A method and system for road intelligent induction and dynamic early warning with meteorological perception

Abstract

The application discloses a kind of road intelligent induction and dynamic early warning method and system fusing weather perception, it is related to intelligent transportation and road safety technical field.The method is by obtaining real-time weather, traffic and road data, using improved Kalman filtering and attention mechanism for multi-source data fusion, generate unified state estimation;Combining bayesian inference and Markov model for dynamic risk assessment, and based on safety, traffic efficiency and multi-objective optimization of energy consumption realizes speed limit adaptive adjustment;Further calculate the length and position of dynamic warning area, control intelligent spike and other equipment release induction information.System includes data acquisition, fusion estimation, risk prediction, speed adjustment, warning calculation and induction unit.The application realizes the real-time monitoring of road traffic environment under complex weather, risk prediction and intelligent control, improves driving safety and traffic efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent transportation systems and road safety control, and particularly relates to a road intelligent induction and dynamic warning method and system fusing meteorological perception. BACKGROUND

[0002] Under complex meteorological conditions such as dense fog, heavy rain, and snow, the road traffic environment changes dramatically, with significant shortening of the driver's visibility, decrease in road adhesion performance, and increase in vehicle braking distance, greatly increasing the probability of traffic accidents. In particular, at special road sections such as highways, bridges, and tunnels, due to complex terrain and weather conditions, sudden weather may cause more serious traffic safety hazards.

[0003] Currently, traditional traffic induction and warning systems mainly rely on cloud platforms or traffic management centers to provide static or periodically updated traffic information, and issue weather conditions and traffic tips to drivers through variable message signs, radio, and other means. However, such systems have obvious limitations in dealing with complex and changing weather environments, and are difficult to meet the requirements of real-time and precision, thereby affecting the effective guidance and safety control of driving behavior.

[0004] Specifically, the existing technology has the following main problems:

[0005] 1. Response lag, poor timeliness: the frequency of meteorological and traffic data collection is low, and the processing and transmission process is cumbersome, making it difficult for the system to quickly respond to changes in sudden weather events.

[0006] 2. Limited coverage, insufficient refinement: most systems only deploy warning devices at key road sections or critical nodes, lacking dynamic monitoring of the entire road network and refined management of local areas.

[0007] 3. Information transmission delay, missing linkage mechanism: the information transmission link from meteorological monitoring to road warning devices is long, and the system is difficult to achieve efficient coordination and rapid linkage between the cloud and the road terminal.

[0008] 4. Fixed strategy, poor adaptability: existing induction systems usually operate based on preset rules, and fail to develop differentiated warning and induction strategies according to different weather types such as heavy fog, heavy rain, and snow, and road section characteristics, resulting in poor induction effect in changing environments.

[0009] 5. Low level of intelligence, weak decision-making ability: most systems lack deep analysis capabilities for multi-source heterogeneous data, and have not introduced intelligent algorithms such as machine learning and Bayesian inference for risk prediction and adaptive optimization, making it difficult to achieve early judgment and dynamic regulation of traffic conditions under adverse weather conditions.

[0010] In summary, the existing traffic guidance and warning system faces the problems of slow response speed, incomplete coverage, single strategy and low intelligence level under complex weather conditions, and an intelligent traffic solution that can realize weather perception, risk assessment, speed limit regulation, dynamic warning and path guidance is urgently needed. SUMMARY

[0011] To solve the above problems, the application proposes a road intelligent guidance and dynamic warning method and system integrating weather perception; the method integrates weather monitoring, traffic state perception, dynamic risk assessment and intelligent guidance control functions to realize real-time monitoring, risk prediction and adaptive regulation of road traffic environment under complex weather conditions, thereby improving road traffic safety and operation efficiency.

[0012] The technical solution adopted by the application is:

[0013] A road intelligent guidance and dynamic warning method integrating weather perception, comprising the following steps:

[0014] S101 acquiring real-time data: collecting multi-source heterogeneous data of weather, traffic and road environment to provide basic input for intelligent analysis;

[0015] S102 data fusion estimation: based on multi-source heterogeneous data, using improved Kalman filtering combined with attention mechanism for data fusion to generate unified state estimation results;

[0016] S103 risk prediction and evaluation: based on the fused state estimation results, the current road section is dynamically evaluated and trend predicted by Bayesian inference and Markov chain model;

[0017] S104 real-time speed limit adjustment: based on the risk evaluation and trend prediction results, the speed limit value of the current road section is dynamically optimized with safety, traffic efficiency and energy consumption as objective functions;

[0018] S105 dynamic warning area calculation: according to the vehicle operating state and weather conditions, the appropriate dynamic warning area length and position are calculated;

[0019] S106 macro guidance information release: according to the results, the macro road network warning information is generated, the corresponding guidance scheme and information are released, and the studs are controlled for guidance to guide the driving vehicles.

[0020] Further, in S101 acquiring real-time data, the weather data, traffic data and road data are collected based on weather sensors, cameras, radars, floating car GPS and V2X communication;

[0021] The weather data includes visibility, precipitation intensity, wind speed and air temperature;

[0022] Traffic data includes vehicle speed, traffic flow, traffic density, acceleration;

[0023] Road data includes geometric characteristics, slope, curvature, historical accident rate.

[0024] Further, in the S102 data fusion estimation, the improved Kalman filter and attention mechanism are used to fuse the feature vectors of meteorological data, traffic data and road data, and a unified data group containing space-time information is generated, and the expression is:

[0025] ;

[0026] In the formula, is the predicted state value, which is used to construct the prior state of the filter, and is combined with to update the fusion; is the state transition matrix, which represents the change rule of the state over time; is the state value at the previous time; is the control input matrix, which represents the influence degree of the control input on the state; is the system control input of speed limit and induction control; is the noise of the state process and the measurement process; is the real-time measurement value of meteorological data, road data and traffic data; is the observation matrix, which represents the relationship between the true state and the observation value; is the attention weight, which represents the contribution degree of different data sources to the fusion result; is the weight matrix in the attention mechanism; is the road feature vector, including road geometric parameters; is the state estimation value at time t after fusion, that is, the state estimation result; is the gain matrix of Kalman filter, which dynamically adjusts the filter weight.

[0027] Further, in the S103 risk prediction and evaluation, the Bayesian inference model is used to output the risk posterior probability at the current time as the basis for risk level division;

[0028] Prior probability: represents the historical probability of traffic accidents or congestion on a certain road section under specific weather conditions; obtained based on historical data analysis, reflecting the influence of weather on risk;

[0029] Likelihood probability: ;

[0030] represents the possibility of traffic state under known weather conditions; The input comes from the state estimation result after fusion real-time traffic information extracted from the traffic state estimation result

[0031] Posterior probability: calculated by Bayes formula:

[0032]

[0033] where, is the conditional probability of traffic state given the current weather condition, which is calculated from the traffic state estimation result extracted from the traffic state estimation result is the prior probability of weather, which represents the likelihood of a certain weather occurring in a specific time period; is the marginal probability of traffic state, which represents the likelihood of a specific traffic state occurring.

[0034] Further, in the risk prediction evaluation of S103, the risk evolution trend at future time is predicted by Markov chain state transition modeling;

[0035] First, risk state definition is performed, and the road segment risk is divided into three levels: low risk, medium risk, high risk;

[0036] Then, state transition probability calculation is performed:

[0037]

[0038] where, is the probability of state transitioning to state ; is the fuzzy membership degree of state at time ; is the fuzzy membership degree of state at time ; , is the observed data of two consecutive time steps, which is the feature value combination extracted from the traffic state estimation result ; is the time step.

[0039] Further, in the risk prediction evaluation of S103, the steady-state risk probability of the road segment is also calculated by the fuzzy membership function, which is used as an input parameter for real-time speed limit adjustment;

[0040] First, the steady-state distribution is solved: through the state transition matrix of Markov chain , the steady-state distribution vector is solved: ;​​

[0041] wherein, the state the long-term stability probability, correspond to low, medium and high risk levels, respectively;

[0042] Then the fuzzy membership function is constructed: ;

[0043] The fuzzy membership function is constructed by comprehensively considering the visibility , precipitation intensity , traffic density variables, wherein the visibility , precipitation intensity , traffic density are the decomposition quantities of the state estimation result ;

[0044] Finally, the steady-state risk probability is calculated, and the calculation formula is:

[0045] ;

[0046] wherein, is the steady-state risk probability, indicating the overall risk level of the road section after comprehensively considering the historical trend and the current environmental factors.

[0047] Further, in the real-time speed limit adjustment S104, based on the risk assessment and trend prediction results, i.e. the steady-state risk probability , a multi-objective optimization function is constructed to calculate and issue control of the dynamic speed limit value of the current road section;

[0048] The multi-objective optimization function is as follows:

[0049] The objective function minimization: ;

[0050] Constraint condition: ;

[0051] ;

[0052] Weight dynamic adjustment mechanism: ;

[0053] wherein, is the target speed to be optimized; is the dynamic weight, which is dynamically adjusted with the steady-state risk probability ; is the maximum traffic flow of the road; is the actual traffic flow of the road under the current speed; is the actual energy consumption of the vehicle under the current speed; is the standard energy consumption at the reference vehicle speed; , is the upper and lower limit of the allowable vehicle speed; is a key variable representing the rate of change of road traffic flow over time; is the road traffic flow, i.e., the number of vehicles passing through a section per unit time; is time; is the sensitivity coefficient of traffic flow to changes in vehicle speed difference, i.e., the vehicle speed response coefficient; is the current average road speed; is the traffic density; is the risk sensitivity adjustment coefficient, which determines the degree of sensitivity of the weight to changes in risk, i.e., the risk sensitivity coefficient; is the summation index variable, =1 to 3 represent three different target items, =1 corresponds to the target safety, =2 corresponds to the target traffic efficiency, =3 corresponds to the target energy consumption control;

[0054] The optimal vehicle speed is output by minimizing the objective function Based on the optimal vehicle speed and the steady-state risk probability obtained by risk prediction and evaluation Generate control instructions for intelligent stud light color adjustment.

[0055] Further, in the S105 dynamic warning area calculation, according to the vehicle operating state after speed limit adjustment and the current meteorological environment, the appropriate warning area length and position are dynamically calculated;

[0056] The warning area length calculation formula is:

[0057] ;

[0058] In the formula, is the appropriate dynamic warning area length; is the weather adjustment factor, which dynamically adjusts the warning length under different weather conditions; is the number of vehicles in the vehicle group; is the real-time speed of the th vehicle after speed limit adjustment, ; is the maximum safe braking deceleration of the vehicle; is the safety factor; is the average reaction time of the driver; is the amplification coefficient of vehicle speed fluctuation on the warning area; is the standard deviation of the vehicle speed in the vehicle group; The amplification coefficient of the acceleration fluctuation to the early warning area is: The standard deviation of the acceleration in the vehicle group is:

[0059] The early warning area position calculation formula is:

[0060] The early warning starting point : ;

[0061] The early warning ending point : ;

[0062] In the formula, The real-time position of the vehicle; The environmental influence factor; The driver compensation factor.

[0063] A vehicle tracking system based on an intelligent road stud array is used in the above-mentioned road intelligent induction and dynamic early warning method based on meteorological perception, and the vehicle tracking system based on the intelligent road stud array comprises:

[0064] A data acquisition unit: acquiring real-time meteorological data, traffic data and road data;

[0065] A data fusion estimation unit: fusing meteorological data, traffic data and road data through a fusion algorithm to generate fused state estimation data;

[0066] A risk prediction unit: calculating the risk state under the current condition according to the fused state data through a Bayesian model and predicting the dynamic risk according to a Markov model;

[0067] A speed adjustment unit: based on the risk assessment result, adjusting the road speed limit in real time, with safety, traffic efficiency and energy consumption as the optimization targets;

[0068] An early warning area calculation unit: calculating the dynamic early warning area length of the corresponding road section according to the real-time vehicle speed after the speed limit, and calculating the early warning starting point and ending point position;

[0069] An induction unit: generating macroscopic road network early warning information according to the result, controlling the road stud to perform induction and guiding the driving vehicle.

[0070] An electronic device, comprising a memory, a processor, and a computer program stored on the memory and capable of running on the processor; wherein the processor implements the method steps of the above-mentioned road intelligent induction and dynamic early warning method based on meteorological perception when executing the computer program.

[0071] A computer readable storage medium, the computer readable storage medium has a computer program stored therein, the computer program is executed by a processor to realize the method steps of the above-mentioned road intelligent induction and dynamic warning method of fusion weather perception.

[0072] The application provides a road intelligent induction and dynamic warning method and system fusing weather perception, which realizes real-time monitoring, risk prediction and self-adaptive regulation of road traffic environment under complex weather conditions by integrating weather monitoring, traffic state perception, dynamic risk assessment and intelligent induction control functions, and significantly improves road traffic safety and operation efficiency. The following are the main beneficial effects:

[0073] 1. Realize multi-source data fusion and improve perception accuracy: use improved Kalman filtering and attention mechanism to model heterogeneous data of weather, traffic and road; introduce road feature vector to participate in the fusion process, enhance the adaptability of the model to special road sections; improve the accuracy and robustness of state estimation, and provide reliable data support for subsequent risk assessment.

[0074] 2. Build a dynamic risk assessment model to realize accurate risk identification: calculate the risk posterior probability of the current road section based on the Bayesian inference model; predict the evolution trend of future risk level combining with the Markov chain model; introduce fuzzy membership function and steady-state distribution analysis to output comprehensive risk indicators; support differentiated risk modeling under different weather conditions.

[0075] 3. Introduce multi-objective speed limit optimization mechanism, taking into account safety, efficiency and energy consumption: build a speed limit adjustment model with safety, traffic efficiency and energy consumption control as objective functions; dynamic weight mechanism automatically adjusts the priority of each sub-target according to the current risk level; realize the upgrade from fixed speed limit to adaptive speed limit, improve the ability to respond to sudden weather.

[0076] 4. Establish a dynamic warning area model to improve the accuracy of induction: dynamically calculate the length of the warning area according to vehicle speed, acceleration, driver reaction time and other parameters; introduce fluctuation compensation term and weather correction factor to enhance the robustness of the model; accurately calculate the starting point and ending point of the warning area for the start control of intelligent studs and other devices; realize the transition from fixed warning to dynamic guidance.

[0077] 5. Support macro road network induction to improve overall traffic efficiency: based on single-point warning, generate a macro-induction scheme for the entire road network; publish path recommendations through variable message boards, electronic signs, V2X communication and other means; support various induction strategies such as detour, split flow, lane guidance; effectively alleviate local congestion under adverse weather conditions and improve overall traffic flow stability.

[0078] 6. Good engineering adaptability and expansion potential: can be deployed in complex terrain sections such as highways, bridges, tunnels, and mountainous areas; supports edge computing architecture to meet low latency and high concurrency scenario requirements; modular design facilitates access to city traffic brain, autonomous driving system and other platforms; subsequent combination of reinforcement learning and large models can further improve intelligent level. BRIEF DESCRIPTION OF DRAWINGS

[0079] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art based on these drawings.

[0080] Figure 1 Flowchart of the road intelligent induction and dynamic warning method of the present application combined with weather perception;

[0081] Figure 2 Flowchart of risk prediction and evaluation of the present application;

[0082] Figure 3 Flowchart of dynamic warning area calculation of the present application;

[0083] Figure 4 Framework diagram of the vehicle tracking system based on the intelligent stud array of the present application;

[0084] Figure 5 Embodiment schematic diagram of the electronic device of the present application;

[0085] Figure 6 Embodiment schematic diagram of the computer readable storage medium of the present application. DETAILED DESCRIPTION

[0086] The technical solutions in the embodiments of the present application will be described clearly and completely with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0087] In order to improve the road traffic safety and operation efficiency under complex weather conditions, the present embodiment provides a road intelligent induction and dynamic warning method combined with weather perception; as shown in Figure 1 The road intelligent induction and dynamic warning method combined with weather perception includes the following steps:

[0088] S101, acquiring real-time data:

[0089] The multi-source heterogeneous data of weather, traffic and road environment are collected to provide basic input for intelligent analysis. Among them, the weather data, traffic data and road data are collected based on weather sensors, cameras, radars, floating car GPS and V2X communication.

[0090] The weather data includes visibility, precipitation intensity, wind speed and air temperature.

[0091] The traffic data includes vehicle speed, traffic flow, traffic density and acceleration.

[0092] The road data includes geometric characteristics, slope, curvature and historical accident rate.

[0093] S102 data fusion estimation:

[0094] Based on multi-source heterogeneous data, improved Kalman filtering combined with attention mechanism is used for data fusion to generate a unified state estimation result. By fusing data from different sources, a unified state estimation data with spatio-temporal characteristics is generated to improve the perception accuracy.

[0095] Specifically, improved Kalman filtering and attention mechanism are used to fuse the feature vectors of weather data, traffic data and road data to generate a unified data group containing spatio-temporal information, represented by:

[0096] ;

[0097] In the formula, is the predicted state value, which is used to construct the prior state of the filter, and is combined with for fusion update; is the state transition matrix, which represents the change rule of the state over time; is the state value at the previous time; is the control input matrix, which represents the influence degree of control input on the state; is the system control input of speed limit and induction control; is the noise of state process and measurement process; is the real-time measurement value of weather data, road data and traffic data; is the observation matrix, which represents the relationship between the true state and the observation value; is the attention weight, which represents the contribution degree of different data sources to the fusion result; is the weight matrix in the attention mechanism; is the road feature vector, including road geometric parameters; is the state estimation value at time t after fusion, i.e. the state estimation result; is the gain matrix of Kalman filtering, which dynamically adjusts the filtering weight.

[0098] wherein the road feature vector contains: the inverse of the radius of the road curve, which describes the curvature affecting the driving stability of the vehicle; the slope of the road, which describes the slope affecting the acceleration and deceleration ability of the vehicle; the road surface type, which describes the material such as cement / asphalt affecting the friction coefficient; the accident-prone point identifier, which describes whether it is a historical accident-prone section; the speed limit sign, which describes the original speed limit value of the current section; and the number of lanes, which describes the traffic capacity and the distribution of traffic density.

[0099] The fusion process mainly includes: a prediction stage, an observation update, an attention calculation, and a state correction. In the prediction stage, the state value at the previous moment is used to predict the state value at the current moment. In the observation update stage, real-time measurement values of meteorological data, road data, and traffic data are obtained. In the attention calculation stage, the real-time measurement values of meteorological data, road data, and traffic data are combined with the road feature vector to calculate the attention weight . In the state correction stage, the gain matrix of the Kalman filter and the attention-weighted residual error are used to correct the predicted state, and the state estimation value at the moment t after fusion is obtained.

[0100] The fusion algorithm improves the Kalman filter framework and introduces the attention mechanism, realizes the efficient fusion of multi-source heterogeneous data such as meteorological data, traffic data, and road data, and improves the state estimation accuracy and system robustness. Among them, the Kalman filter is used to model the state evolution and observation relationship; the attention mechanism enhances the model's attention ability to key data sources; and the road feature vector as context information participates in the fusion, improving the algorithm's adaptability to complex environments.

[0101] S103 risk prediction and evaluation:

[0102] As shown in Figure 2 , based on the state estimation result after fusion, the dynamic risk level evaluation and trend prediction of the current section are performed through Bayesian inference and Markov chain model. Through the dynamic risk level evaluation and trend prediction of the current section, decision basis can be provided for subsequent speed limit adjustment and guidance control.

[0103] Specifically, the Bayesian inference model in the embodiment is as follows:

[0104] Prior probability: represents the historical probability of traffic accidents or congestion occurring on a certain section under specific weather conditions such as heavy fog and heavy rain; which is obtained based on historical data analysis and reflects the influence of weather on risk.

[0105] Likelihood probability: ;

[0106] represents the likelihood of the traffic state occurring given the known weather conditions; Input from the fused state estimation results extracted real-time traffic information.

[0107] Posterior probability: calculated by Bayes' theorem:

[0108] ;

[0109] where, is the conditional probability of the traffic state given the current weather conditions, extracted from the state estimation results , which is the input term for evaluating the posterior probability; is the prior probability of the weather, representing the likelihood of a certain weather occurring within a specific time period; is the marginal probability of the traffic state, representing the likelihood of a specific traffic state occurring.

[0110] Finally, the posterior probability of risk at the current time is output by the Bayesian inference model as the basis for risk level classification.

[0111] The Markov chain state transition modeling in this embodiment is as follows:

[0112] First, risk state definition is performed, dividing the road segment risk into three levels: low risk, medium risk, high risk.

[0113] Then, state transition probability calculation is performed:

[0114] ;

[0115] where, is the probability of transitioning from state to state ; is the fuzzy membership degree of state at time ; is the fuzzy membership degree of state at time ; , is the observation data of two consecutive times, which is the feature value combination extracted from the state estimation results ; is the time step.

[0116] The risk evolution trend at the future time is predicted by the state transition probability calculation.

[0117] Further, the steady-state risk probability calculation is performed:

[0118] The steady-state risk probability of the road section is calculated by the fuzzy membership function as the input parameter of real-time speed limit adjustment.

[0119] First, the steady-state distribution is solved: the steady-state distribution vector of the Markov chain is solved by the state transition matrix ;

[0120] In the formula, represents the long-term stable probability of state , and corresponds to the low, medium and high risk levels respectively.

[0121] Then, the fuzzy membership function is constructed: ;

[0122] The fuzzy membership function considering the visibility , precipitation intensity , and traffic density variables, wherein the visibility , precipitation intensity , and traffic density are the decomposition quantities of the state estimation result .

[0123] Finally, the steady-state risk probability is calculated, and the calculation formula is:

[0124] ;

[0125] In the formula, is the steady-state risk probability, which represents the overall risk level of the road section considering the historical trend and the current environmental factors.

[0126] The risk prediction and evaluation realizes the quantitative evaluation and trend prediction of the road risk under complex weather conditions by combining Bayesian inference and Markov chain modeling; the Bayesian model is used to fuse weather and traffic state to generate the current risk posterior probability; the Markov chain can capture the evolution law of the risk state and support the prediction of the future risk; the fuzzy membership function enhances the adaptability of the model to continuous variables and uncertain environment; and the finally output steady-state risk probability provides a scientific basis for subsequent speed limit regulation and path induction.

[0127] S104 real-time speed limit adjustment:

[0128] ​Based on the risk assessment and trend prediction results, i.e., the steady-state risk probability The speed limit value of the current road section is dynamically optimized with safety, traffic efficiency and energy consumption as the objective functions, so as to improve road traffic safety, improve traffic efficiency and reduce vehicle energy consumption.

[0129] Specifically, the multi-objective optimization function is as follows:

[0130] Objective function minimization:

[0131] Constraint condition:

[0132]

[0133] Weight dynamic adjustment mechanism:

[0134] In the formula, is the target vehicle speed to be optimized; is a dynamic weight, which is adjusted dynamically with the steady-state risk probability ; is the maximum traffic flow of the road; is the actual traffic flow of the road at the current speed; is the actual energy consumption of the vehicle at the current speed; is the standard energy consumption at the reference speed; , is the upper and lower limit of the allowable speed; is a key variable, representing the rate of change of road traffic flow with time; is the road traffic flow, i.e., the number of vehicles passing through a certain section per unit time; is time; is the sensitivity coefficient of traffic flow to speed difference change, i.e., the speed response coefficient; is the current average speed of the road; is the traffic density; is the risk sensitivity adjustment coefficient, which determines the sensitivity of the weight to the risk, i.e., the risk sensitivity coefficient; is the summation index variable, =1 to 3 represent three different target items, =1 corresponds to the target safety, =2 corresponds to the target traffic efficiency, =3 corresponds to the target energy consumption control.

[0135] Among them, the three optimization objectives are: Safety, the lower the speed, the higher the safety coefficient, the larger the value; the weight increases when the risk is high, prompting the speed limit to decrease; ​​​​For traffic efficiency, it is used to measure the road utilization rate at the current speed, and the smaller the value is, the higher the efficiency is; For energy consumption control, it is used to measure the unit energy consumption level at the current speed, which is used for energy saving and emission reduction optimization.

[0136] According to the current risk level Automatically adjust the priority of the three targets; when the risk is high, Significantly increase, and the speed limit is automatically reduced; when the risk is low, And Predominate, pay more attention to traffic efficiency and energy saving; risk sensitivity adjustment coefficient Control the sensitivity of weight change, which can be set according to the actual scene.

[0137] Real-time speed limit adjustment outputs the optimal vehicle speed by minimizing the objective function ; Based on the optimal speed And the steady-state risk probability obtained by risk prediction and evaluation Generate control instructions for intelligent stud light color adjustment, and the specific mapping process is as follows:

[0138]

[0139] Suppose after optimization calculation: optimal speed limit ; Current steady-state risk probability ; The control instructions issued to the intelligent stud are as follows:

[0140] Instruction content: Numerical / form: Speed limit value 50 km / h; Stud color Yellow, indicating medium risk; Flashing frequency Medium frequency flashing, to alert the driver; Communication protocol V2X protocol frame + device ID + control byte

[0141] Real-time speed limit adjustment realizes dynamic adaptive adjustment of speed limit value by constructing a multi-objective optimization model combined with the risk evaluation result of the current road section , Its core advantage is: multi-objective collaborative optimization of safety, efficiency and energy consumption, weight automatically adjusted according to risk level, enhanced system robustness, integrated into intelligent transportation system, supported edge computing and cloud linkage, realized complete closed-loop process from data perception to control execution.

[0142] S105 dynamic warning area calculation:

[0143] As Figure 3 shown, according to the running state of the vehicle after speed limit adjustment such as speed and acceleration, and the current meteorological environment, the appropriate warning area length and position are dynamically calculated to ensure that the driver has enough reaction time and distance in complex weather, thereby effectively avoiding traffic accidents such as rear-end and side slip.

[0144] Specifically, according to the running state of the vehicle after speed limit adjustment and the current meteorological environment, the appropriate warning area length and position are dynamically calculated;

[0145] The length of the early warning area is calculated by the following formula:

[0146]

[0147] In the formula, is the appropriate dynamic early warning area length; is a weather adjustment factor, which dynamically adjusts the early warning length under different weather conditions; is the number of vehicles in the vehicle group; is the real-time speed of the first vehicle after speed limit, is the maximum safe braking deceleration of the vehicle; is the safety factor; is the average reaction time of the driver; is the amplification factor of speed fluctuation on the early warning area; is the standard deviation of the speed of the vehicle in the vehicle group; is the amplification factor of acceleration fluctuation on the early warning area; is the standard deviation of acceleration in the vehicle group.

[0148] Through the early warning area length calculation, the appropriate dynamic early warning area length is output ; Then, the early warning area position calculation is performed:

[0149] The early warning area position calculation formula is:

[0150] The early warning start point :

[0151] The early warning end point :

[0152] In the formula, is the real-time position of the vehicle; is the environmental influence factor; is the driver compensation factor.

[0153] Through the early warning area position calculation, the early warning start point and the early warning end point are output. The dynamic early warning area calculation dynamically calculates the early warning area length and position by establishing a mathematical model containing braking distance, fluctuation compensation, and weather correction, combined with the real-time state of the vehicle and environmental factors, to provide precise control instructions for road surface induction equipment, significantly improving driving safety and guidance efficiency under complex weather conditions.

[0154] S106 Macroscopic Induction Information Release:

[0155] ​​​​​According to the results, macroscopic road network early warning information is generated, corresponding induction scheme and information publishing are carried out, and the studs are controlled to guide the driving vehicles. The macroscopic road network early warning information includes dynamic adjustment of the stud color and information screen prompt.

[0156] In summary, the road intelligent induction and dynamic early warning method fusing meteorological perception, by acquiring real-time meteorological data, traffic data and road data, fusing the meteorological data, traffic data and road data through a fusion algorithm to generate fusion state estimation data, calculating the risk state under the current condition through a Bayesian model according to the fusion state data and predicting the dynamic risk according to a Markov model, adjusting the road speed limit in real time based on the risk assessment result, taking safety, traffic efficiency and energy consumption as optimization targets, calculating the dynamic early warning area length of the corresponding road section according to the real-time vehicle speed after the speed limit, calculating the early warning start and end positions, generating macroscopic road network early warning information according to the results, corresponding induction scheme and information publishing, controlling the studs to guide the driving vehicles. The present application provides a meteorological environment perception and road intelligent induction and warning method, which can realize real-time monitoring and early warning of severe weather in a wide range and precise warning of key road sections by integrating meteorological monitoring, real-time data transmission and intelligent induction functions.

[0157] On the other hand, to cooperate with the implementation of the road intelligent induction and dynamic early warning method fusing meteorological perception, the present embodiment provides a vehicle tracking system based on an intelligent stud array. As shown in the Figure 4 The vehicle tracking system based on the intelligent stud array includes:

[0158] The data acquisition unit 201 acquires real-time meteorological data, traffic data and road data; the data fusion estimation unit 202 fuses the meteorological data, traffic data and road data through a fusion algorithm to generate fusion state estimation data; the risk prediction unit 203 calculates the risk state under the current condition through a Bayesian model according to the fusion state data and predicts the dynamic risk according to a Markov model; the speed adjustment unit 204 adjusts the road speed limit in real time based on the risk assessment result, taking safety, traffic efficiency and energy consumption as optimization targets; the early warning area calculation unit 205 calculates the dynamic early warning area length of the corresponding road section according to the real-time vehicle speed after the speed limit, and calculates the early warning start and end positions; the induction unit 206 generates macroscopic road network early warning information according to the results, controls the studs to guide the driving vehicles.

[0159] In addition, the road intelligent induction and dynamic early warning method fusing meteorological perception can also be implemented through an electronic device or a computer readable storage medium; specifically, as shown in Figure 5 and Figure 6As shown, the electronic device 300 includes a memory 310, a processor 320, and a computer program 311 stored on the memory and executable on the processor; wherein the processor 320 implements the method steps of the above-mentioned road intelligent induction and dynamic early warning method fusing weather perception when executing the computer program 311. The computer program 411 is stored in the computer readable storage medium 400, and the computer program 411 implements the method steps of the above-mentioned road intelligent induction and dynamic early warning method fusing weather perception when executed by the processor.

[0160] The basic principles, main features and advantages of the present application are shown and described above. It should be understood by those skilled in the art that the present application is not limited by the above-mentioned embodiments, and the above-mentioned embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for road intelligent induction and dynamic early warning with meteorological perception, characterized in that, The method comprises the following steps: S101 acquiring real-time data: collecting multi-source heterogeneous data of weather, traffic and road environment to provide basic input for intelligent analysis; S102 data fusion estimation: based on multi-source heterogeneous data, using improved Kalman filtering combined with attention mechanism for data fusion to generate unified state estimation results; S103 risk prediction and evaluation: based on the fused state estimation results, using Bayesian inference and Markov chain model to evaluate and predict the trend of the dynamic risk level of the current road section; S104 real-time speed limit adjustment: based on the risk evaluation and trend prediction results, taking safety, traffic efficiency and energy consumption as objective functions, dynamically optimizing the speed limit value of the current road section; S105 dynamic warning area calculation: calculating the appropriate dynamic warning area length and position according to the vehicle operating state and weather conditions; S106 macroscopic guidance information release: generating macroscopic road network warning information according to the results, releasing corresponding guidance scheme and information, and guiding the driving vehicles through the control studs; In S105 dynamic warning area calculation, the appropriate warning area length and position are dynamically calculated according to the vehicle operating state after speed limit adjustment and the current meteorological environment; The warning area length calculation formula is: ； In the formula, is a suitable dynamic warning area length; is a weather adjustment factor, dynamically adjusting the warning length under different weather conditions; is the number of vehicles in the vehicle group; is the real-time speed of the first vehicle after speed limit, ; is the maximum safe braking deceleration of the vehicle; is the safety factor; is the average reaction time of the driver; is the amplification factor of vehicle speed fluctuation on the warning area; is the standard deviation of the vehicle speed in the vehicle group; is the amplification factor of acceleration fluctuation on the warning area; is the standard deviation of the acceleration in the vehicle group; The warning area position calculation formula is: Early warning starting point : ; Early warning end point : ; In the formula, is the real-time position of the vehicle; is the environmental impact factor; is the driver compensation factor.

2. The method of claim 1, wherein the method is characterized by: In S101 acquiring real-time data, the weather data, traffic data and road data are collected based on weather sensors, cameras, radars, floating car GPS and V2X communication; The weather data includes visibility, precipitation intensity, wind speed and air temperature; The traffic data includes speed, traffic flow, traffic density and acceleration; The road data includes geometric characteristics, slope, curvature and historical accident rate; In S102 data fusion estimation, the improved Kalman filtering and attention mechanism are used to fuse the feature vectors of weather data, traffic data and road data to generate a unified data group containing space-time information, and the expression is: ； In the formula, is the predicted state value, the prior state used to construct the filter, combined with fusion update; is the state transition matrix, representing the law of state change over time; is the state value at the previous time; is the control input matrix, representing the degree of influence of the control input on the state; is the system control input of speed limit and induction control; is the noise of the state process and the measurement process; is the real-time measurement value of meteorological data, road data and traffic data; is the observation matrix, representing the relationship between the true state and the observation value; is the attention weight, representing the contribution degree of different data sources to the fusion result; is the weight matrix in the attention mechanism; is the road feature vector, including road geometric parameters; is the state estimation value at time t after fusion, i.e. the state estimation result; is the gain matrix of Kalman filter, dynamically adjusting the filter weight. 3.The method of claim 1, wherein the method further comprises: determining a current location of the vehicle; and determining a current weather condition of the vehicle. In S103 risk prediction and evaluation, the risk posterior probability at the current time is output through the Bayesian inference model as the basis for risk level division; Prior probability: represents the historical probability of traffic accidents or congestion on a certain road segment under certain weather conditions; obtained based on historical data analysis, reflecting the impact of weather on risk; Likelihood probability: ; representing a likelihood of a traffic state occurring under known weather conditions; input from the fused state estimation results real-time traffic information extracted in the The posterior probability is calculated through the Bayesian formula: ； wherein is the conditional probability of the traffic state given the current weather conditions, estimated from the state estimation results extracted from the weather data as input terms for evaluating the posterior probability; is the prior probability of the weather, indicating the likelihood of a certain weather to occur in a certain time period; is the marginal probability of the traffic state, indicating the likelihood of a certain traffic state to occur.

4. The method of claim 1, wherein the method further comprises: determining a current location of the vehicle; and determining a current weather condition of the vehicle. In S103 risk prediction and evaluation, the risk evolution trend at the future time is predicted through Markov chain state transition modeling; First, the risk state definition is performed, and the road segment risk is divided into three levels: is low risk, is medium risk, is high risk; Then, the state transition probability is calculated: ； wherein is the probability of transition from state to state ; is the fuzzy membership of state at time ; is the fuzzy membership of state at time ; , is the observation data at two consecutive times, and is the feature value combination extracted from the state estimation result ; is the time step.

5. The method of claim 1, wherein the method further comprises: determining a current location of the vehicle; determining a current time; determining a current weather condition; determining a current traffic condition; and determining a current road condition. In the risk prediction evaluation of S103, the steady-state risk probability of the road section is calculated by a fuzzy membership function as an input parameter of real-time speed limit adjustment; First solve the steady state distribution: Solve its steady state distribution vector by the state transition matrix of Markov chain ;​ In the formula, indicates the state of long-term stability probability, respectively corresponding to low, medium and high risk levels; Then the fuzzy membership function is constructed: ; to comprehensively consider the visibility , precipitation intensity , traffic density variables after the fuzzy membership function, wherein the visibility , precipitation intensity , traffic density is the decomposition of the state estimation results ; Finally, the steady-state risk probability is calculated, and the calculation formula is: ； In the formula, is the steady-state risk probability, representing the overall risk level of the road section considering both historical trends and current environmental factors.

6. The method of claim 1, wherein the method further comprises: In the real-time speed limit adjustment S104, based on the risk assessment and trend prediction results, i.e., the steady-state risk probability , a multi-objective optimization function is constructed to calculate and issue control of the dynamic speed limit value for the current road section. The multi-objective optimization function is as follows: Objective function minimization: ; Constraints: ; ； Weight dynamic adjustment mechanism: ; wherein, is the target speed to be optimized; is the dynamic weight, which varies with the steady-state risk probability is dynamically adjusted; is the maximum traffic flow on the road; is the actual traffic flow on the road at the current speed; is the actual energy consumption of the vehicle at the current speed; is the standard energy consumption at the reference speed; , is the upper and lower limit of the allowed speed; is a key variable representing the rate of change of the road traffic flow over time; is the road traffic flow, i.e., the number of vehicles passing through a certain section per unit time; is time; is the sensitivity coefficient of traffic flow to speed difference change, i.e., the speed response coefficient; is the current average speed of the road; is the traffic density; is the risk sensitivity adjustment coefficient, which determines the sensitivity degree of the weight to the risk, i.e., the risk sensitivity coefficient; is the summation index variable, = 1 to 3 represent three different target items, = 1 corresponds to the target safety, = 2 corresponds to the target traffic efficiency, = 3 corresponds to the target energy consumption control; Output optimal vehicle speed by minimizing objective function ; based on optimal vehicle speed and steady state risk probability resulting from risk prediction evaluation Generate control instructions for intelligent stud light color adjustment.

7. A smart-stud array-based vehicle tracking system for the fusion of weather-aware road intelligent guidance and dynamic warning method of any one of claims 1-6, wherein, The vehicle tracking system based on the intelligent stud array comprises: A data acquisition unit for acquiring real-time weather data, traffic data and road data; A data fusion estimation unit for fusing weather data, traffic data and road data through a fusion algorithm to generate fused state estimation data; A risk prediction unit for calculating the risk state under the current condition through a Bayesian model according to the fused state data and predicting the dynamic risk through a Markov model; A speed adjustment unit for adjusting the road speed limit in real time based on the risk evaluation results, taking safety, traffic efficiency and energy consumption as optimization objectives; A warning area calculation unit for calculating the dynamic warning area length of the corresponding road section according to the real-time speed after speed limit adjustment and calculating the start and end positions of the warning area. The induction unit generates macroscopic road network early warning information according to the result, controls the studs to perform induction, and guides the driving vehicle.

8. An electronic device, comprising: The electronic device comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor; wherein the processor implements the method steps of the road intelligent induction and dynamic early warning method fusing weather perception according to any one of claims 1-6 when executing the computer program.

9. A computer-readable storage medium, characterized in that: The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the method steps of the road intelligent induction and dynamic early warning method fusing weather perception according to any one of claims 1-6.

Images