A road condition forecasting and warning method and system based on intelligent transportation

Abstract

The present invention relates to the field of traffic technology, and discloses a road condition forecasting and warning method and system based on intelligent traffic. By constructing a road feature network and an adaptive prediction model including road environment features, road historical traffic features and feature similarity values ​​between roads, the features of each road in the road feature network are learned through the adaptive prediction model, and the development trend of abnormal traffic events is predicted based on the identified abnormal traffic events, the real-time road information of the target road, and the road features of the target road and similar roads. According to the prediction results, the traffic diversion strategy is dynamically adjusted. The present invention enables the traffic management department to formulate a more reasonable traffic diversion strategy based on the historical experience of similar roads, improve the response efficiency of the entire traffic system, and effectively reduce the scope of accident hazards.

Description

Technical Field

[0001] The present invention relates to the field of traffic technology, and in particular to a road condition forecasting and warning method and system based on intelligent traffic. Background Art

[0002] Most of the current traffic warning systems focus on traffic flow anomalies or accident identification, and are unable to respond to complex and changeable traffic conditions in a timely and effective manner, especially in situations such as natural disasters or sudden traffic accidents at night, when the visibility of the development trend of the accident is low. However, it is crucial to adjust the traffic diversion strategy in time according to the subsequent development trend of the accident. The response speed and processing capacity of the existing traffic warning system are still greatly limited. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide a road condition forecasting and warning method and system based on intelligent transportation to solve the current problem of low visibility of subsequent development trends of accidents and inability to adjust diversion strategies in time to provide effective warnings to drivers in the process of dealing with traffic accidents.

[0004] The first aspect of the present invention discloses a road condition forecasting and warning method based on intelligent transportation, the method comprising the following steps:

[0005] Collecting road information; the road information includes road traffic information and road environment information;

[0006] Constructing a road feature network and an adaptive prediction model, and learning the features of each road in the road feature network through the adaptive prediction model; the road feature network includes road environment features, road historical traffic features, and feature similarity values ​​between each road;

[0007] Monitor and analyze the road traffic information and road environment information collected in real time, identify abnormal traffic events, determine the road where the abnormal traffic event is located as the target road, and determine the road whose similarity value with the target road reaches a preset threshold based on the road feature network as the first road;

[0008] Predicting the development trend of the abnormal traffic event based on the identified abnormal traffic event, the real-time road information of the target road, and the road characteristics of the target road and the first road through an adaptive prediction model;

[0009] According to the prediction results of the adaptive prediction model, the traffic diversion strategy is dynamically adjusted and warning information is sent to drivers;

[0010] The construction process of the road feature network includes learning the relevance of each road through a graph neural network, and extracting the environmental characteristics, historical traffic characteristics of each road and the characteristic similarity values ​​between roads;

[0011] The training process of the adaptive prediction model includes inputting the road feature network into the adaptive prediction model, and training the model based on historical abnormal traffic events, road information changes, and road feature network information through a deep learning algorithm and a time series prediction algorithm.

[0012] Furthermore, the road traffic information includes vehicle driving data, traffic flow data, and traffic camera data; the road environment information includes meteorological data, geological monitoring data, and soil moisture data, and the meteorological data also includes temperature data, air humidity data, rainfall data, and wind speed and direction data.

[0013] Furthermore, the abnormal traffic events include abnormal speed of traveling vehicles, vehicle stagnation, road obstacles, sudden accidents and natural disasters.

[0014] Furthermore, the development trend of the event includes the impact scope, duration and potential impact on traffic flow of the event.

[0015] Furthermore, the method further includes collecting and analyzing feedback data; the feedback data includes changes in traffic flow, changes in road environment, and duration of the event after the abnormal traffic event occurs;

[0016] Compare the feedback data with the prediction results of the adaptive prediction model and calculate the prediction error;

[0017] Based on the prediction error, model weights and parameters are adjusted through an adaptive optimization algorithm.

[0018] Furthermore, the dynamic adjustment of the traffic diversion strategy includes adjusting the diversion strategy in real time based on real-time road information of the target road.

[0019] Furthermore, the dynamic adjustment of the traffic diversion strategy also includes making adjustments according to the characteristics of the first road when it is determined that the first road has an abnormal traffic event similar to that of the target road.

[0020] The second aspect of the present invention discloses a road condition forecasting and warning system based on intelligent transportation. The system is implemented based on the method disclosed in the first aspect. The system includes a data acquisition module, a construction module, an identification module and a guidance module; wherein:

[0021] The data acquisition module is used to collect road information; the road information includes road traffic information and road environment information;

[0022] The construction module is used to construct a road feature network and an adaptive prediction model, and learn the characteristics of each road in the road feature network through the adaptive prediction model; the road feature network includes road environment characteristics, road historical traffic characteristics and characteristic similarity values ​​between each road;

[0023] The recognition module is used to monitor and analyze the road traffic information and road environment information collected in real time, identify abnormal traffic events, determine the road where the abnormal traffic event is located as the target road, and determine the road whose similarity value with the target road reaches a preset threshold based on the road feature network as the first road;

[0024] Predicting the development trend of the abnormal traffic event based on the identified abnormal traffic event, the real-time road information of the target road, and the road characteristics of the target road and the first road through an adaptive prediction model;

[0025] The traffic diversion module is used to dynamically adjust the traffic diversion strategy according to the prediction results of the adaptive prediction model and send warning information to the driver.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] The present invention constructs a road feature network and an adaptive prediction model including road environment features, road historical traffic features, and feature similarity values ​​between each road, learns the features of each road in the road feature network through the adaptive prediction model, and predicts the development trend of abnormal traffic events based on the identified abnormal traffic events, the real-time road information of the target road, and the road features of the target road and similar roads, and dynamically adjusts the traffic diversion strategy according to the prediction results. By combining the constructed road feature network with the adaptive prediction model, the road traffic conditions are analyzed and predicted in real time, and the feature similarity values ​​in the road feature network are used to quickly identify other roads similar to the target road, providing more accurate reference information for the prediction of the development trend of abnormal traffic events, and in this way, the traffic management department can formulate a more reasonable traffic diversion strategy based on the historical experience of similar roads, improve the response efficiency of the entire traffic system, and effectively reduce the scope of accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of the economic application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:

[0029] Figure 1 The present invention is a flowchart of a method for predicting and warning road conditions based on intelligent transportation disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments.

[0031] Embodiment 1

[0032] The first aspect of the present invention discloses a road condition forecasting and warning method based on intelligent transportation. Figure 1 , Figure 1 The present invention discloses a method for predicting and warning road conditions based on intelligent transportation, which includes the following steps:

[0033] Collect road information, which includes road traffic information and road environment information.

[0034] Specifically, in the embodiment of the present invention, the road traffic information includes but is not limited to vehicle driving data, traffic flow data, and traffic camera data. The road environment information includes but is not limited to meteorological data, geological monitoring data, and soil moisture data, and the meteorological data also includes temperature data, air humidity data, rainfall data, and wind speed and direction data.

[0035] As a preferred implementation mode of Example 1 of the present invention, the present invention further introduces a special road information collection mechanism for nighttime. When collecting vehicle driving data, it includes but is not limited to collecting the vehicle speed, lane departure information and braking frequency. Especially for low visibility at night, additional data such as the vehicle's automatic headlight opening status and headlight lighting intensity are collected to evaluate the interaction information between the vehicle and the road. Optionally, if installation conditions exist, road surface sensors, night visible light cameras and thermal imaging cameras are set to collect vehicle flow information at night, and infrared cameras and high-sensitivity light-sensing cameras with enhanced image acquisition capabilities under low-light conditions are used to ensure real-time monitoring of traffic flow, lane occupancy and emergencies under nighttime conditions, effectively overcoming the camera monitoring blind spots caused by insufficient light at night.

[0036] For road environment information, we further add data such as nighttime light intensity and nighttime road surface temperature changes to the nighttime environmental information collection to reflect the impact of nighttime driving conditions on traffic. In addition, if the installation conditions exist, ground sensors are set up to further collect information such as the slipperiness of nighttime roads to effectively assess the traffic safety risks that may be caused by slippery roads at night.

[0037] A road feature network and an adaptive prediction model are constructed, and the characteristics of each road in the road feature network are learned through the adaptive prediction model. The road feature network includes road environment characteristics, road historical traffic characteristics, and characteristic similarity values ​​between each road.

[0038] Monitor and analyze the road traffic information and road environment information collected in real time, identify abnormal traffic events, determine the road where the abnormal traffic event occurs as the target road, and determine the road whose similarity value with the target road reaches a preset threshold based on the road feature network as the first road.

[0039] Specifically, in the embodiment of the present invention, abnormal traffic events include, but are not limited to, abnormal speed of a moving vehicle, vehicle stagnation, road obstacles, sudden accidents, and natural disasters.

[0040] The development trend of the abnormal traffic event is predicted based on the identified abnormal traffic event, the real-time road information of the target road, and the road characteristics of the target road and the first road through an adaptive prediction model.

[0041] In the embodiment of the present invention, the development trend of the event includes but is not limited to the impact scope, duration and potential impact of the event on traffic flow.

[0042] According to the prediction results of the adaptive prediction model, the traffic diversion strategy is dynamically adjusted and early warning information is sent to drivers.

[0043] Furthermore, the construction process of the road feature network includes learning the correlation between each road through a graph neural network, and extracting the environmental characteristics, historical traffic characteristics of each road and the feature similarity values ​​between roads.

[0044] In the embodiment of the present invention, the road feature network includes the road environment features of different roads, such as mountain sections, tunnel sections, rainy sections, etc., and the road historical traffic accident features, such as accident type, subsequent development trend of the accident, accident rescue strategy, traffic diversion strategy and effect, etc. By assigning values ​​to different road environment features and road historical traffic features of each road, the similarity values ​​between different roads are calculated.

[0045] Specifically, in the process of learning the relevance of each road through the graph neural network, the road feature network is set to be a graph structure containing nodes and edges, where each node represents a road and the edge represents the relevance between different roads. Each node contains not only the basic information of the road, such as the road location, environmental characteristics (such as mountain sections, tunnel sections, rainy sections, etc.), but also historical traffic characteristics (such as traffic accident types, accident development trends, rescue strategies, traffic diversion strategies and their effects, etc.).

[0046] Since graph neural networks are suitable for processing complex graph structure data, the present invention uses the node feature aggregation mechanism of graph neural networks to learn the feature propagation and influence between different roads. The characteristics of each road node not only depend on its own environment and historical traffic information, but also are affected by the characteristics of its adjacent roads. Specifically, the graph neural network uses the message passing mechanism to propagate and aggregate each node through the connection between the edges. Through the iterative calculation of the multi-layer neural network, the characteristics of each road are fused with the information of the surrounding roads to obtain a comprehensive feature description. This feature contains the local and global characteristics of the road, so that the road feature network has good characterization capabilities.

[0047] In addition, in order to calculate the similarity between different roads, the present invention uses the road features extracted by the graph neural network, where the similarity depends not only on the static features of the road (such as geographical location, road environment), but also on dynamic features (such as historical traffic accidents, traffic diversion effects, etc.). Through the feature learning of the graph neural network, the feature similarity between roads is automatically calculated, and then the edge weights of the road feature network are generated. Roads with higher similarity values ​​mean that when similar events occur in the future, they have similar development trends and traffic impacts, which will help to optimize the diversion strategies for related roads in the future.

[0048] In the present invention, the training of the graph neural network is based on a large amount of historical traffic data and road information. Through retrospective analysis of a large number of abnormal traffic events, the weights in the graph neural network are continuously adjusted so that it can accurately capture the correlation and similarity between roads. The training process combines time series data (such as the temporal development process of traffic events) and road spatial characteristics (such as the geometric structure of the road and the surrounding environment). Through iterative optimization, the graph neural network can dynamically learn the changes in road characteristics and improve the adaptability of the road feature network.

[0049] Furthermore, the training process of the adaptive prediction model includes inputting the road feature network into the adaptive prediction model, and training it based on historical abnormal traffic events, road information changes, and road feature network information through a deep learning algorithm and a time series prediction algorithm.

[0050] Furthermore, the method further includes collecting and analyzing feedback data; the feedback data includes changes in traffic flow, changes in road environment, and duration of the event after the abnormal traffic event occurs;

[0051] Compare the feedback data with the prediction results of the adaptive prediction model and calculate the prediction error;

[0052] Based on the prediction error, model weights and parameters are adjusted through an adaptive optimization algorithm.

[0053] Specifically, in the embodiment of the present invention, the core of the adaptive prediction model is to realize the dynamic prediction of the development trend of abnormal traffic events by combining the deep learning algorithm and the time series prediction algorithm. The basic structure of the model can include the following parts:

[0054] Feature extraction unit: extracts key features from abnormal traffic events, road information, and road feature networks, such as event type, traffic flow, vehicle speed, environmental factors, and road similarity.

[0055] Time series prediction unit: Use time series prediction algorithms to model the time series of historical abnormal traffic events and predict the development trend of current events. The time series model predicts the impact range, duration, and changes in traffic flow of the event by learning the time evolution pattern after the abnormal traffic event occurs.

[0056] Adaptive learning unit: This unit uses an adaptive optimization algorithm to continuously adjust the model's weights and parameters based on actual feedback data (such as changes in traffic flow) to improve the accuracy of predictions.

[0057] Furthermore, the training process of the adaptive prediction model includes the following steps:

[0058] Data preparation: Normalize and preprocess historical abnormal traffic events, road traffic information, road environment information, and road feature network data to build training sets and test sets. The data of historical abnormal traffic events includes the time, location, type, development trend of the accident, and the implementation effect of traffic diversion strategies.

[0059] Feature input: Input the processed historical data into the adaptive prediction model.

[0060] Time series modeling: Through time series prediction algorithms (such as LSTM or GRU), historical abnormal traffic events are modeled. And through the recursive neural network structure, the model uses the development pattern of past abnormal events to predict the evolution trend of future traffic events.

[0061] Model training: During the training process, the model continuously adjusts its internal parameters to minimize the prediction error and uses optimization algorithms such as gradient descent to update the model parameters through back propagation.

[0062] Adaptive optimization: During the training and application of the model, the model's hyperparameters and weights are dynamically adjusted through the adaptive optimization algorithm to improve the model's prediction performance. Adaptive adjustments are made based on the actual effect of the traffic diversion strategy after execution to ensure that the model can maintain a high prediction accuracy under different traffic conditions.

[0063] Adaptive mechanism: During the application of the model, real-time traffic information and event feedback data (such as the execution effect of the diversion strategy) are input into the adaptive prediction model. The model analyzes these feedback data, calculates the prediction error, and then adjusts the model parameters through the adaptive optimization algorithm to enable it to better adapt to new traffic conditions. For example, when the traffic diversion strategy of a certain road is better than the prediction, the model will adjust its prediction parameters for similar roads based on this feedback.

[0064] As a preferred implementation of the first embodiment of the present invention, set X as a road feature matrix, then the road feature matrix X can be expressed as:

[0065]

[0066] in, For the The feature vector of each road; is the road feature matrix containing the feature information of all roads, is the number of roads; is the feature dimension.

[0067] Aggregate features through graph neural network:

[0068]

[0069] in, Indicated in Layer The feature vector of the road node, that is, the new feature of the node after being processed by the graph neural network; Indicates The neighbor node set of a road node is used to represent other nodes directly connected to the node; is the normalization factor, where For the The degree of the node, For the The degree of each node represents the number of neighboring nodes adjacent to it, which is used for normalization operation to ensure the uniform contribution of node features; For the The weight matrix of the layer; For the Tier The feature vector of each road node; For the The bias term of the layer; is the fusion weight; For the A collection of other roads that are similar to the road; is the similarity weight, indicating the The road node features are The influence of road node characteristics; Indicated in Layer The feature vector of each road node; is the activation function of nonlinear transformation.

[0070] Combining historical abnormal traffic events and road information for time series modeling:

[0071]

[0072] in for The hidden state of the moment; for Input features at each moment, including road traffic features and road environment features; for The weight parameters of the model; is the learning parameter; The feature inputs that affect the forecast results include but are not limited to weather features and time features.

[0073] The predicted output is:

[0074]

[0075] in, Indicates that from the current time step To the future The prediction output of traffic events at the time step; is the current time step A set of other road features similar to the target road; is the weight parameter of the output layer; is the nonlinear adjustment parameter of the output; is the time dynamic weight.

[0076] In the embodiment of the invention, the constructed road feature network is combined with the adaptive prediction model to analyze and predict the road traffic conditions in real time. The feature similarity values ​​in the road feature network can be used to quickly identify other roads similar to the target road, providing more accurate reference information for the prediction of the development trend of abnormal traffic events. In this way, the traffic management department can formulate more reasonable traffic diversion strategies based on the historical experience of similar roads, and improve the response efficiency of the entire traffic system.

[0077] Furthermore, the dynamic adjustment of the traffic diversion strategy includes adjusting the diversion strategy in real time based on real-time road information of the target road.

[0078] Furthermore, the dynamic adjustment of the traffic diversion strategy also includes making adjustments according to the characteristics of the first road when it is determined that the first road has an abnormal traffic event similar to that of the target road.

[0079] When the adaptive prediction model detects an abnormal traffic event on the input target road, it finds the first similar road based on the information of the road feature network and predicts the development trend of the abnormal event. Based on the prediction results, the traffic diversion strategy is dynamically adjusted. This adjustment not only considers the real-time data of the target road, but also refers to the historical traffic event handling experience of similar roads, so as to optimize the diversion strategy.

[0080] Embodiment 2

[0081] The second aspect of the present invention discloses a road condition forecasting and warning system based on intelligent transportation, the system includes a data acquisition module, a construction module, an identification module and a guidance module; wherein,

[0082] The data acquisition module is used to collect road information; the road information includes road traffic information and road environment information;

[0083] The construction module is used to construct a road feature network and an adaptive prediction model, and learn the characteristics of each road in the road feature network through the adaptive prediction model; the road feature network includes road environment characteristics, road historical traffic characteristics and characteristic similarity values ​​between each road;

[0084] The recognition module is used to monitor and analyze the road traffic information and road environment information collected in real time, identify abnormal traffic events, determine the road where the abnormal traffic event is located as the target road, and determine the road whose similarity value with the target road reaches a preset threshold based on the road feature network as the first road;

[0085] Predicting the development trend of the abnormal traffic event based on the identified abnormal traffic event, the real-time road information of the target road, and the road characteristics of the target road and the first road through an adaptive prediction model;

[0086] The traffic diversion module is used to dynamically adjust the traffic diversion strategy according to the prediction results of the adaptive prediction model and send warning information to the driver.

[0087] It should be noted that the specific implementation process of the second embodiment is similar to that of the first embodiment and will not be repeated in the second embodiment.

[0088] Finally, it should be noted that the road condition forecasting and warning method and system based on intelligent transportation disclosed in the embodiment of the present invention only discloses the preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, it should be understood by those skilled in the art that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A road condition forecasting and warning method based on intelligent transportation, characterized in that: The method comprises the following steps: Collecting road information; the road information includes road traffic information and road environment information; Constructing a road feature network and an adaptive prediction model, and learning the features of each road in the road feature network through the adaptive prediction model; the road feature network includes road environment features, road historical traffic features, and feature similarity values ​​between each road; Monitor and analyze the road traffic information and road environment information collected in real time, identify abnormal traffic events, determine the road where the abnormal traffic event is located as the target road, and determine the road whose similarity value with the target road reaches a preset threshold based on the road feature network as the first road; Predicting the development trend of the abnormal traffic event based on the identified abnormal traffic event, the real-time road information of the target road, and the road characteristics of the target road and the first road through an adaptive prediction model; According to the prediction results of the adaptive prediction model, the traffic diversion strategy is dynamically adjusted and warning information is sent to drivers; The construction process of the road feature network includes learning the relevance of each road through a graph neural network, and extracting the environmental characteristics, historical traffic characteristics of each road and the characteristic similarity values ​​between roads; The training process of the adaptive prediction model includes inputting the road feature network into the adaptive prediction model, and training the model based on historical abnormal traffic events, road information changes, and road feature network information through a deep learning algorithm and a time series prediction algorithm; The training process is: set X as the road feature matrix, then the road feature matrix X is expressed as: ； in, For the The feature vector of each road; is the road feature matrix containing the feature information of all roads, is the number of roads; is the feature dimension; Aggregate features through graph neural network: ； in, Indicated in Layer The feature vector of the road node, that is, the new feature of the node after being processed by the graph neural network; Indicates The neighbor node set of a road node is used to represent other nodes directly connected to the node; is the normalization factor, where For the The degree of the node, For the The degree of each node represents the number of neighboring nodes adjacent to it, which is used for normalization operation to ensure the uniform contribution of node features; For the The weight matrix of the layer; For the Tier The feature vector of each road node; For the The bias term of the layer; is the fusion weight; For the A collection of other roads that are similar to the road; is the similarity weight, indicating the The road node features are The influence of road node characteristics; Indicated in Layer The feature vector of each road node; is the activation function of nonlinear transformation; Combining historical abnormal traffic events and road information for time series modeling: ； in, for The hidden state of the moment; for Input features at each moment, including road traffic features and road environment features; for The weight parameters of the model; is the learning parameter; Input features that affect the forecast results, including weather features and time features; The predicted output is: ; in, Indicates that from the current time step To the future The prediction output of traffic events at the time step; is the current time step A set of other road features similar to the target road; is the weight parameter of the output layer; is the nonlinear adjustment parameter of the output; is the time dynamic weight.

2. The road condition forecasting and warning method based on intelligent transportation according to claim 1 is characterized in that: The road traffic information includes vehicle driving data, traffic flow data, and traffic camera data; the road environment information includes meteorological data, geological monitoring data, and soil moisture data. The meteorological data also includes temperature data, air humidity data, rainfall data, and wind speed and direction data.

3. The road condition forecasting and warning method based on intelligent transportation according to claim 1 is characterized in that: The abnormal traffic events include abnormal vehicle speed, vehicle stagnation, road obstacles, sudden accidents and natural disasters.

4. The road condition forecasting and warning method based on intelligent transportation according to claim 1 is characterized in that: The development trend of the event includes the impact scope, duration and potential impact on traffic flow.

5. The road condition forecasting and warning method based on intelligent transportation according to claim 1 is characterized in that: The method further includes collecting and analyzing feedback data; the feedback data includes changes in traffic flow, changes in road environment, and duration of the event after the abnormal traffic event occurs; Compare the feedback data with the prediction results of the adaptive prediction model and calculate the prediction error; Based on the prediction error, model weights and parameters are adjusted through an adaptive optimization algorithm.

6. The road condition forecasting and warning method based on intelligent transportation according to any one of claims 1 to 5, characterized in that: The dynamically adjusting the traffic diversion strategy includes adjusting the diversion strategy in real time based on real-time road information of the target road.

7. The road condition forecasting and warning method based on intelligent transportation according to claim 6 is characterized in that: The dynamically adjusting traffic diversion strategy also includes making adjustments based on the characteristics of the first road when it is determined that an abnormal traffic event similar to that on the target road exists on the first road.

8. A road condition forecasting and warning system based on intelligent transportation, the system is implemented based on the method described in claim 1, characterized in that: The system includes a data acquisition module, a construction module, an identification module and a guidance module; wherein, The data acquisition module is used to collect road information; the road information includes road traffic information and road environment information; The construction module is used to construct a road feature network and an adaptive prediction model, and learn the characteristics of each road in the road feature network through the adaptive prediction model; the road feature network includes road environment characteristics, road historical traffic characteristics and characteristic similarity values ​​between each road; The recognition module is used to monitor and analyze the road traffic information and road environment information collected in real time, identify abnormal traffic events, determine the road where the abnormal traffic event is located as the target road, and determine the road whose similarity value with the target road reaches a preset threshold based on the road feature network as the first road; Predicting the development trend of the abnormal traffic event based on the identified abnormal traffic event, the real-time road information of the target road, and the road characteristics of the target road and the first road through an adaptive prediction model; The traffic diversion module is used to dynamically adjust the traffic diversion strategy according to the prediction results of the adaptive prediction model and send warning information to the driver.

Images