and Fig. 9 As shown, as a preferred embodiment of the present invention, the steps of obtaining surrounding vehicle information through wireless communication technology, determining whether an abnormal situation occurs, the abnormal situation includes sudden deceleration, stopping and signal disappearance, and sending an abnormal reminder to the rear vehicle terminal specifically include: Step 301, obtaining surrounding vehicle information through wireless communication technology, the wireless communication technology adopts vehicle-to-vehicle communication, and determining whether an abnormal situation occurs; In this step, the surrounding vehicle information is obtained through wireless communication technology. For example, vehicle-to-vehicle communication (V2V) is a wireless communication technology that enables vehicles to exchange information with each other, such as speed, location, direction, etc. This technology can help monitor the status and behavior of surrounding vehicles in real time, and can transmit the vehicle's location information, speed, driving trajectory and acceleration in real time; At the same time, mobile phone-to-vehicle communication (V2P) can be used to enable smartphones to communicate with vehicles, usually through Bluetooth or wireless networks. When the vehicle detects potential danger or abnormality, an alert is sent to the driver through the mobile phone, allowing the user to share the current location and synchronize with other vehicles or traffic management systems; Use abnormal behaviors detected by sensors, such as sudden braking, sharp turns, abnormal driving trajectories, and long periods of stagnation, to determine whether an abnormal situation has occurred; Step 302, if it is determined to be an abnormal situation, obtain the vehicle information around the abnormal vehicle, obtain the vehicle driving trajectory around the abnormal vehicle, and determine whether it is a single accidental event; In this step, if it is determined to be an abnormal situation, if the vehicle ahead brakes suddenly in a short period of time, the system will mark it as a potential problem. If the vehicle is stationary in a non-parking area for a long time, the system will warn of a possible fault. Frequent deviations of the vehicle ahead from the lane, that is, trajectory deviation, may also be a sign of a problem. After determining it as an abnormal situation, it is also combined with the judgment of the probability of landslide; Obtain information about vehicles around the abnormal vehicle, obtain the driving trajectories of the surrounding vehicles through V2P, and determine the driving trajectories of the surrounding vehicles. When the driving trajectory of one of the vehicles changes, obtain the driving trajectories of the remaining vehicles to determine whether the driving trajectories of the remaining vehicles have changed. If the trajectories of the remaining vehicles have not changed, it is determined to be a single accidental event. If the trajectories of the remaining vehicles have changed, it is determined to be a non-single accidental event. Step 303: If it is determined to be a single accidental event, a reminder is sent to the rear vehicle terminal; if it is determined to be a non-single accidental event and is determined to be a landslide in combination with the risk level, an alarm signal is sent to the rear vehicle terminal; In this step, if it is determined to be a single accidental event, it means that a certain vehicle has an accident, and it is determined not to be a landslide risk, then a reminder is sent to the rear vehicle terminal to remind that there is an accident vehicle ahead and to pay attention to driving safety; If it is determined to be a non-single accidental event, the driving trajectories of multiple groups of vehicles change, and the vehicle signals disappear. Since cars now use lane-level navigation, the vehicles will not sink immediately after falling under the highway, and the navigation will not disappear immediately. In the period before the mobile phone signal or navigation is lost, the threshold is set to 20 seconds. If the vehicle signal disappears for more than 20 seconds, and the navigation software detects that the vehicle's driving trajectory has also undergone a major abnormality, combined with the predicted risk of landslides, it is determined that a landslide has occurred, and an alarm signal is sent to the rear vehicle terminal to stop the rear vehicle; like Figure 4 As shown, as a preferred embodiment of the present invention, the steps of obtaining vehicle information of a stopped vehicle, obtaining the on-board laser radar authority and visual authority of the stopped vehicle, determining the abnormal range according to the laser radar data, and sending the dangerous range to the rear vehicle terminal specifically include: Step 401, obtaining vehicle information of a stopped vehicle, where the stopped vehicle is a vehicle not affected by the landslide, and obtaining the onboard laser radar authority and visual authority of the stopped vehicle; In this step, the vehicle information of the stopped vehicle is obtained. When a landslide has been determined, the vehicle-mounted laser radar authority and visual authority of the vehicle stopped outside the landslide range are obtained. The range of the landslide can be obtained through the vehicle's laser radar and optical radar; Step 402, fusing multiple sets of laser radar data and visual data through wireless communication technology to obtain second data, determining the abnormal range based on the second data, and identifying the range and location of the landslide; In this step, multiple sets of LiDAR data and visual data are integrated through wireless communication technology to obtain the second data. After the landslide, there will be a black hole under the illumination of the car lights, and the road ahead will be different in the shape of the LiDAR, with low reflection and abnormality. When the vehicle approaches the landslide area, the on-board LiDAR system starts to scan the road ahead; LiDAR will emit laser pulses, measure the light returned from different angles, and generate three-dimensional point cloud data of the landslide area. The abnormal range is determined according to the second data, and the range and location of the landslide are identified. Multiple vehicles approaching the landslide area can collect data at the same time, covering different angles and ranges, to ensure full coverage of the landslide area and form a more complete landslide area model; Step 403, sending the danger range to the rear vehicle terminal, uploading the danger range to the terminal, and obtaining the processing result; In this step, the danger range is sent to the rear vehicle terminal, and the landslide area model is preprocessed to remove noise and irrelevant data points, such as other vehicles or non-landslide related objects. By processing the point cloud data, a three-dimensional topographic map of the landslide area is generated; these maps can show the undulations of the surface and the specific form of the landslide; The data can be transmitted in real time to the traffic management center or other vehicles through V2V or vehicle-to-infrastructure (V2I) communication technology, uploading the danger range to the terminal, obtaining the processing results, determining the length of the road section that needs to be closed and the detour route, and notifying the emergency rescue team to rush to the scene to clear the landslide.

[0028] Embodiment 2: In this embodiment, the traffic condition is: the pier of the highway bridge breaks at night, and the bridge deck falls into the river; the vehicle driving warning method for sudden landslide on the highway under severe conditions includes the following steps: Step 1: Obtain information about highway vehicles. Use the global positioning system (GPS) technology to obtain the exact location of each vehicle on the highway in real time, including latitude and longitude, lane information, and relative distance. Obtain real-time speed information through the vehicle's speed sensor and GPS. Obtain soil type and hydrological conditions along the highway to determine the risk of landslides. Obtain the slope of the highway, especially steep slopes, which are prone to landslides after heavy rain. Real-time monitoring of current rainfall, especially in the case of heavy rain or continuous rainfall. Historical temperature and humidity data can help determine under what conditions soil saturation increases, leading to an increased risk of landslides. By obtaining and integrating highway vehicle information, historical highway data, geological environment data, and meteorological environment data, a comprehensive risk assessment can be conducted before a landslide occurs. Step 2: Build a prediction model based on historical highway data. The model uses a logistic regression model. The logistic regression model is suitable for predicting the probability of occurrence of binary events and judging whether a landslide will occur. The logistic regression model is trained with historical highway data. The logistic regression model is used to predict the possibility of landslides on highways. The logistic regression model P(y=1|X)= ,The parameters of the prediction model include topography, water system distribution, vegetation and land use type, highway structure and meteorological environment data; the high-speed data acquired in real time is input into the logistic regression model to calculate the probability of problems with the vehicle ahead and conduct risk assessment. Based on the calculated probability value, the system determines whether it is necessary to issue an alarm to the vehicle behind and divide the risk level; Step 3: Obtain information about surrounding vehicles, such as speed, location, direction, etc., through vehicle-to-vehicle communication (V2V) technology, monitor the status and behavior of surrounding vehicles in real time, and be able to transmit the vehicle's location information, speed, driving trajectory and acceleration in real time; use sensors to detect abnormal behaviors, such as sudden braking, sharp turns, abnormal driving trajectory and long-term stagnation; the vehicle analyzes this data through the on-board computer to determine whether there is an abnormality; if the vehicle in front brakes suddenly in a short period of time, the system will mark it as a potential problem. If the vehicle is stagnant in a non-parking area for a long time, the system will warn of possible faults. Frequent lane deviations of the vehicle in front, that is, trajectory deviations, may also be signs of problems; Step 4: Obtain vehicle information of stopped vehicles. When the driving trajectories of multiple vehicles change and multiple groups of vehicles stop, obtain the on-board laser radar authority and visual authority of the stopped vehicles, and accurately measure the distance between the object and the sensor by emitting laser pulses and optical radars and measuring their return time; vehicles approaching the landslide area through multiple groups of vehicles can collect data at the same time, covering different angles and ranges to ensure full coverage of the landslide area, identify the range and location of the landslide, send the danger range to the rear vehicle terminal, and remind the rear vehicle of the distance from the landslide range; send the danger range to the rear vehicle terminal, pre-process the landslide area model, remove noise and irrelevant data points, such as other vehicles or non-landslide related objects, and generate a three-dimensional topographic map of the landslide area by processing point cloud data; these maps can show the undulations of the surface and the specific form of the landslide; the data can be transmitted to the traffic management center or other vehicles in real time through V2V or vehicle-to-infrastructure V2I communication technology, upload the danger range to the terminal, obtain the processing results, determine the length of the road section to be closed and the detour route, and notify the emergency rescue team to rush to the scene to clear the landslide.

[0029] Embodiment 3: In this embodiment, the traffic condition is: road collapse condition during the day; the vehicle driving warning method for sudden landslide on highway under severe conditions includes the following steps: Step (I): Obtain information about vehicles on the highway. Use the global positioning system (GPS) technology to obtain the exact location of each vehicle on the highway in real time, including latitude and longitude, lane information, and relative distance. Obtain real-time speed information through the vehicle's speed sensor and GPS. Obtain soil type and hydrological conditions along the highway to determine the risk of landslides. Obtain the slope of the highway, especially steep slopes, which are prone to landslides after heavy rain. Monitor the current rainfall in real time, especially in the case of heavy rain or continuous rainfall. Historical temperature and humidity data can help determine under what conditions soil saturation increases, leading to an increased risk of landslides. By obtaining and integrating information about vehicles on the highway, historical highway data, geological environment data, and meteorological environment data, a comprehensive risk assessment can be conducted before a landslide occurs. Step (2): Build a prediction model based on historical highway data. The model uses a logistic regression model. The logistic regression model is suitable for predicting the probability of occurrence of binary events and judging whether a landslide will occur. The logistic regression model is trained with historical highway data. The logistic regression model is used to predict the possibility of landslides on highways. The logistic regression model P(y=1|X)= ,The parameters of the prediction model include topography, water system distribution, vegetation and land use type, highway structure and meteorological environment data; the high-speed data acquired in real time is input into the logistic regression model to calculate the probability of problems with the vehicle ahead and conduct risk assessment. Based on the calculated probability value, the system determines whether it is necessary to issue an alarm to the vehicle behind and divide the risk level; Step (III): Obtain information about surrounding vehicles, such as speed, location, direction, etc., through vehicle-to-vehicle communication (V2V) technology, monitor the status and behavior of surrounding vehicles in real time, and be able to transmit the vehicle's location information, speed, driving trajectory and acceleration in real time; use sensors to detect abnormal behaviors, such as sudden braking, sharp turns, abnormal driving trajectory and long-term stagnation; the vehicle analyzes this data through the on-board computer to determine whether there is an abnormality; if the vehicle in front brakes suddenly in a short period of time, the system will mark it as a potential problem. If the vehicle is stagnant in a non-parking area for a long time, the system will warn of possible faults. Frequent lane deviations of the vehicle in front, that is, trajectory deviations, may also be signs of problems; Step (IV): Obtain vehicle information of stopped vehicles. When the driving trajectories of multiple vehicles change and multiple groups of vehicles stop moving, obtain visual authority of the stopped vehicles. According to the visual information of the stopped vehicles, manually assess the scope and location of the landslide, send the danger range to the rear vehicle terminal, remind the rear vehicle of the distance from the landslide range, and transmit it to the traffic management center or other vehicles in real time through V2V or vehicle-to-infrastructure V2I communication technology. Upload the danger range to the terminal, obtain the processing results, determine the length of the road section that needs to be closed and the detour route, and notify the emergency rescue team to rush to the scene to clear the landslide.

[0030] Embodiment 4: In this embodiment, the traffic conditions are: all the vehicles on the road are intelligent driving vehicles, and most of the vehicles have laser radar and intelligent driving functions; the vehicle driving warning method for sudden landslide on the highway under severe conditions includes the following steps: Step 10, obtain high-speed vehicle information, and use the global positioning system (GPS) technology to obtain the accurate position of each vehicle on the highway in real time, including latitude and longitude, lane information and relative distance, and obtain real-time speed information through the vehicle's speed sensor and GPS; obtain historical high-speed data and meteorological environment data, the historical high-speed data includes geological environment data and highway structure data, and the meteorological environment data includes historical meteorological data, meteorological real-time monitoring data and meteorological forecast data; specifically, including soil types along the highway, such as clay, sand, rock, etc. Different soil types react differently to rainfall and geological activities, affecting the risk of landslides. The slope of the highway, especially the steep slope section, is prone to landslides after heavy rain. Real-time monitoring of current rainfall, especially in the case of heavy rain and continuous rainfall, historical temperature and humidity data can help determine under what conditions soil saturation increases, leading to increased landslide risks. By obtaining and integrating high-speed vehicle information, historical high-speed data, geological environment data and meteorological environment data, a comprehensive risk assessment can be conducted before a landslide occurs; Step 20, construct a prediction model based on historical highway data. The parameters of the prediction model include topography, water system distribution, vegetation and land use type, highway structure and meteorological environment data. Real-time highway data is obtained, and the current road collapse risk is calculated through the model, and risk assessment is performed to classify the risk level. Step 30, obtain information about surrounding vehicles through vehicle-to-vehicle communication (V2V) technology to determine whether there are any abnormal conditions, including sudden deceleration, stopping, trajectory deviation and signal loss, and send abnormal reminders to the rear vehicle terminal; help monitor the status and behavior of surrounding vehicles in real time, and transmit the vehicle's location information, speed, driving trajectory and acceleration in real time; use sensors to detect abnormal behaviors, such as sudden braking, sharp turns, abnormal driving trajectory and long-term stagnation; the vehicle analyzes these data through the on-board computer to determine whether there are any abnormalities; if the vehicle in front brakes suddenly in a short period of time, the system will mark it as a potential problem. If the vehicle is stagnant in a non-parking area for a long time, the system will warn of possible faults. Frequent lane deviations of the vehicle in front, that is, trajectory deviations, may also be signs of problems; Specifically, step 30 includes: Step 31, obtain information about surrounding vehicles through vehicle-to-vehicle communication (V2V) technology to determine whether there are any abnormal conditions, such as speed, location, direction, etc.; this technology can help monitor the status and behavior of surrounding vehicles in real time, and can transmit the vehicle's location information, speed, driving trajectory and acceleration, etc. in real time; at the same time, mobile phone-to-vehicle communication (V2P) can be used to enable the smartphone to communicate with the vehicle, usually through Bluetooth or wireless network. When the vehicle detects potential danger or abnormality, it sends an alarm to the driver through the mobile phone, allowing the user to share the current location and synchronize with other vehicles or traffic management systems; use abnormal behaviors detected by sensors, such as sudden braking, sharp turns, abnormal driving trajectory and long-term stagnation, to determine whether there are any abnormal conditions; Step 32, if it is determined to be an abnormal situation, obtain the vehicle information around the abnormal vehicle, obtain the driving trajectory of the vehicles around the abnormal vehicle, and determine whether it is a single accidental event; if it is determined to be an abnormal situation, if the vehicle in front brakes suddenly in a short time, the system will mark it as a potential problem. If the vehicle is stagnant in a non-parking area for a long time, the system will warn of possible faults. Frequent deviations of the vehicle in front from the lane, that is, trajectory deviation, may also be a sign of a problem. After determining it to be an abnormal situation, it is combined with the judgment of the probability of landslide; obtain the vehicle information around the abnormal vehicle, obtain the driving trajectory of the surrounding vehicles through V2P, and determine the driving trajectory of the surrounding vehicles. When the driving trajectory of one of the vehicles changes, obtain the driving trajectory of the remaining vehicles to determine whether the driving trajectory of the remaining vehicles changes. If the trajectory of the remaining vehicles does not change, it is determined to be a single accidental event. If the trajectory of the remaining vehicles changes, it is determined to be a non-single accidental event; Step 33, if it is determined to be a single accidental event, a reminder is sent to the rear vehicle terminal; if it is determined to be a non-single accidental event, it is determined to be a landslide in combination with the risk level, and an alarm signal is sent to the rear vehicle terminal; that is, if it is determined to be a single accidental event, it means that a certain vehicle has an accident, and it is determined not to be a risk of landslide, a reminder is sent to the rear vehicle terminal to remind that there is an accident vehicle ahead and to pay attention to driving safety; if it is determined to be a non-single accidental event, the driving trajectories of multiple groups of vehicles change, and the vehicle signals disappear at the same time. Since cars now have lane-level navigation, after the vehicle falls under the highway, the vehicle will not sink immediately, and the navigation will not disappear immediately. In the period before the mobile phone signal or navigation is lost, the threshold is set to 20 seconds. If the vehicle signal disappears for more than 20 seconds, and the navigation software detects that the vehicle's driving trajectory has also undergone a large abnormality, combined with the predicted risk of landslide, it is determined that a landslide has occurred, and an alarm signal is sent to the rear vehicle terminal to stop the rear vehicle; Step 40, obtain the vehicle information of the stopped vehicle, obtain the on-board laser radar authority and visual authority of the stopped vehicle, determine the abnormal range according to the laser radar data and the visual data, and send the dangerous range to the rear vehicle terminal; specifically, obtain the vehicle information of the stopped vehicle, when the driving trajectories of multiple vehicles change and multiple groups of vehicles are stationary, obtain the on-board laser radar authority and visual authority of the stopped vehicle, and accurately measure the distance between the object and the sensor by emitting laser pulses and optical radars and measuring their return time; vehicles approaching the landslide area through multiple groups of vehicles can collect data at the same time, covering different angles and ranges, to ensure full coverage of the landslide area, identify the range and location of the landslide, send the dangerous range to the rear vehicle terminal, and remind the rear vehicle of the distance from the landslide range; Specifically, step 40 includes: Step 41, obtaining vehicle information of the stopped vehicle, the stopped vehicle is a vehicle not affected by the landslide, and obtaining the on-board laser radar authority and visual authority of the stopped vehicle; specifically, obtaining the vehicle information of the stopped vehicle, when a landslide has been determined, obtaining the on-board laser radar authority and visual authority of the vehicle stopped outside the range of the landslide, and being able to obtain the range of the landslide through the laser radar and optical radar of the vehicle; Step 42, fusing multiple sets of laser radar data and visual data through wireless communication technology to obtain second data, determining the abnormal range based on the second data, and identifying the range and location of the landslide; specifically, fusing multiple sets of laser radar data and visual data through wireless communication technology to obtain second data, after the landslide, there will be a black hole under the illumination of the car lights, and the road ahead will have a different shape in the laser radar, and the reflection will be very low, which is abnormal. When the vehicle approaches the landslide area, the on-board LiDAR system starts to scan the road ahead; LiDAR will emit laser pulses, measure the light returned from different angles, and generate three-dimensional point cloud data of the landslide area. Determine the abnormal range based on the second data, and identify the range and location of the landslide. Multiple vehicles approaching the landslide area can collect data at the same time, covering different angles and ranges, to ensure full coverage of the landslide area and form a more complete landslide area model; Step 43, sending the danger range to the rear vehicle terminal, and uploading the danger range to the terminal to obtain the processing result; specifically, sending the danger range to the rear vehicle terminal, pre-processing the landslide area model, removing noise and irrelevant data points, such as other vehicles or non-landslide related objects, and generating a three-dimensional topographic map of the landslide area by processing the point cloud data; these maps can show the undulations of the surface and the specific form of the landslide; the data can be transmitted to the traffic management center or other vehicles in real time through V2V or vehicle-to-infrastructure V2I communication technology, uploading the danger range to the terminal, obtaining the processing result, determining the length of the road section that needs to be closed and the detour route, and notifying the emergency rescue team to rush to the scene to clear the landslide.

[0031] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0032] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A method for early warning of sudden landslides on highways under severe conditions, characterized in that: The following steps are involved: S100: Acquire vehicle information, historical highway data, and meteorological environment data of the target highway section; S200: constructing a prediction model, using the historical high-speed data, the meteorological environment data, the real-time high-speed data and the prediction model to obtain a risk level; S300: Acquire surrounding vehicle information, perform abnormality determination and single incident determination based on the surrounding vehicle information, and obtain reminder information; S400: obtaining vehicle stop information and vehicle stop sensor authority according to the reminder information; According to the stopping vehicle sensor authority, the danger range is obtained by utilizing information fusion technology.

2. The method for early warning of sudden landslides on highways under severe conditions according to claim 1 is characterized in that: Step S200 specifically includes: S201: construct a prediction model, and train the prediction model using the historical high-speed data to obtain a trained prediction model; S202: acquiring real-time high-speed data, and obtaining the probability of landslide occurrence by using the trained prediction model according to the real-time high-speed data and the meteorological environment data; S203: Perform risk assessment and classification on the probability of occurrence of the landslide to obtain a risk level.

3. The method for early warning of sudden landslides on highways under severe conditions according to claim 2 is characterized in that: Step S201 specifically includes: constructing a prediction model, training the prediction model using the historical high-speed data, and obtaining a trained prediction model; the prediction model is a logistic regression model, such as the formula: ， in, is the probability of landslide, β0 is the intercept of the model, β1, β2, …, β n is the regression coefficient of each feature.

4. The method for early warning of sudden landslides on highways under adverse conditions according to claim 1 is characterized in that: Step S300 specifically includes: S301: Using wireless communication technology to obtain surrounding vehicle information, where the surrounding vehicle information is used to determine whether surrounding vehicles have abnormal conditions; S302: performing abnormality determination based on the surrounding vehicle information to obtain surrounding vehicle status; obtaining vehicle information around the abnormal vehicle based on the surrounding vehicle status, performing single contingency determination of the event based on the vehicle information around the abnormal vehicle to obtain a single contingency status of the event; S303: Obtain reminder information according to the single sporadic state of the event.

5. The method for early warning of sudden landslides on highways under adverse conditions according to claim 1, characterized in that: Step S400 specifically includes: S401: acquiring the stopped vehicle information according to the reminder information, and obtaining the stopped vehicle sensor authority according to the stopped vehicle information; S402: obtaining stopped vehicle sensor data according to the stopped vehicle sensor authority, fusing the stopped vehicle sensor data using wireless communication technology to obtain second data; obtaining a landslide range and a landslide location according to the second data; S403: Obtain a danger range according to the landslide range and the landslide position.

6. A vehicle driving warning system for sudden landslides on highways under severe conditions, characterized in that: The vehicle driving warning system for sudden landslides on highways under severe conditions includes the following modules: A data acquisition module is configured to: acquire vehicle information, historical highway data and meteorological environment data of a target highway section; The risk module is configured to: construct a prediction model, and obtain a risk level by using the historical high-speed data, the meteorological environment data, the real-time high-speed data and the prediction model; The abnormal module is configured to: obtain surrounding vehicle information, perform abnormality determination and single incident determination of the event according to the surrounding vehicle information, and obtain reminder information; A range module is configured to: obtain the vehicle stop information and the vehicle stop sensor authority according to the reminder information; According to the stopping vehicle sensor authority, the danger range is obtained by utilizing information fusion technology.

7. The vehicle driving warning system for sudden landslide on highway under severe conditions according to claim 6 is characterized in that: The risk module includes: A model unit is configured to: construct a prediction model, and train the prediction model using the historical high-speed data to obtain a trained prediction model; A computing unit is configured to: obtain real-time high-speed data, and obtain the probability of landslide occurrence by using the trained prediction model according to the real-time high-speed data and the meteorological environment data; The evaluation unit is configured to: perform risk evaluation and grade classification on the probability of the landslide to obtain a risk grade.

8. The vehicle driving warning system for sudden landslide on highway under severe conditions according to claim 7 is characterized in that: The model unit is specifically configured to: construct a prediction model, train the prediction model using the historical high-speed data, and obtain a trained prediction model; the prediction model is a logistic regression model, such as the formula: ， in, is the probability of landslide, β0 is the intercept of the model, β1, β2, …, β n is the regression coefficient of each feature.

9. The vehicle driving warning system for sudden landslides on highways under severe conditions according to claim 6 is characterized in that: The abnormal module includes: The abnormality unit is configured to: obtain surrounding vehicle information by using wireless communication technology, wherein the surrounding vehicle information is used to determine whether surrounding vehicles have abnormal conditions; The trajectory unit is configured to: perform abnormality determination based on the surrounding vehicle information to obtain the surrounding vehicle status; obtain vehicle information around the abnormal vehicle based on the surrounding vehicle status, perform single contingency determination of the event based on the vehicle information around the abnormal vehicle, and obtain the single contingency status of the event; The determination unit is configured to obtain reminder information according to the single sporadic state of the event.

10. The vehicle driving warning system for sudden landslide on highway under severe conditions according to claim 6, characterized in that: The scope module includes: The permission unit is configured to: obtain the stopped vehicle information according to the reminder information, and obtain the stopped vehicle sensor permission according to the stopped vehicle information; The fusion unit is configured to: obtain the stopped vehicle sensor data according to the stopped vehicle sensor authority, fuse the stopped vehicle sensor data using wireless communication technology to obtain second data; and obtain the landslide range and landslide location according to the second data; The processing unit is configured to obtain a dangerous range according to the landslide range and the landslide position.

Images