Traffic data processing method and system based on artificial intelligence

Abstract

The invention relates to the technical field of artificial intelligence data processing, provides a traffic data processing method and system based on artificial intelligence, and solves the problems of data missing and abnormity by processing diversified traffic data through big data analysis and a deep learning algorithm. The system combines real-time data and simulation data for prediction, automatically evaluates prediction errors, optimizes a prediction model, and improves prediction precision. The traffic prediction capability is enhanced, the development of an intelligent traffic system is supported, the operation stability and management efficiency of the traffic system are improved, powerful support is provided for urban traffic planning and management, and the problem that the traffic data processing cannot be carried out if data processing errors or data processing timeout occur in the traffic data processing process is solved. And the operation stability of the traffic system is influenced.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence data processing, and provides a traffic data processing method and system based on artificial intelligence. Background Art

[0002] With the rapid development of science and technology, artificial intelligence (AI) is increasingly used in the field of traffic data processing. Currently, AI technology is being used to analyze massive amounts of traffic data, including traffic flow, vehicle speed, road conditions, etc., to optimize traffic management and improve road safety. By processing this data in real time, AI can intelligently adjust the timing of traffic lights and optimize the distribution of traffic flow, thereby effectively reducing traffic congestion and improving traffic efficiency. At the same time, AI also plays a key role in autonomous driving technology. Through deep learning algorithms and sensor technology, it enables autonomous perception and decision-making of vehicles, laying the foundation for future intelligent transportation systems.

[0003] However, artificial intelligence faces the following technical pain points in the field of traffic data processing: the sources of traffic data are diverse and the types are complicated, including structured data and unstructured data, such as sensor data, video image data, etc. The real-time and accuracy requirements of traffic data are extremely high, and AI algorithms are required to process large amounts of data in a short period of time and make accurate decisions. In addition, with the continuous expansion of urban traffic, the surge in data volume has also brought difficulties to the performance and efficiency of AI algorithms. In the process of traffic data processing, if data processing errors or data processing timeouts occur, it will affect the operational stability of the traffic system. For this reason, the present invention provides a traffic data processing method and system based on artificial intelligence. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a traffic data processing method and system based on artificial intelligence to solve the problem that if data processing errors or data processing timeouts occur during the traffic data processing process, the operating stability of the traffic system will be affected.

[0005] In order to solve the above technical problems, the specific technical solutions of the present invention are as follows:

[0006] In a first aspect, the present invention provides a traffic data processing method based on artificial intelligence, comprising:

[0007] Step S101, acquiring traffic data, the traffic data including traffic flow, vehicle speed, traffic video images and traffic sensor data;

[0008] Step S102, preprocessing the acquired traffic data to obtain traffic data to be processed, copying the preprocessed traffic data to obtain a traffic data group to be processed, the traffic data group to be processed including first traffic data to be processed, second traffic data to be processed and third traffic data to be processed;

[0009] Step S103, using a big data analysis method, deeply analyzing the first traffic data to be processed, extracting traffic congestion mode data features and vehicle driving trajectory data features through a deep learning algorithm, performing data integrity detection on the traffic congestion mode data features and vehicle driving trajectory data features, and if there is data missing in the traffic congestion mode data features and vehicle driving trajectory data features, determining the missing data time information, and according to the missing data time information, substituting the second traffic data to be processed and the third traffic data to be processed into a preset traffic data simulation model to obtain traffic simulation data;

[0010] Step S104, fusing the traffic simulation data with the first traffic data to be processed to obtain corrected processed traffic data, substituting the corrected processed traffic data into a preset traffic data prediction model to obtain a real-time traffic operation data prediction result;

[0011] Step S105, collect the data time characteristics in the real-time traffic operation data prediction results, retrieve the corresponding traffic data according to the data time characteristics, compare the real-time traffic operation data prediction results with the corresponding traffic data retrieved according to the data time characteristics, if the error value of the comparison result is greater than the preset allowable range, optimize the traffic data prediction model, use the optimized traffic data prediction model to process the traffic data, and obtain the optimized traffic data processing results.

[0012] Furthermore, the traffic data processing method based on artificial intelligence provided by the present invention, step S103, comprises:

[0013] The data analysis module aggregates and summarizes the data to generate a statistical overview and trend analysis of the data, and uses association rule mining technology to extract the associations between traffic data items. The associations between traffic data items include the association between traffic congestion in a certain time period and location and the weather, and the association between traffic congestion in a certain time period and location and special events;

[0014] Use deep learning algorithms to train and learn traffic data to automatically extract data features of traffic congestion patterns and vehicle driving trajectories. Deep learning models identify periodic patterns, abnormal patterns in traffic flow, and regular features of vehicle driving trajectories.

[0015] The extracted traffic congestion pattern data features and vehicle driving trajectory data features were subjected to integrity verification to check whether the features were complete and whether there were missing values ​​or outliers;

[0016] If the data are missing, the missing values ​​are handled using interpolation, filling or deletion. If the data are outliers, the outliers are corrected.

[0017] Furthermore, the traffic data processing method based on artificial intelligence provided by the present invention, step S104, comprises:

[0018] Obtaining generated traffic simulation data, which includes traffic flow predictions and congestion simulation data for different time periods;

[0019] The real-time processing module collects the first traffic data to be processed from the traffic monitoring system, vehicle GPS tracking, and mobile phone signaling;

[0020] The real-time captured traffic data and the simulated data are fused, including time alignment, spatial matching and data normalization. The real-time captured traffic data and the simulated data are fused through the attention mechanism in deep learning to obtain the fused data.

[0021] Loading a preset traffic data prediction model, inputting the corrected processed traffic data into the preset traffic data prediction model, the preset traffic data prediction model predicts real-time traffic operation data based on the input data, combined with the model's internal parameters and the knowledge learned during training;

[0022] The preset traffic data prediction model outputs the prediction results of real-time traffic operation data, and the prediction results of real-time traffic operation data include traffic flow, congestion and average vehicle speed.

[0023] Furthermore, the traffic data processing method based on artificial intelligence provided by the present invention, step S105, comprises:

[0024] Collecting real-time traffic operation data prediction results, including traffic flow, congestion index and average vehicle speed;

[0025] Retrieving corresponding actual traffic data from a traffic database according to the time characteristics of the real-time traffic operation data prediction results, where the time characteristics include date characteristics and time period characteristics;

[0026] Compare the predicted results of real-time traffic operation data with the actual traffic data retrieved, and the comparison indicators include traffic flow, average vehicle speed and congestion duration;

[0027] Calculate the error value of the comparison result, and determine whether the error value of the comparison result is greater than a preset allowable range;

[0028] If the error value is greater than the preset allowable range, the prediction accuracy of the current traffic data prediction model does not meet the requirements.

[0029] According to the error analysis results, the sources of deviation in the traffic data prediction model are identified, the traffic data prediction model parameters are adjusted, the traffic data prediction model structure is updated, the optimized traffic data prediction model is obtained, and the optimized traffic data prediction model is retrained on the new data set.

[0030] In a second aspect, the present invention provides a traffic data processing system based on artificial intelligence, which is applied to the traffic data processing method based on artificial intelligence, including: a server end, a traffic equipment end, a data collection end and a background control end, the server end establishes a communication connection with the traffic equipment end, the data collection end and the background control end, and the server end includes:

[0031] A data acquisition module, which acquires traffic data, including traffic flow, vehicle speed, traffic video images, and traffic sensor data;

[0032] A data preprocessing module preprocesses the acquired traffic data to obtain traffic data to be processed, and copies the preprocessed traffic data to obtain a traffic data group to be processed, wherein the traffic data group to be processed includes first traffic data to be processed, second traffic data to be processed, and third traffic data to be processed;

[0033] The data analysis module uses a big data analysis method to conduct an in-depth analysis of the first traffic data to be processed, extracts traffic congestion mode data features and vehicle driving trajectory data features through a deep learning algorithm, performs data integrity detection on the traffic congestion mode data features and vehicle driving trajectory data features, and if there is data missing in the traffic congestion mode data features and vehicle driving trajectory data features, determines the missing data time information, and substitutes the second traffic data to be processed and the third traffic data to be processed into a preset traffic data simulation model according to the missing data time information to obtain traffic simulation data;

[0034] Real-time processing module: fusing the traffic simulation data with the first traffic data to be processed to obtain corrected processed traffic data, substituting the corrected processed traffic data into a preset traffic data prediction model to obtain a real-time traffic operation data prediction result;

[0035] The traffic data processing optimization unit collects the data time characteristics in the real-time traffic operation data prediction results, retrieves the corresponding traffic data according to the data time characteristics, and compares the real-time traffic operation data prediction results with the corresponding traffic data retrieved according to the data time characteristics. If the error value of the comparison result is greater than a preset allowable range, the traffic data prediction model is optimized, and the traffic data is processed using the optimized traffic data prediction model to obtain an optimized traffic data processing result.

[0036] Furthermore, in the traffic data processing system based on artificial intelligence provided by the present invention, the data analysis module is also used for:

[0037] The data analysis module aggregates and summarizes the data to generate a statistical overview and trend analysis of the data, and uses association rule mining technology to extract the associations between traffic data items. The associations between traffic data items include the association between traffic congestion in a certain time period and location and the weather, and the association between traffic congestion in a certain time period and location and special events;

[0038] Use deep learning algorithms to train and learn traffic data to automatically extract data features of traffic congestion patterns and vehicle driving trajectories. Deep learning models identify periodic patterns, abnormal patterns in traffic flow, and regular features of vehicle driving trajectories.

[0039] The extracted traffic congestion pattern data features and vehicle driving trajectory data features were subjected to integrity verification to check whether the features were complete and whether there were missing values ​​or outliers;

[0040] If the data are missing, the missing values ​​are handled using interpolation, filling or deletion. If the data are outliers, the outliers are corrected.

[0041] Furthermore, in the traffic data processing system based on artificial intelligence provided by the present invention, the real-time processing module is also used for:

[0042] Obtaining generated traffic simulation data, which includes traffic flow predictions and congestion simulation data for different time periods;

[0043] The real-time processing module collects the first traffic data to be processed from the traffic monitoring system, vehicle GPS tracking, and mobile phone signaling;

[0044] The real-time captured traffic data and the simulated data are fused, including time alignment, spatial matching and data normalization. The real-time captured traffic data and the simulated data are fused through the attention mechanism in deep learning to obtain the fused data.

[0045] Loading a preset traffic data prediction model, inputting the corrected processed traffic data into the preset traffic data prediction model, the preset traffic data prediction model predicts real-time traffic operation data based on the input data, combined with the model's internal parameters and the knowledge learned during training;

[0046] The preset traffic data prediction model outputs the prediction results of real-time traffic operation data, and the prediction results of real-time traffic operation data include traffic flow, congestion and average vehicle speed.

[0047] Furthermore, the traffic data processing system based on artificial intelligence provided by the present invention, the traffic data processing optimization unit is also used for:

[0048] Collecting real-time traffic operation data prediction results, including traffic flow, congestion index and average vehicle speed;

[0049] Retrieving corresponding actual traffic data from a traffic database according to the time characteristics of the real-time traffic operation data prediction results, where the time characteristics include date characteristics and time period characteristics;

[0050] Compare the predicted results of real-time traffic operation data with the actual traffic data retrieved, and the comparison indicators include traffic flow, average vehicle speed and congestion duration;

[0051] Calculate the error value of the comparison result, and determine whether the error value of the comparison result is greater than a preset allowable range;

[0052] If the error value is greater than the preset allowable range, the prediction accuracy of the current traffic data prediction model does not meet the requirements.

[0053] According to the error analysis results, the sources of deviation in the traffic data prediction model are identified, the traffic data prediction model parameters are adjusted, the traffic data prediction model structure is updated, the optimized traffic data prediction model is obtained, and the optimized traffic data prediction model is retrained on the new data set.

[0054] The beneficial effects of the present invention are mainly reflected in the following aspects:

[0055] By combining big data analysis and deep learning algorithms, the present invention can efficiently and accurately process a variety of traffic data, including traffic flow, vehicle speed, traffic video images and sensor data, effectively solving the problems of missing data and outliers that may be encountered in the process of traffic data processing.

[0056] By using a preset traffic data prediction model, the present invention can predict real-time traffic operation data based on real-time traffic data and simulation data, including key indicators such as traffic flow, congestion and average vehicle speed, thereby improving the accuracy and real-time performance of traffic prediction.

[0057] The present invention automatically evaluates the prediction error by comparing the real-time prediction results with the actual traffic data, and optimizes the traffic data prediction model according to the error analysis results, thereby ensuring the accuracy and adaptability of the model prediction and further improving the prediction effect. By timely discovering and handling data processing errors and continuously optimizing the prediction model, the present invention helps to improve the operational stability and reliability of the entire transportation system and reduce traffic congestion and improper management caused by data processing problems.

[0058] The implementation of the present invention provides technical support for building a more intelligent and efficient transportation system, helps promote the development of intelligent traffic signal control, autonomous driving technology, etc., and improves the overall efficiency of urban transportation and user experience.

[0059] In summary, the present invention significantly improves the efficiency and accuracy of traffic data processing through innovative data processing methods and system architecture, enhances traffic prediction capabilities, and lays a solid foundation for the development of intelligent transportation systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0061] Figure 1 A schematic diagram of a traffic data processing method based on artificial intelligence provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described in conjunction with the specific embodiments of the present invention and the corresponding drawings. 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. The technical solutions provided by the embodiments of the present invention are described in detail below in conjunction with the drawings.

[0063] In order to better understand the purpose of the present invention, the present invention is described in further detail below.

[0064] First, see Figure 1 The present invention provides a traffic data processing method based on artificial intelligence, comprising:

[0065] Step S101, acquiring traffic data, the traffic data including traffic flow, vehicle speed, traffic video images and traffic sensor data;

[0066] In step S101, the present invention obtains diversified traffic data, which is the basis for subsequent analysis, processing and prediction. The specific data types obtained include:

[0067] Traffic flow: Traffic flow is an important indicator to measure the degree of road traffic congestion, usually collected through road monitoring equipment, geomagnetic sensors and other equipment. Traffic flow data can reflect the number of vehicles passing through a certain road section or intersection in real time within a specific time period.

[0068] Vehicle speed: Vehicle speed data can reflect the smoothness of road traffic. It can be obtained through vehicle GPS tracking systems, speed radars, cameras and other technical means. The change in vehicle speed is of great significance for assessing road congestion and formulating traffic management strategies.

[0069] Traffic video images: Video image data provides intuitive traffic scene information, including vehicle trajectories, pedestrian dynamics, traffic accidents, etc. Through video image analysis, traffic violations can be identified, traffic flow and density can be monitored, and even traffic congestion trends can be predicted. Modern traffic monitoring systems widely use high-definition cameras and intelligent video analysis technology to improve the efficiency and accuracy of traffic management.

[0070] Traffic sensor data: Traffic sensors include geomagnetic sensors, microwave detectors, infrared sensors and other types. They can detect the presence, speed, model and other information of vehicles on the road in real time. By integrating different types of sensor data, a more comprehensive traffic status perception network can be built.

[0071] When acquiring these data, the method of the present invention relies on multi-source data acquisition equipment and technical means, such as traffic monitoring cameras, vehicle GPS tracking systems, geomagnetic sensors, microwave detectors, etc., which are transmitted to a data processing center or cloud platform via wired or wireless means for subsequent big data analysis, deep learning and model prediction.

[0072] Acquiring high-quality traffic data is the key to ensuring the accuracy of subsequent analysis and prediction. Therefore, in step S101, the present invention may also involve data quality verification and preprocessing to improve the integrity, accuracy and consistency of the data.

[0073] Step S102, preprocessing the acquired traffic data to obtain traffic data to be processed, copying the preprocessed traffic data to obtain a traffic data group to be processed, the traffic data group to be processed including first traffic data to be processed, second traffic data to be processed and third traffic data to be processed;

[0074] First, the raw traffic data obtained from various data sources (such as traffic monitoring systems, vehicle GPS tracking, traffic sensors, etc.) are preprocessed. The main purpose of preprocessing is to clean the data, correct errors, fill missing values, and convert the data into a format suitable for subsequent analysis.

[0075] Data cleaning: remove duplicate records, correct format errors (such as inconsistent date formats, incorrect data types, etc.), and handle outliers (such as unreasonable vehicle speeds, traffic peaks, etc.).

[0076] Data conversion: Convert text data into numerical data (such as converting dates and times into timestamps), and standardize or normalize the data for use in subsequent analysis.

[0077] Data integration: Integrate data from different data sources according to time and space dimensions to form a unified traffic data set.

[0078] After the preprocessing is completed, the obtained traffic data to be processed is copied to form multiple independent data copies. The copies are organized into traffic data groups to be processed, usually including the first traffic data to be processed, the second traffic data to be processed, and the third traffic data to be processed. The purpose of this step is to:

[0079] Improve data redundancy: By duplicating data, data redundancy is increased and the risk of a single data copy being damaged or lost is reduced.

[0080] Support parallel processing: Multiple data copies can support parallel processing and analysis, improving data processing speed and efficiency.

[0081] Providing materials for data simulation: In the subsequent steps, if the first traffic data to be processed is missing or abnormal, the second and third traffic data to be processed can be used for data simulation to supplement or correct the missing data.

[0082] Through data preprocessing and replication grouping strategy, the present invention provides high-quality and highly redundant traffic data sets for subsequent data analysis and model prediction. This strategy not only improves the availability and accuracy of data, but also provides an effective solution for dealing with problems such as missing and abnormal data. At the same time, the grouping strategy also facilitates possible parallel processing and data analysis, which helps to improve the efficiency of overall data processing and analysis.

[0083] Step S103, using a big data analysis method, deeply analyzing the first traffic data to be processed, extracting traffic congestion mode data features and vehicle driving trajectory data features through a deep learning algorithm, performing data integrity detection on the traffic congestion mode data features and vehicle driving trajectory data features, and if there is data missing in the traffic congestion mode data features and vehicle driving trajectory data features, determining the missing data time information, and according to the missing data time information, substituting the second traffic data to be processed and the third traffic data to be processed into a preset traffic data simulation model to obtain traffic simulation data;

[0084] In step S103, the present invention conducts in-depth analysis and processing on the first traffic data to be processed, aiming to extract key traffic congestion patterns and vehicle driving trajectory features, and ensure the integrity and availability of the data through a series of measures. The specific steps and logic are as follows:

[0085] In-depth analysis of the first unprocessed traffic data, using big data analysis methods to conduct in-depth mining of the first unprocessed traffic data. Big data analysis methods include statistical analysis, association rule mining, cluster analysis, etc. These methods can help researchers identify patterns, associations and anomalies in the data.

[0086] Through deep learning algorithms (such as convolutional neural networks (CNN), recurrent neural networks (RNN), long short-term memory networks (LSTM), etc.), traffic data is trained and learned to automatically extract data features of traffic congestion patterns and vehicle driving trajectories. These features can reflect information such as periodic changes in traffic flow, abnormal congestion, and regularity of vehicle driving, which are of great significance for subsequent traffic prediction and management decisions.

[0087] After extracting the features, the traffic congestion pattern data features and vehicle driving trajectory data features are checked for data integrity. The purpose of this step is to ensure that the extracted feature data is complete and can fully reflect the traffic conditions. If missing or abnormal data is detected (such as no congestion data in certain time periods, discontinuous driving trajectories, etc.), further processing is required.

[0088] When it is found that the traffic congestion mode data features or vehicle driving trajectory data features are missing, the present invention adopts an innovative solution to use the second traffic data to be processed and the third traffic data to be processed for data simulation. The specific steps are as follows:

[0089] To determine the time information of missing data, first, determine the specific time period and location of the missing data based on the results of the data integrity check.

[0090] Select a simulation data source, and select data with a time period close to the missing data and similar traffic conditions from the second to-be-processed traffic data and the third to-be-processed traffic data as the simulation data source.

[0091] Substitute the selected simulated data source into the preset traffic data simulation model. This model may be built based on statistical methods, machine learning algorithms or deep learning models, and can generate simulated traffic conditions based on existing data.

[0092] Generate traffic simulation data, and generate traffic simulation data corresponding to the missing data time period through the simulation model. These data are consistent with the real data in statistical characteristics and trends, and can effectively fill the missing parts of the original data.

[0093] Filling missing values ​​through data simulation methods can ensure the integrity and continuity of traffic data and provide a reliable data foundation for subsequent data analysis and model prediction.

[0094] When faced with errors or timeouts that may occur during traffic data processing, the method of the present invention can effectively deal with data loss problems and avoid system performance degradation or increased prediction errors caused by incomplete data. A complete traffic data set helps to improve the accuracy of traffic operation data prediction. By deeply analyzing the features and patterns in the complete data set, a more accurate prediction model can be established, providing a more scientific basis for traffic management and decision-making.

[0095] In summary, step S103 effectively solves the integrity problems that may arise in the process of traffic data processing through a series of measures such as extracting traffic characteristics through big data analysis and deep learning algorithms, performing data integrity detection, and filling missing values ​​with simulated data, providing strong support for the stable operation and efficient prediction of the traffic system.

[0096] Step S104, fusing the traffic simulation data with the first traffic data to be processed to obtain corrected processed traffic data, substituting the corrected processed traffic data into a preset traffic data prediction model to obtain a real-time traffic operation data prediction result;

[0097] In step S104, the present invention fuses the traffic simulation data with the first traffic data to be processed to obtain a more complete and accurate traffic data set, and then uses this data set to predict real-time traffic operation data. The specific steps and logic are as follows:

[0098] Traffic simulation data corresponding to the missing data time period has been generated, and the processing and analysis of the first traffic data to be processed has been completed. The traffic simulation data is fused with the first traffic data to be processed. This step usually includes time alignment, spatial matching and data normalization of the data to make the two sets of data consistent in time and space dimensions. At the same time, the two sets of data are effectively fused using methods such as the attention mechanism in deep learning to obtain the corrected processed traffic data.

[0099] Substitute the corrected processed traffic data into the preset traffic data prediction model. This prediction model may be built based on machine learning, deep learning or other statistical methods, and can predict future traffic operation conditions based on the input traffic data. The traffic data prediction model predicts real-time traffic operation data based on the input corrected processed traffic data, combined with the model's internal parameters and the knowledge learned during training. The prediction results may include future traffic flow, congestion, average vehicle speed and other indicators. The model outputs the prediction results of real-time traffic operation data. These prediction results will provide important reference for traffic management, signal control, route planning, etc.

[0100] By fusing traffic simulation data with original traffic data, a more complete and accurate data set is obtained, thereby improving the accuracy of traffic operation data prediction. In the face of missing or abnormal data, through data simulation and fusion processing, the system can maintain stable prediction performance and enhance the robustness of the system. Real-time traffic operation data prediction results can provide immediate feedback for traffic management and decision-making, which helps to optimize traffic signal control, alleviate traffic congestion, and improve road capacity.

[0101] In summary, step S104 effectively combines the traffic simulation data with the first traffic data to be processed through data fusion and real-time prediction methods, thereby improving the accuracy of traffic operation data prediction and the robustness of the system, and providing strong support for real-time traffic management and decision-making.

[0102] Step S105, collect the data time characteristics in the real-time traffic operation data prediction results, retrieve the corresponding traffic data according to the data time characteristics, compare the real-time traffic operation data prediction results with the corresponding traffic data retrieved according to the data time characteristics, if the error value of the comparison result is greater than the preset allowable range, optimize the traffic data prediction model, use the optimized traffic data prediction model to process the traffic data, and obtain the optimized traffic data processing results.

[0103] In step S105, the present invention further verifies and optimizes the prediction results of real-time traffic operation data to improve the accuracy and reliability of the prediction results. The specific steps and logic are as follows:

[0104] Extract data time features, such as date and time, from the prediction results of real-time traffic operation data. This information is crucial for subsequent retrieval of actual traffic data for comparison. According to the extracted data time features, retrieve the corresponding actual traffic data from the traffic database. This data should be the real traffic condition record corresponding to the prediction result time, which is used for subsequent comparison with the prediction result.

[0105] Compare the prediction results of real-time traffic operation data with the actual traffic data retrieved. The comparison indicators may include traffic flow, average vehicle speed, congestion duration, etc. These indicators can fully reflect the traffic operation status. Compare the difference between the prediction results and the actual data and calculate the error value. The size of the error value directly reflects the accuracy of the prediction results. Compare the calculated error value with the preset allowable range. If the error value is greater than the preset allowable range, it means that the prediction accuracy of the current traffic data prediction model does not meet the requirements and needs to be optimized.

[0106] Based on the error analysis results, the sources of deviation in the traffic data prediction model are identified. Possible reasons include improper model parameter settings, unreasonable model structure, insufficient training data, etc. To address these issues, the model parameters are adjusted, the model structure is updated, and the optimized traffic data prediction model is retrained on a new data set.

[0107] The traffic data is reprocessed using the optimized traffic data prediction model to obtain optimized traffic data processing results. These results will be more accurate and reliable, providing more powerful support for traffic management and decision-making. By verifying and optimizing the prediction results, the present invention can continuously improve the accuracy of the traffic data prediction model, making the prediction results closer to the actual traffic conditions. By comparing and verifying the actual data with the prediction results, deviations and errors in the prediction model can be discovered and corrected in a timely manner to ensure the reliability and accuracy of the prediction data. The optimization of the prediction model is a continuous process. As traffic conditions change and data accumulate, the model can be continuously adjusted and improved to adapt to new traffic environments and needs.

[0108] Specifically, the traffic data processing method based on artificial intelligence provided by the present invention, step S103 includes:

[0109] The data analysis module aggregates and summarizes the data to generate a statistical overview and trend analysis of the data, and uses association rule mining technology to extract the associations between traffic data items. The associations between traffic data items include the association between traffic congestion in a certain time period and location and the weather, and the association between traffic congestion in a certain time period and location and special events;

[0110] Use deep learning algorithms to train and learn traffic data to automatically extract data features of traffic congestion patterns and vehicle driving trajectories. Deep learning models identify periodic patterns, abnormal patterns in traffic flow, and regular features of vehicle driving trajectories.

[0111] The extracted traffic congestion pattern data features and vehicle driving trajectory data features were subjected to integrity verification to check whether the features were complete and whether there were missing values ​​or outliers;

[0112] If the data are missing, the missing values ​​are handled using interpolation, filling or deletion. If the data are outliers, the outliers are corrected.

[0113] Specifically, the traffic data processing method based on artificial intelligence provided by the present invention, step S104 includes:

[0114] Obtaining generated traffic simulation data, which includes traffic flow predictions and congestion simulation data for different time periods;

[0115] The real-time processing module collects the first traffic data to be processed from the traffic monitoring system, vehicle GPS tracking, and mobile phone signaling;

[0116] The real-time captured traffic data and the simulated data are fused, including time alignment, spatial matching and data normalization. The real-time captured traffic data and the simulated data are fused through the attention mechanism in deep learning to obtain the fused data.

[0117] Loading a preset traffic data prediction model, inputting the corrected processed traffic data into the preset traffic data prediction model, the preset traffic data prediction model predicts real-time traffic operation data based on the input data, combined with the model's internal parameters and the knowledge learned during training;

[0118] The preset traffic data prediction model outputs the prediction results of real-time traffic operation data, and the prediction results of real-time traffic operation data include traffic flow, congestion and average vehicle speed.

[0119] Specifically, the traffic data processing method based on artificial intelligence provided by the present invention, step S105, includes:

[0120] Collecting real-time traffic operation data prediction results, including traffic flow, congestion index and average vehicle speed;

[0121] Retrieving corresponding actual traffic data from a traffic database according to the time characteristics of the real-time traffic operation data prediction results, where the time characteristics include date characteristics and time period characteristics;

[0122] Compare the predicted results of real-time traffic operation data with the actual traffic data retrieved, and the comparison indicators include traffic flow, average vehicle speed and congestion duration;

[0123] Calculate the error value of the comparison result, and determine whether the error value of the comparison result is greater than a preset allowable range;

[0124] If the error value is greater than the preset allowable range, the prediction accuracy of the current traffic data prediction model does not meet the requirements.

[0125] According to the error analysis results, the sources of deviation in the traffic data prediction model are identified, the traffic data prediction model parameters are adjusted, the traffic data prediction model structure is updated, the optimized traffic data prediction model is obtained, and the optimized traffic data prediction model is retrained on the new data set.

[0126] In a second aspect, the present invention provides a traffic data processing system based on artificial intelligence, which is applied to the traffic data processing method based on artificial intelligence, including: a server end, a traffic equipment end, a data collection end and a background control end, the server end establishes a communication connection with the traffic equipment end, the data collection end and the background control end, and the server end includes:

[0127] A data acquisition module, which acquires traffic data, including traffic flow, vehicle speed, traffic video images, and traffic sensor data;

[0128] A data preprocessing module preprocesses the acquired traffic data to obtain traffic data to be processed, and copies the preprocessed traffic data to obtain a traffic data group to be processed, wherein the traffic data group to be processed includes first traffic data to be processed, second traffic data to be processed, and third traffic data to be processed;

[0129] The data analysis module uses a big data analysis method to conduct an in-depth analysis of the first traffic data to be processed, extracts traffic congestion mode data features and vehicle driving trajectory data features through a deep learning algorithm, performs data integrity detection on the traffic congestion mode data features and vehicle driving trajectory data features, and if there is data missing in the traffic congestion mode data features and vehicle driving trajectory data features, determines the missing data time information, and substitutes the second traffic data to be processed and the third traffic data to be processed into a preset traffic data simulation model according to the missing data time information to obtain traffic simulation data;

[0130] Real-time processing module: fusing the traffic simulation data with the first traffic data to be processed to obtain corrected processed traffic data, substituting the corrected processed traffic data into a preset traffic data prediction model to obtain a real-time traffic operation data prediction result;

[0131] The traffic data processing optimization unit collects the data time characteristics in the real-time traffic operation data prediction results, retrieves the corresponding traffic data according to the data time characteristics, and compares the real-time traffic operation data prediction results with the corresponding traffic data retrieved according to the data time characteristics. If the error value of the comparison result is greater than a preset allowable range, the traffic data prediction model is optimized, and the traffic data is processed using the optimized traffic data prediction model to obtain an optimized traffic data processing result.

[0132] Specifically, the traffic data processing system based on artificial intelligence provided by the present invention, the data analysis module is also used for:

[0133] The data analysis module aggregates and summarizes the data to generate a statistical overview and trend analysis of the data, and uses association rule mining technology to extract the associations between traffic data items. The associations between traffic data items include the association between traffic congestion in a certain time period and location and the weather, and the association between traffic congestion in a certain time period and location and special events;

[0134] Use deep learning algorithms to train and learn traffic data to automatically extract data features of traffic congestion patterns and vehicle driving trajectories. Deep learning models identify periodic patterns, abnormal patterns in traffic flow, and regular features of vehicle driving trajectories.

[0135] The extracted traffic congestion pattern data features and vehicle driving trajectory data features were subjected to integrity verification to check whether the features were complete and whether there were missing values ​​or outliers;

[0136] If the data are missing, the missing values ​​are handled using interpolation, filling or deletion. If the data are outliers, the outliers are corrected.

[0137] Specifically, the traffic data processing system based on artificial intelligence provided by the present invention, the real-time processing module is also used for:

[0138] Obtaining generated traffic simulation data, which includes traffic flow predictions and congestion simulation data for different time periods;

[0139] The real-time processing module collects the first traffic data to be processed from the traffic monitoring system, vehicle GPS tracking, and mobile phone signaling;

[0140] The real-time captured traffic data and the simulated data are fused, including time alignment, spatial matching and data normalization. The real-time captured traffic data and the simulated data are fused through the attention mechanism in deep learning to obtain the fused data.

[0141] Loading a preset traffic data prediction model, inputting the corrected processed traffic data into the preset traffic data prediction model, the preset traffic data prediction model predicts real-time traffic operation data based on the input data, combined with the model's internal parameters and the knowledge learned during training;

[0142] The preset traffic data prediction model outputs the prediction results of real-time traffic operation data, and the prediction results of real-time traffic operation data include traffic flow, congestion and average vehicle speed.

[0143] Specifically, the traffic data processing system based on artificial intelligence provided by the present invention, the traffic data processing optimization unit is also used for:

[0144] Collecting real-time traffic operation data prediction results, including traffic flow, congestion index and average vehicle speed;

[0145] Retrieving corresponding actual traffic data from a traffic database according to the time characteristics of the real-time traffic operation data prediction results, where the time characteristics include date characteristics and time period characteristics;

[0146] Compare the predicted results of real-time traffic operation data with the actual traffic data retrieved, and the comparison indicators include traffic flow, average vehicle speed and congestion duration;

[0147] Calculate the error value of the comparison result, and determine whether the error value of the comparison result is greater than a preset allowable range;

[0148] If the error value is greater than the preset allowable range, the prediction accuracy of the current traffic data prediction model does not meet the requirements.

[0149] According to the error analysis results, the sources of deviation in the traffic data prediction model are identified, the traffic data prediction model parameters are adjusted, the traffic data prediction model structure is updated, the optimized traffic data prediction model is obtained, and the optimized traffic data prediction model is retrained on the new data set.

[0150] The technical solution of the present invention effectively solves the data processing errors or data processing timeout problems that may occur in the process of traffic data processing through a series of innovative methods and system designs, thereby improving the operational stability of the traffic system. The specific solutions are as follows:

[0151] First, the acquired traffic data is preprocessed, including cleaning, formatting and other operations to ensure the initial quality of the data. The preprocessed traffic data is copied into multiple copies (such as the first traffic data to be processed, the second traffic data to be processed, and the third traffic data to be processed) for subsequent processing and analysis. This grouping strategy increases the redundancy of the data and provides an alternative solution for handling missing or erroneous data.

[0152] The first traffic data to be processed is analyzed in depth using big data analysis and deep learning algorithms to extract data features of traffic congestion patterns and vehicle driving trajectories, and to perform data integrity verification. If data missing is detected, the second and third traffic data to be processed are used to generate traffic simulation data based on the time information of the missing data to supplement the missing part. This method can effectively deal with the problem of missing data and ensure data integrity.

[0153] The real-time processing module captures real-time traffic data from multiple data sources (such as traffic monitoring systems, vehicle GPS tracking, mobile phone signaling, etc.) and fuses it with simulation data to obtain corrected processed traffic data. The corrected processed traffic data is substituted into the preset traffic data prediction model to obtain the real-time traffic operation data prediction results, and the prediction results are compared and verified with the actual traffic data. If the error between the prediction results and the actual data exceeds the preset range, the traffic data prediction model is optimized, including adjusting model parameters, updating model structure, etc., to improve the prediction accuracy.

[0154] The system includes the server side, traffic equipment side, data collection side and background control side. Communication connections are established between the modules to work together to complete the acquisition, processing, analysis and prediction of traffic data. Through measures such as data grouping, simulation data generation and model optimization, the redundancy and fault tolerance of the system are built, reducing the risk of data processing errors and timeouts. The entire data processing process is highly automated, reducing human intervention and the possibility of human error. The application of deep learning algorithms and big data analysis enables the system to intelligently identify traffic patterns, predict traffic conditions, and provide strong support for traffic management and decision-making.

[0155] In summary, the technical solution of the present invention effectively solves the error and timeout problems in the process of traffic data processing through a series of innovative designs, and ensures the stable operation of the traffic system.

[0156] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations, and the above-described embodiments of the present invention do not constitute a limitation on the scope of protection of the present invention.

Claims

1. A traffic data processing method based on artificial intelligence, characterized in that: include: Step S101, acquiring traffic data, the traffic data including traffic flow, vehicle speed, traffic video images and traffic sensor data; Step S102, preprocessing the acquired traffic data to obtain traffic data to be processed, copying the preprocessed traffic data to obtain a traffic data group to be processed, the traffic data group to be processed including first traffic data to be processed, second traffic data to be processed and third traffic data to be processed; Step S103, using a big data analysis method, deeply analyzing the first traffic data to be processed, extracting traffic congestion mode data features and vehicle driving trajectory data features through a deep learning algorithm, performing data integrity detection on the traffic congestion mode data features and vehicle driving trajectory data features, and if there is data missing in the traffic congestion mode data features and vehicle driving trajectory data features, determining the missing data time information, and according to the missing data time information, substituting the second traffic data to be processed and the third traffic data to be processed into a preset traffic data simulation model to obtain traffic simulation data; Step S104, fusing the traffic simulation data with the first traffic data to be processed to obtain corrected processed traffic data, substituting the corrected processed traffic data into a preset traffic data prediction model to obtain a real-time traffic operation data prediction result; Step S105, collect the data time characteristics in the real-time traffic operation data prediction results, retrieve the corresponding traffic data according to the data time characteristics, compare the real-time traffic operation data prediction results with the corresponding traffic data retrieved according to the data time characteristics, if the error value of the comparison result is greater than the preset allowable range, optimize the traffic data prediction model, use the optimized traffic data prediction model to process the traffic data, and obtain the optimized traffic data processing results.

2. The traffic data processing method based on artificial intelligence according to claim 1, characterized in that: The step S103 includes: The data analysis module aggregates and summarizes the data to generate a statistical overview and trend analysis of the data, and uses association rule mining technology to extract the associations between traffic data items. The associations between traffic data items include the association between traffic congestion in a certain time period and location and the weather, and the association between traffic congestion in a certain time period and location and special events; Use deep learning algorithms to train and learn traffic data to automatically extract data features of traffic congestion patterns and vehicle driving trajectories. Deep learning models identify periodic patterns, abnormal patterns in traffic flow, and regular features of vehicle driving trajectories. The extracted traffic congestion pattern data features and vehicle driving trajectory data features were subjected to integrity verification to check whether the features were complete and whether there were missing values ​​or outliers; If the data are missing, the missing values ​​are handled using interpolation, filling or deletion. If the data are outliers, the outliers are corrected.

3. The traffic data processing method based on artificial intelligence according to claim 1, characterized in that: The step S104 includes: Obtaining generated traffic simulation data, which includes traffic flow predictions and congestion simulation data for different time periods; The real-time processing module collects the first traffic data to be processed from the traffic monitoring system, vehicle GPS tracking, and mobile phone signaling; The real-time captured traffic data and the simulated data are fused, including time alignment, spatial matching and data normalization. The real-time captured traffic data and the simulated data are fused through the attention mechanism in deep learning to obtain the fused data. Loading a preset traffic data prediction model, inputting the corrected processed traffic data into the preset traffic data prediction model, the preset traffic data prediction model predicts real-time traffic operation data based on the input data, combined with the model's internal parameters and the knowledge learned during training; The preset traffic data prediction model outputs the prediction results of real-time traffic operation data, and the prediction results of real-time traffic operation data include traffic flow, congestion and average vehicle speed.

4. The traffic data processing method based on artificial intelligence according to claim 1, characterized in that: The step S105 includes: Collecting real-time traffic operation data prediction results, including traffic flow, congestion index and average vehicle speed; Retrieving corresponding actual traffic data from a traffic database according to the time characteristics of the real-time traffic operation data prediction results, where the time characteristics include date characteristics and time period characteristics; Compare the predicted results of real-time traffic operation data with the actual traffic data retrieved, and the comparison indicators include traffic flow, average vehicle speed and congestion duration; Calculate the error value of the comparison result, and determine whether the error value of the comparison result is greater than a preset allowable range; If the error value is greater than the preset allowable range, the prediction accuracy of the current traffic data prediction model does not meet the requirements. According to the error analysis results, the sources of deviation in the traffic data prediction model are identified, the traffic data prediction model parameters are adjusted, the traffic data prediction model structure is updated, the optimized traffic data prediction model is obtained, and the optimized traffic data prediction model is retrained on the new data set.

5. A traffic data processing system based on artificial intelligence, applied to the traffic data processing method based on artificial intelligence according to any one of claims 1 to 4, characterized in that: include: The server side, the traffic equipment side, the data collection side and the background control side establish communication connections with the traffic equipment side, the data collection side and the background control side. The server side includes: A data acquisition module, which acquires traffic data, including traffic flow, vehicle speed, traffic video images, and traffic sensor data; A data preprocessing module preprocesses the acquired traffic data to obtain traffic data to be processed, and copies the preprocessed traffic data to obtain a traffic data group to be processed, wherein the traffic data group to be processed includes first traffic data to be processed, second traffic data to be processed, and third traffic data to be processed; The data analysis module uses a big data analysis method to conduct an in-depth analysis of the first traffic data to be processed, extracts traffic congestion mode data features and vehicle driving trajectory data features through a deep learning algorithm, performs data integrity detection on the traffic congestion mode data features and vehicle driving trajectory data features, and if there is data missing in the traffic congestion mode data features and vehicle driving trajectory data features, determines the missing data time information, and substitutes the second traffic data to be processed and the third traffic data to be processed into a preset traffic data simulation model according to the missing data time information to obtain traffic simulation data; Real-time processing module: fusing the traffic simulation data with the first traffic data to be processed to obtain corrected processed traffic data, substituting the corrected processed traffic data into a preset traffic data prediction model to obtain a real-time traffic operation data prediction result; The traffic data processing optimization unit collects the data time characteristics in the real-time traffic operation data prediction results, retrieves the corresponding traffic data according to the data time characteristics, and compares the real-time traffic operation data prediction results with the corresponding traffic data retrieved according to the data time characteristics. If the error value of the comparison result is greater than a preset allowable range, the traffic data prediction model is optimized, and the traffic data is processed using the optimized traffic data prediction model to obtain an optimized traffic data processing result.

6. The traffic data processing system based on artificial intelligence as claimed in claim 5, characterized in that: The data analysis module is also used for: The data analysis module aggregates and summarizes the data to generate a statistical overview and trend analysis of the data, and uses association rule mining technology to extract the associations between traffic data items. The associations between traffic data items include the association between traffic congestion in a certain time period and location and the weather, and the association between traffic congestion in a certain time period and location and special events; Use deep learning algorithms to train and learn traffic data to automatically extract data features of traffic congestion patterns and vehicle driving trajectories. Deep learning models identify periodic patterns, abnormal patterns in traffic flow, and regular features of vehicle driving trajectories. The extracted traffic congestion pattern data features and vehicle driving trajectory data features were subjected to integrity verification to check whether the features were complete and whether there were missing values ​​or outliers; If the data are missing, the missing values ​​are handled using interpolation, filling or deletion. If the data are outliers, the outliers are corrected.

7. The traffic data processing system based on artificial intelligence as claimed in claim 5, characterized in that: The real-time processing module is also used for: Obtaining generated traffic simulation data, which includes traffic flow predictions and congestion simulation data for different time periods; The real-time processing module collects the first traffic data to be processed from the traffic monitoring system, vehicle GPS tracking, and mobile phone signaling; The real-time captured traffic data and the simulated data are fused, including time alignment, spatial matching and data normalization. The real-time captured traffic data and the simulated data are fused through the attention mechanism in deep learning to obtain the fused data. Loading a preset traffic data prediction model, inputting the corrected processed traffic data into the preset traffic data prediction model, the preset traffic data prediction model predicts real-time traffic operation data based on the input data, combined with the model's internal parameters and the knowledge learned during training; The preset traffic data prediction model outputs the prediction results of real-time traffic operation data, and the prediction results of real-time traffic operation data include traffic flow, congestion and average vehicle speed.

8. The traffic data processing system based on artificial intelligence as claimed in claim 5, characterized in that: The traffic data processing optimization unit is further used for: Collecting real-time traffic operation data prediction results, including traffic flow, congestion index and average vehicle speed; Retrieving corresponding actual traffic data from a traffic database according to the time characteristics of the real-time traffic operation data prediction results, where the time characteristics include date characteristics and time period characteristics; Compare the predicted results of real-time traffic operation data with the actual traffic data retrieved, and the comparison indicators include traffic flow, average vehicle speed and congestion duration; Calculate the error value of the comparison result, and determine whether the error value of the comparison result is greater than a preset allowable range; If the error value is greater than the preset allowable range, the prediction accuracy of the current traffic data prediction model does not meet the requirements. According to the error analysis results, the sources of deviation in the traffic data prediction model are identified, the traffic data prediction model parameters are adjusted, the traffic data prediction model structure is updated, the optimized traffic data prediction model is obtained, and the optimized traffic data prediction model is retrained on the new data set.

Images