Traffic operation safety situation dynamic prediction system and method based on AI technology
Through the traffic operation safety situation dynamic prediction system based on AI technology, the problems of traffic safety hazards and low operational efficiency in the existing technology are solved, and safety incident monitoring and early warning, risk modeling and data analysis are realized, which improves the overall level of traffic operations. .
Patent Information
- Application Number
- CN202410728537.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-06
- Publication Date
- 2024-12-10
- Estimated Expiration
- 2044-06-06
AI Technical Summary
The existing technology cannot realize the functions of monitoring and early warning of safety incidents, modeling of traffic risk prediction models, processing and analysis of traffic data, and visual display of safety situations in traffic operations, resulting in large traffic safety hazards, low operational efficiency, low safety and decision-making quality.
Design a dynamic prediction system for traffic operation safety situation based on AI technology, including a prediction modeling module and a safety situation warning module. Through abnormal behavior detection, dynamic traffic event warning and comprehensive safety situation warning, combined with deep learning, timing models and machine learning, in real time Monitor and predict traffic safety situations and provide visual decision information through the decision support module.
Real-time monitoring and early warning of traffic safety events, accurate modeling of traffic risks and efficient data analysis are achieved, the safety and efficiency of traffic operations are improved, traffic scheduling and emergency decision-making are optimized, and the risk of accidents is reduced.
Smart Images
Figure CN118675323B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of traffic safety technology, and specifically to a dynamic prediction system and method for traffic operation safety situation based on AI technology. Background Art
[0002] With the acceleration of urbanization and the continuous growth of transportation demand, modern urban transportation faces many challenges and problems. Frequent traffic accidents, serious congestion, poor traffic control and other phenomena not only affect the efficiency of traffic operations, but also threaten traffic safety and the quality of life of urban residents. Traditional traffic management methods have been unable to cope with the complex and changing traffic operation environment. There is an urgent need for an intelligent traffic management system to improve safety, reduce congestion, optimize traffic scheduling, and achieve the goal of smart city transportation.
[0003] Therefore, it is very necessary to propose a dynamic prediction system and method for traffic operation safety situation based on AI technology in this application to solve the problems of large traffic safety hazards, low traffic operation efficiency, low safety, and low decision-making efficiency and quality in existing traffic operations.
[0004] 1. Patent document CN116935700B discloses a sector traffic situation prediction method based on multi-source features. The above patent achieves comprehensive and effective capture of multi-source features of air traffic situation and recognition and prediction of situation level, but the above patent cannot realize the monitoring and early warning function of safety incidents in traffic operations.
[0005] 2. Patent document CN115049167B discloses a traffic situation prediction method, device, equipment and storage medium. The above patent realizes the ability to inform users who are about to travel of the prediction results, reduce the probability of traffic congestion by reducing travel demand, and improve the traffic travel experience, but the above patent cannot realize the modeling function of the traffic risk prediction model.
[0006] 3. Patent document CN109191852B discloses a vehicle-road-cloud collaborative traffic flow situation prediction method. The above patent realizes the real-time prediction of traffic flow situation, thereby providing effective support for improving traffic flow management efficiency and reducing energy expenditure. It has good applicability, but the above patent cannot realize the processing and analysis function of traffic data.
[0007] 4. Patent document CN113140108B discloses a cloud-based traffic situation prediction method in a networked intelligent transportation system. The above patent realizes the rapid prediction of the status of the traffic network, provides support for the path planning of smart cars, the autonomous guidance and management of traffic flow, and the priority guidance of emergency vehicles, but the above patent cannot realize the visual display function of safety situation and emergency decision-making.
[0008] In summary, the above patents cannot realize the monitoring and early warning functions of safety incidents in traffic operations, the modeling function of traffic risk prediction models, the processing and analysis functions of traffic data, and the visual display functions of safety situations and emergency decisions, resulting in major traffic safety hazards, low traffic operation efficiency, low safety, and low decision-making efficiency and quality.
[0009] To this end, the present application proposes a dynamic prediction system and method for traffic operation safety situation based on AI technology, which can realize the monitoring and early warning functions of safety incidents in traffic operations, the modeling function of traffic risk prediction models, the processing and analysis function of traffic data, and the visualization display function of safety situation and emergency decision-making. Summary of the invention
[0010] The purpose of the present invention is to provide a dynamic prediction system and method for traffic operation safety situation based on AI technology, so as to solve the technical problems proposed in the above background technology that the monitoring and early warning functions of safety incidents in traffic operations, the modeling function of traffic risk prediction models, the processing and analysis functions of traffic data, and the visualization display function of safety situation and emergency decision-making cannot be realized, resulting in great traffic safety hazards, low traffic operation efficiency, low safety, and low decision-making efficiency and quality.
[0011] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a dynamic prediction system for traffic operation safety situation based on AI technology, comprising a prediction modeling module and a safety situation warning module, wherein the prediction modeling module provides a detection warning model for the safety situation warning module, and the safety situation warning module is used to predict the risk of traffic operation safety situation;
[0012] The safety situation warning module includes: an abnormal behavior detection warning unit, a dynamic traffic event warning unit and a comprehensive safety situation warning unit, and the results of the abnormal behavior detection warning unit and the dynamic traffic event warning unit are output as comprehensive analysis data of the comprehensive safety situation warning unit;
[0013] The abnormal behavior detection and warning unit detects abnormal behaviors in traffic scenes in real time through the CNN deep learning model, and immediately issues a warning notification when abnormal behaviors are found, supplemented by video evidence and trajectory data;
[0014] The dynamic traffic incident warning unit combines real-time data and historical event statistics, uses time series models and probability models to predict potential traffic accidents, and makes early warnings and dispatch arrangements;
[0015] The comprehensive safety situation warning unit comprehensively considers the abnormal behavior detection, traffic event prediction and weather factor prediction results to form comprehensive safety situation warning information to support system decision-making and user emergency response.
[0016] Preferably, the prediction modeling module is used to establish multiple risk prediction models and continuously update model parameters through real-time data;
[0017] The prediction modeling module includes: a traffic flow prediction modeling unit, a traffic accident risk prediction unit and a weather factor impact prediction unit. The model data output of the traffic flow prediction modeling unit, the traffic accident risk prediction unit and the weather factor impact prediction unit is used as the warning model database of the safety situation warning module;
[0018] The traffic flow prediction modeling unit establishes a traffic flow prediction model based on the long short-term memory network model to predict traffic flow and congestion in the future and continuously updates the model parameters through real-time data;
[0019] The traffic accident risk prediction unit uses the support vector machine algorithm to establish a traffic accident risk prediction model to analyze the possibility and influencing factors of traffic incidents;
[0020] The weather factor impact prediction unit combines machine learning algorithms and meteorological data to model and predict the impact of different weather factors on traffic operation safety, and take countermeasures in advance.
[0021] Preferably, the prediction system is designed with a data acquisition module, which is used to collect traffic status and meteorological status data in real time;
[0022] The data acquisition module includes: a sensor data acquisition unit, a traffic event acquisition unit and a meteorological environment data acquisition unit;
[0023] The sensor data acquisition unit uses advanced high-definition cameras, millimeter-wave radars, and lidars to collect real-time data from traffic scenes in real time;
[0024] The traffic event collection unit combines video surveillance and image recognition technology to monitor traffic events in real time and obtain real-time traffic event data from the traffic management center;
[0025] The meteorological and environmental data collection unit integrates the weather data and environmental monitoring data provided by the Meteorological Bureau to analyze the impact of weather on traffic operations. The meteorological and environmental data collection unit transmits the collected data to the weather factor impact prediction unit via the Internet.
[0026] Preferably, the data acquisition module is connected to a data processing and analysis module provided in the system via the Internet, the data output of the data processing and analysis module is used as the modeling parameter input of the prediction modeling module, and the data processing and analysis module is used to process and analyze the comprehensive data collected by the data acquisition module;
[0027] The data processing and analysis module includes: a data cleaning preprocessing unit, a traffic situation perception unit and a safety situation analysis unit. The data output of the data cleaning preprocessing unit serves as the analysis input of the traffic situation perception unit and the safety situation analysis unit.
[0028] The data cleaning preprocessing unit receives the raw data from the sensor data collection unit, the traffic event collection unit and the meteorological environment data collection unit, processes the raw data collected by the sensor data collection unit through data cleaning and denoising technology, and formats and standardizes the data;
[0029] The traffic situation awareness unit receives the traffic data processed by the data cleaning preprocessing unit, uses the deep learning algorithm to perform real-time analysis on the processed data, identifies traffic flow, vehicle driving trajectory, congestion and abnormal vehicle behavior, and forms awareness of the traffic situation;
[0030] The safety situation analysis unit receives the processed traffic data and the information provided by the traffic situation awareness unit, collects historical data and real-time situation, uses machine learning models to analyze the traffic safety status, and identifies potential safety hazards and risks. The safety situation analysis unit transmits the analysis data to the traffic flow prediction modeling unit and the traffic accident risk prediction unit via the Internet.
[0031] Preferably, the safety situation warning module is connected to a decision support module via the Internet, the decision support module receives the warning information output of the safety situation warning module as decision data input, and the decision support module is used to optimize traffic operations and visualize decision information display;
[0032] The decision support module includes: a safety emergency decision support unit, a traffic dispatch optimization unit and a safety situation visualization display unit. The data output of the safety emergency decision support unit and the traffic dispatch optimization unit is used as the display database source of the safety situation visualization display unit.
[0033] The safety emergency decision support unit provides safety emergency decision support based on early warning information to optimize traffic operation safety;
[0034] The traffic dispatch optimization unit combines the prediction model provided by the prediction modeling module and the real-time data provided by the data processing and analysis module to optimize traffic dispatch;
[0035] The security situation visualization display unit displays the prediction results of the security situation warning module and the security situation information of the security emergency decision support unit and the traffic dispatch optimization unit in a visual manner, providing decision makers with intuitive traffic safety situation information.
[0036] Preferably, the CNN deep learning model used by the abnormal behavior detection and early warning unit is expressed as:
[0037]
[0038] Among them, y is the output, f is the activation function, and w i is the weight, x i is the input, b is the bias term;
[0039] Specific model architecture: Input layer: image and video data in traffic scenes; Convolution layer and pooling layer: used for feature extraction and dimensionality reduction; Activation function layer: introduces nonlinear factors; Fully connected layer: maps the extracted features to the final output layer; Output layer: warning results of abnormal behavior.
[0040] Preferably, the long short-term memory network model used by the traffic flow prediction modeling unit is expressed as: the forget gate output f t =σ(W f ·[h t-1 ,x t ]+b f ), the input gate output i t =σ(W i ·[h t-1 ,x t ]+b i ), the output gate outputs o t =σ(W o ·[h t-1 ,x t ]+b o ), the output of the new candidate unit Cell state Output h t =o t *tanh(C t );
[0041] Specific model architecture: Input layer: time series traffic flow data; LSTM layer: used to process sequence data and remember long-term dependencies; Output layer: predict traffic flow and congestion in the future.
[0042] Preferably, the support vector machine algorithm formula used by the traffic accident risk prediction unit is as follows:
[0043]
[0044] y i (w·x i +b)≥1-ξ i ,ξ i ≥0
[0045] Among them, w is the normal vector of the hyperplane, b is the bias term, C is the penalty parameter, and x i is the sample data, y iis the sample label, ξ i is a slack variable, N is the number of samples, and i is the sample number;
[0046] Specific model architecture: Input layer: feature vector of traffic data; Output layer: risk prediction probability.
[0047] Preferably, the method comprises the following steps:
[0048] S1. First, use the data acquisition module to obtain the vehicle speed, density, trajectory and weather real-time data in the traffic scene;
[0049] S2, then extracting characteristic data of traffic flow, congestion level and vehicle behavior pattern from the cleaned data through the data processing and analysis module;
[0050] S3, then use the modeling and prediction module to model the traffic flow, predict the traffic flow and congestion in the future, and provide a reference for the decision support module;
[0051] S4. Finally, the safety situation warning module detects abnormal behaviors in traffic scenes in real time, issues warning notifications and provides evidence.
[0052] Preferably, the method further comprises the following steps:
[0053] S21. Use the AI algorithm in the security situation analysis unit to perform security situation analysis and identify potential security risks and risk factors;
[0054] S31, establishing a traffic accident risk prediction model through a support vector machine algorithm in a traffic accident risk prediction unit to predict the location and time of a possible accident so as to take preventive measures and emergency treatment;
[0055] S41. Utilize the comprehensive safety situation warning unit to combine real-time data and prediction results to dynamically warn of the traffic operation safety situation and provide it to the decision support module so as to take real-time traffic control and scheduling optimization.
[0056] Compared with the prior art, the present invention has the following beneficial effects:
[0057] 1. The present invention realizes the monitoring and early warning function of safety events in traffic operations by designing a safety situation early warning module, solves the problem that the existing technology cannot provide early warning notifications for safety accidents, improves the traffic accident early warning and emergency response capabilities, and reduces the risk of accidents;
[0058] 2. The present invention realizes the modeling function of the traffic risk prediction model by designing a prediction modeling module, solves the problem that the impact of traffic flow, accident risk and weather factors on traffic operations cannot be modeled and predicted, improves the efficiency and safety of traffic operations, and optimizes traffic scheduling and emergency response plans;
[0059] 3. The present invention realizes the processing and analysis function of traffic data by designing a data processing and analysis module, solves the problem that traffic data cannot be processed and analyzed to obtain important information and features, improves the efficiency of data processing and analysis, reduces the workload of decision makers, and ensures data quality and accuracy;
[0060] 4. The present invention realizes the visualization function of safety situation and emergency decision-making by designing a decision support module, provides decision makers with real-time decision support and visualization information display, optimizes traffic management and safety assurance, improves decision-making efficiency and quality, reduces accidents and traffic congestion, and improves the overall level of traffic operations. BRIEF DESCRIPTION OF THE DRAWINGS
[0061] Figure 1 It is a schematic diagram of the prediction process of the present invention;
[0062] Figure 2 It is a schematic diagram of a data acquisition module of the present invention;
[0063] Figure 3 It is a schematic diagram of a data processing and analysis module of the present invention;
[0064] Figure 4 It is a schematic diagram of the prediction modeling module of the present invention;
[0065] Figure 5 It is a schematic diagram of the security situation warning module of the present invention. DETAILED DESCRIPTION
[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0067] Example 1
[0068] See also Figure 1 and Figure 5, an embodiment provided by the present invention: a dynamic prediction system for traffic operation safety situation based on AI technology, comprising a prediction modeling module and a safety situation warning module, wherein the prediction modeling module provides a detection warning model for the safety situation warning module, and the safety situation warning module is used to predict the risk of traffic operation safety situation;
[0069] The safety situation warning module includes: an abnormal behavior detection warning unit, a dynamic traffic event warning unit and a comprehensive safety situation warning unit, and the results of the abnormal behavior detection warning unit and the dynamic traffic event warning unit are output as comprehensive analysis data of the comprehensive safety situation warning unit;
[0070] The abnormal behavior detection and warning unit detects abnormal behaviors in traffic scenes in real time through the CNN deep learning model, and immediately issues a warning notification when abnormal behaviors are found, supplemented by video evidence and trajectory data;
[0071] The dynamic traffic incident warning unit combines real-time data and historical event statistics, uses time series models and probability models to predict potential traffic accidents, and makes early warnings and dispatch arrangements;
[0072] The comprehensive safety situation warning unit comprehensively considers the abnormal behavior detection, traffic incident prediction and weather factor prediction results to form comprehensive safety situation warning information to support system decision-making and user emergency response;
[0073] Furthermore, the abnormal behavior detection and early warning unit receives processed traffic scene data from the data processing and analysis module, including vehicle form trajectory, speed, acceleration and other information, and uses the CNN deep learning model to extract features from the data to identify abnormal behaviors in the traffic scene, such as sudden braking, illegal lane change and wrong-way driving, etc. Based on the extracted feature data, the model determines whether there is abnormal behavior, such as sudden braking behavior or illegal lane change behavior, etc. Once abnormal behavior is detected, a warning notification is issued, including an alarm sound, a text prompt or a mobile application pushed to relevant personnel;
[0074] The dynamic traffic event warning unit receives processed traffic scene data and forecast information provided by the safety situation analysis unit, such as traffic flow, congestion, etc. It uses time series models and probability models, combined with real-time data and historical event statistics, to predict possible traffic events, such as accidents, traffic control, etc. Once a possible traffic event is predicted, a dynamic warning notification is immediately issued to remind relevant departments and drivers to take corresponding measures;
[0075] The comprehensive safety situation warning unit receives the warning information from the abnormal behavior detection warning unit and the dynamic traffic event warning unit, and combines the results of the safety situation analysis unit to form comprehensive safety situation data. It conducts a comprehensive analysis of the abnormal behavior, dynamic event warning and safety situation analysis results to form comprehensive safety situation warning information. Based on the comprehensive analysis results, it issues a comprehensive safety situation warning notification to convey comprehensive safety situation information to relevant departments and drivers so that they can take corresponding emergency measures and optimize traffic scheduling.
[0076] Example 2
[0077] See also Figure 1 , Figure 3 and Figure 4 , an embodiment provided by the present invention: a dynamic prediction system for traffic operation safety situation based on AI technology, wherein the prediction modeling module is used to establish multiple risk prediction models and continuously update model parameters through real-time data;
[0078] The prediction modeling module includes: a traffic flow prediction modeling unit, a traffic accident risk prediction unit and a weather factor impact prediction unit. The model data output of the traffic flow prediction modeling unit, the traffic accident risk prediction unit and the weather factor impact prediction unit is used as the warning model database of the safety situation warning module;
[0079] The traffic flow prediction modeling unit establishes a traffic flow prediction model based on the long short-term memory network model to predict traffic flow and congestion in the future and continuously updates the model parameters through real-time data;
[0080] The traffic accident risk prediction unit uses the support vector machine algorithm to establish a traffic accident risk prediction model to analyze the possibility and influencing factors of traffic incidents;
[0081] The weather factor impact prediction unit combines machine learning algorithms and meteorological data to model and predict the impact of different weather factors on traffic operation safety, and take countermeasures in advance;
[0082] Furthermore, the traffic flow prediction modeling unit receives processed traffic scene data, including historical traffic flow data, vehicle trajectory data, road condition information, etc., performs data preprocessing, feature extraction and feature selection on the input data, establishes a traffic flow prediction model based on a long short-term memory network model according to the specific requirements of the prediction task, trains the selected prediction model using historical data, optimizes model parameters, uses the trained model to predict traffic flow in the future, generates prediction results and outputs them;
[0083] The traffic accident risk prediction unit receives the processed traffic scene data and the data provided by the safety situation analysis unit, including historical accident data, vehicle driving trajectories and road conditions, performs feature engineering on the input data, extracts features related to accident risks, such as traffic flow, intersection density and intersection congestion, selects a support vector machine to use historical data to train the selected prediction model, optimizes model parameters, and uses the trained model to predict the location and time of traffic accidents that may occur in the future, generates prediction results and outputs them;
[0084] The weather factor impact prediction unit receives weather data and environmental monitoring data provided by the Meteorological Bureau, including temperature, humidity, wind speed and visibility, etc., performs feature engineering on the input data, extracts features related to weather factors, such as temperature changes, rainfall and haze degree, etc., selects models suitable for weather factor impact prediction, such as time series models, regression models, neural network models, etc., uses historical weather data to train the selected prediction model, optimizes model parameters, and uses the trained model to predict the impact of weather factors on traffic operations in the future, generates prediction results and outputs them.
[0085] Example 3
[0086] See also Figure 1 and Figure 2 , an embodiment provided by the present invention: a dynamic prediction system for traffic operation safety situation based on AI technology, wherein the prediction system is designed with a data acquisition module, and the data acquisition module is used to collect traffic status and meteorological status data in real time;
[0087] The data acquisition module includes: a sensor data acquisition unit, a traffic event acquisition unit and a meteorological environment data acquisition unit;
[0088] The sensor data acquisition unit uses advanced high-definition cameras, millimeter-wave radars, and lidars to collect real-time data from traffic scenes in real time;
[0089] The traffic event collection unit combines video surveillance and image recognition technology to monitor traffic events in real time and obtain real-time traffic event data from the traffic management center;
[0090] The meteorological environment data collection unit integrates the weather data and environmental monitoring data provided by the Meteorological Bureau to analyze the impact of weather on traffic operations. The meteorological environment data collection unit transmits the collected data to the weather factor impact prediction unit through the Internet;
[0091] Furthermore, various types of sensor devices are arranged in the traffic scene, such as vehicle-mounted sensors, high-definition cameras for traffic monitoring, millimeter-wave radars and laser radars, etc. The sensor devices collect various data in the traffic scene in real time, including information such as vehicle speed, density, trajectory, acceleration and vehicle type, and the collected raw data is output to the data processing and analysis module through the Internet;
[0092] The traffic event collection unit combines video surveillance and image recognition technology to monitor various events occurring in traffic scenes in real time, such as accidents, congestion, construction, etc., identify and classify the monitored events, distinguish different types of traffic events, and record relevant data, such as the time of occurrence, location, and scope of impact, etc. The identified event data is transmitted to the data processing and analysis module for use by the safety situation analysis module and the predictive modeling module;
[0093] Set up meteorological battle points or install meteorological sensor equipment to monitor meteorological data in the environment, such as temperature, humidity, wind speed and visibility. The meteorological environment data acquisition unit collects meteorological data in real time and records meteorological changes, including daily changes and sudden changes. The collected meteorological data is transmitted to the data processing and analysis module through the communication network for the establishment of prediction models and the analysis of the impact of weather factors on traffic.
[0094] Example 4
[0095] See also Figure 1 , Figure 2 and Figure 3 , an embodiment provided by the present invention: a dynamic prediction system for traffic operation safety situation based on AI technology, wherein the data acquisition module is connected to a data processing and analysis module provided in the system through the Internet, the data output of the data processing and analysis module is used as a modeling parameter input of the prediction modeling module, and the data processing and analysis module is used to process and analyze the comprehensive data collected by the data acquisition module;
[0096] The data processing and analysis module includes: a data cleaning preprocessing unit, a traffic situation perception unit and a safety situation analysis unit. The data output of the data cleaning preprocessing unit serves as the analysis input of the traffic situation perception unit and the safety situation analysis unit.
[0097] The data cleaning preprocessing unit receives the raw data from the sensor data collection unit, the traffic event collection unit and the meteorological environment data collection unit, processes the raw data collected by the sensor data collection unit through data cleaning and denoising technology, and formats and standardizes the data;
[0098] The traffic situation awareness unit receives the traffic data processed by the data cleaning preprocessing unit, uses the deep learning algorithm to perform real-time analysis on the processed data, identifies traffic flow, vehicle driving trajectory, congestion and abnormal vehicle behavior, and forms awareness of the traffic situation;
[0099] The safety situation analysis unit receives the processed traffic data and the information provided by the traffic situation perception unit, collects historical data and real-time situation, uses machine learning models to analyze the traffic safety status, and identifies potential safety hazards and risks. The safety situation analysis unit transmits the analysis data to the traffic flow prediction modeling unit and the traffic accident risk prediction unit through the Internet;
[0100] Furthermore, the data cleaning preprocessing unit cleans and denoises the raw data collected from the data acquisition module, excludes abnormal data and erroneous data, and preprocesses the cleaned data, including data format conversion, missing value filling, data standardization, etc., to integrate and consolidate data from different data sources and unify the format and data structure;
[0101] The traffic situation awareness unit extracts traffic situation-related features from the preprocessed data, such as traffic flow, congestion level, vehicle speed, traffic density, and intersection traffic conditions. It uses machine learning and data mining techniques to analyze and process the extracted features, perceive the current situation in the traffic scene, such as whether the traffic is smooth, whether there is congestion, and whether there is abnormal behavior, etc. It continuously monitors and updates traffic situation information, responds to changes and development trends in traffic scenes in real time, and provides real-time data support for traffic management and dispatching.
[0102] The safety situation analysis unit extracts features related to the safety situation from the pre-processed data, including traffic event data, vehicle driving trajectories, road congestion and weather factors, etc. It uses machine learning and data analysis techniques to perform safety situation analysis on the extracted features, identify potential safety hazards and risk factors, such as the possibility of traffic accidents, traffic violations and traffic event predictions, and issue safety situation warning notifications based on the safety situation analysis results to remind relevant departments and drivers to pay attention to safety and take corresponding preventive measures and emergency response.
[0103] Example 5
[0104] See also Figure 1-Figure 5 , an embodiment provided by the present invention: a dynamic prediction system for traffic operation safety situation based on AI technology, wherein the safety situation warning module is connected to a decision support module via the Internet, the decision support module receives the warning information output of the safety situation warning module as decision data input, and the decision support module is used to optimize traffic operations and visualize decision information display;
[0105] The decision support module includes: a safety emergency decision support unit, a traffic dispatch optimization unit and a safety situation visualization display unit. The data output of the safety emergency decision support unit and the traffic dispatch optimization unit is used as the display database source of the safety situation visualization display unit.
[0106] The safety emergency decision support unit provides safety emergency decision support based on early warning information to optimize traffic operation safety;
[0107] The traffic dispatch optimization unit combines the prediction model provided by the prediction modeling module and the real-time data provided by the data processing and analysis module to optimize traffic dispatch;
[0108] The safety situation visualization display unit visualizes the prediction results of the safety situation warning module and the safety situation information of the safety emergency decision support unit and the traffic dispatch optimization unit, providing decision makers with intuitive traffic safety situation information;
[0109] Furthermore, the safety emergency decision support unit receives real-time safety warning information from the safety situation analysis unit, including traffic accident risk warning and abnormal behavior detection warning, analyzes and evaluates the received safety warning information, determines the urgency and impact scope of the incident, and provides a basis for the emergency response. According to the event analysis results and warning information, it provides emergency decision support to relevant decision makers, including intersection traffic signal control adjustment, road section traffic control, accident rescue dispatch, etc.;
[0110] The traffic dispatch optimization unit receives the prediction results of future traffic flow, congestion, etc. from the traffic flow prediction modeling unit as the basic data for dispatch optimization. It uses optimization algorithms such as genetic algorithms, simulated annealing algorithms, and rule-based algorithms to optimize traffic dispatch, including traffic signal control optimization, road section flow adjustment, traffic management strategy formulation, etc., to generate optimized traffic dispatch plans to improve traffic operation efficiency and reduce congestion and accident probability;
[0111] The safety situation visualization display unit obtains real-time data and analysis results from the data processing and analysis module and other modules, integrates and formats the data, and designs a visualization display interface for the safety situation, including traffic status maps, safety warning information, and real-time monitoring charts of traffic incidents. The integrated data is displayed to decision makers and operators in real time through the visualization display interface, providing intuitive traffic safety situation information and prediction results, and supporting users to interact with the display interface, such as zooming in and out of the map, viewing detailed information, and adjusting parameters, to meet users' personalized needs.
[0112] Working principle: First, various data collection modules are used to collect data in traffic scenarios in real time, including vehicle speed, density, meteorological factors, and traffic events. The collected raw data is cleaned, denoised, and formatted. Then, feature extraction and data integration are performed to generate a data set suitable for modeling.
[0113] Then the prediction modeling module uses deep learning, machine learning and statistical analysis technologies to construct various prediction models, including traffic flow prediction models, traffic accident risk prediction models and weather factor impact prediction models. By learning and training historical data, the model parameters are optimized to provide prediction accuracy and generalization capabilities. At the same time, the safety situation warning module analyzes and evaluates the prediction results to identify key information and trends in traffic and safety situations.
[0114] Finally, the system transmits the results of the prediction model and the analysis results to the decision support module, including safety emergency decision support, traffic scheduling optimization and security situation visualization. Based on real-time data and prediction results, the system provides safety warning and emergency decision support, optimizes traffic scheduling plans, and realizes safe, efficient and intelligent management of traffic operations.
[0115] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.
Claims
1. Dynamic prediction system of traffic operation safety situation based on AI technology, Features: It includes a prediction modeling module and a safety situation warning module. The prediction modeling module provides a detection warning model for the safety situation warning module. The safety situation warning module is used to predict the risk of traffic operation safety situation. The safety situation warning module includes: an abnormal behavior detection warning unit, a dynamic traffic event warning unit and a comprehensive safety situation warning unit, and the results of the abnormal behavior detection warning unit and the dynamic traffic event warning unit are output as comprehensive analysis data of the comprehensive safety situation warning unit; The abnormal behavior detection and warning unit detects abnormal behaviors in traffic scenes in real time through the CNN deep learning model, and immediately issues a warning notification when abnormal behaviors are found, supplemented by video evidence and trajectory data; The dynamic traffic incident warning unit combines real-time data and historical event statistics, uses time series models and probability models to predict potential traffic accidents, and makes early warnings and dispatch arrangements; The comprehensive safety situation warning unit comprehensively considers abnormal behavior detection, dynamic traffic event warning and weather factor prediction results to form comprehensive safety situation warning information to support system decision-making and user emergency response; The prediction modeling module is used to establish multiple risk prediction models and continuously update model parameters through real-time data; The prediction modeling module includes: a traffic flow prediction modeling unit, a traffic accident risk prediction unit and a weather factor impact prediction unit. The model data output of the traffic flow prediction modeling unit, the traffic accident risk prediction unit and the weather factor impact prediction unit is used as the warning model database of the safety situation warning module; The traffic flow prediction modeling unit establishes a traffic flow prediction model based on the long short-term memory network model to predict traffic flow and congestion in the future and continuously updates the model parameters through real-time data; The traffic accident risk prediction unit uses the support vector machine algorithm to establish a traffic accident risk prediction model to analyze the possibility and influencing factors of traffic incidents; The weather factor impact prediction unit combines machine learning algorithms and meteorological data to model and predict the impact of different weather factors on traffic operation safety, and take countermeasures in advance.
2. According to claim 1, the dynamic prediction system for traffic operation safety situation based on AI technology, Features: The prediction system is designed with a data acquisition module, which is used to collect traffic status and meteorological status data in real time; The data acquisition module includes: a sensor data acquisition unit, a traffic event acquisition unit and a meteorological environment data acquisition unit; The sensor data acquisition unit uses advanced high-definition cameras, millimeter-wave radars, and lidars to collect real-time data from traffic scenes in real time; The traffic event collection unit combines video surveillance and image recognition technology to monitor traffic events in real time and obtain real-time traffic event data from the traffic management center; The meteorological and environmental data collection unit integrates the weather data and environmental monitoring data provided by the Meteorological Bureau to analyze the impact of weather on traffic operations. The meteorological and environmental data collection unit transmits the collected data to the weather factor impact prediction unit via the Internet.
3. The traffic operation safety situation dynamic prediction system based on AI technology according to claim 2, Features: The data acquisition module is connected to a data processing and analysis module set in the system through the Internet. The data output of the data processing and analysis module is used as the modeling parameter input of the prediction modeling module. The data processing and analysis module is used to process and analyze the comprehensive data collected by the data acquisition module; The data processing and analysis module includes: a data cleaning preprocessing unit, a traffic situation perception unit and a safety situation analysis unit. The data output of the data cleaning preprocessing unit serves as the analysis input of the traffic situation perception unit and the safety situation analysis unit. The data cleaning preprocessing unit receives the raw data from the sensor data collection unit, the traffic event collection unit and the meteorological environment data collection unit, processes the raw data collected by the sensor data collection unit through data cleaning and denoising technology, and formats and standardizes the data; The traffic situation awareness unit receives the traffic data processed by the data cleaning preprocessing unit, uses the deep learning algorithm to perform real-time analysis on the processed data, identifies traffic flow, vehicle driving trajectory, congestion and abnormal vehicle behavior, and forms awareness of the traffic situation; The safety situation analysis unit receives the processed traffic data and the information provided by the traffic situation awareness unit, collects historical data and real-time situation, uses machine learning models to analyze the traffic safety status, and identifies potential safety hazards and risks. The safety situation analysis unit transmits the analysis data to the traffic flow prediction modeling unit and the traffic accident risk prediction unit via the Internet.
4. The traffic operation safety situation dynamic prediction system based on AI technology according to claim 1, Features: The safety situation warning module is connected to a decision support module via the Internet. The decision support module receives the warning information output of the safety situation warning module as decision data input. The decision support module is used to optimize traffic operations and visualize decision information display; The decision support module includes: a safety emergency decision support unit, a traffic dispatch optimization unit and a safety situation visualization display unit. The data output of the safety emergency decision support unit and the traffic dispatch optimization unit is used as the display database source of the safety situation visualization display unit. The safety emergency decision support unit provides safety emergency decision support based on early warning information to optimize traffic operation safety; The traffic dispatch optimization unit combines the prediction model provided by the prediction modeling module and the real-time data provided by the data processing and analysis module to optimize traffic dispatch; The security situation visualization display unit displays the prediction results of the security situation warning module and the security situation information of the security emergency decision support unit and the traffic dispatch optimization unit in a visual manner, providing decision makers with intuitive traffic safety situation information.
5. The traffic operation safety situation dynamic prediction system based on AI technology according to claim 1, Features: The CNN deep learning model used by the abnormal behavior detection and early warning unit is expressed as: ， Among them, y is the output, f is the activation function, and w i is the weight, x i is the input, b is the bias term; Specific model architecture: Input layer: image and video data in traffic scenes; Convolution layer and pooling layer: used for feature extraction and dimensionality reduction; Activation function layer: introduces nonlinear factors; Fully connected layer: maps the extracted features to the final output layer; Output layer: warning results of abnormal behavior.
6. The traffic operation safety situation dynamic prediction system based on AI technology according to claim 1, Features: The long short-term memory network model used by the traffic flow prediction modeling unit is expressed as: the forget gate output , input gate output , the output gate outputs , the output of the new candidate unit , cell status , output ; Among them, f t Represents the output of the forget gate, with a value between 0 and 1, indicating that C t-1 The degree of forgetting; σ represents the activation function, which maps the output to the range [0,1]; W f represents the weight matrix of the forget gate; h t-1 represents the hidden state of the previous time step; x t represents the input of the current time step; b f Represents the bias vector of the forget gate; i t Represents the output of the input gate, indicating the importance of the input information, with a value between 0 and 1; W i represents the weight matrix of the input gate; b i represents the bias vector of the input gate; Represents the new candidate unit state of the current time step, and its value range is [-1, 1] after the tanh activation function; W C represents the weight matrix of the new candidate unit; b C represents the bias vector of the new candidate unit; C t represents the cell state at the current time step; f t *C t-1 Indicates that the cell state information of the previous time step is retained; i t * Indicates the status information of adding new candidate units; o t Represents the output of the output gate, which determines what information can be output; h t represents the hidden state of the current time step, that is, the output of LSTM; W o represents the weight matrix of the output gate; b o Represents the bias vector of the output gate; tanh(C t ) means applying the tanh function to the cell state and mapping it to the range of [-1, 1]; Specific model architecture: Input layer: time series traffic flow data; LSTM layer: used to process sequence data and remember long-term dependencies; Output layer: predict traffic flow and congestion in the future.
7. The traffic operation safety situation dynamic prediction system based on AI technology according to claim 1, Features: The support vector machine algorithm formula used by the traffic accident risk prediction unit is as follows: ， Among them, w is the normal vector of the hyperplane, B is the bias term, C is the penalty parameter, and X i is the sample data, y i is the sample label, is a slack variable, N is the number of samples, and i is the sample number; Specific model architecture: Input layer: feature vector of traffic data; Output layer: risk prediction probability.
8. A method for dynamically predicting the safety situation of traffic operations based on AI technology, applicable to a system for dynamically predicting the safety situation of traffic operations based on AI technology as described in any one of claims 1 to 7, Features: The method comprises the following steps: S1. First, use the data acquisition module to obtain the vehicle speed, density, trajectory and weather real-time data in the traffic scene; S2, then extracting characteristic data of traffic flow, congestion level and vehicle behavior pattern from the cleaned data through the data processing and analysis module; S3, then use the modeling and prediction module to model the traffic flow, predict the traffic flow and congestion in the future, and provide a reference for the decision support module; S4. Finally, the safety situation warning module detects abnormal behaviors in traffic scenes in real time, issues warning notifications and provides evidence.
9. The method for dynamic prediction of traffic operation safety situation based on AI technology according to claim 8, Features: The method further comprises the following steps: S21. Use the AI algorithm in the security situation analysis unit to perform security situation analysis and identify potential security risks and risk factors; S31, establishing a traffic accident risk prediction model through a support vector machine algorithm in a traffic accident risk prediction unit to predict the location and time of a possible accident so as to take preventive measures and emergency treatment; S41. Utilize the comprehensive safety situation warning unit to combine real-time data and prediction results to dynamically warn of the traffic operation safety situation and provide it to the decision support module so as to take real-time traffic control and scheduling optimization.