A multi-source data fusion traffic dynamic guidance method and device and storage medium

Abstract

The application belongs to the technical field of dynamic traffic, and specifically discloses a traffic dynamic induction method based on multi-source data fusion, which comprises the following steps: S1, acquiring traffic information from multi-source heterogeneous data, wherein the data comprises real-time traffic information, navigation map data, road network topology structure and public transportation timetable; S2, performing data fusion processing on the traffic information acquired in S1 to generate fused traffic state estimation; S3, constructing a personalized travel guide model based on the travel history records and preferences of users to provide optimal travel suggestions; S4, monitoring traffic conditions in real time, dynamically adjusting based on user feedback and sudden events, and optimizing travel paths. Through the integration of real-time information from multiple data sources, comprehensive perception of road traffic conditions is ensured, traffic changes in space and time can be processed at the same time, and the influence of data islands and information missing is reduced by dynamically adjusting the weight according to the reliability of the data source.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of dynamic traffic, in particular to a traffic dynamic induction method and device based on multi-source data fusion and a storage medium. BACKGROUND

[0002] With the acceleration of global urbanization, urban transportation systems are facing increasingly severe challenges. Traffic congestion, emergencies (such as traffic accidents, construction), and asymmetric traffic information lead to uneven distribution of road resources, greatly reducing the efficiency of urban transportation systems. At the same time, with the continuous improvement of people's demand for travel experience, how to provide real-time and personalized traffic induction solutions for users to reduce travel time and cost has become a problem to be solved.

[0003] Traditional traffic induction systems usually rely on a single data source (such as traffic monitoring cameras or navigation map data), and due to the lack of multi-source data fusion, it is difficult to comprehensively and accurately perceive traffic conditions, resulting in poor traffic induction effect. In addition, existing systems often cannot flexibly respond to traffic flow fluctuations, accidents and other emergencies, resulting in insufficient dynamic adjustment capability of the system and poor user travel experience. Especially in the case of increasingly strong individualized demand, existing traffic induction methods are difficult to provide travel suggestions that meet individualized preferences and needs, and the travel plan lacks flexibility and pertinence.

[0004] On the other hand, with the advancement of technology, the diversification of data sources provides opportunities for the optimization of intelligent transportation systems. Traffic information can come from multiple heterogeneous data sources, including traffic monitoring cameras, GPS data, road sensors, navigation maps, road network topology, public transportation schedules, etc. However, due to different data sources and heterogeneous formats, how to effectively fuse these multi-source data becomes a key to improving traffic induction effect. At the same time, how to dynamically guide users by real-time adjusting travel plans based on multi-source data fusion through intelligent algorithms is also a major problem faced by existing technologies. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides a traffic dynamic induction method and device based on multi-source data fusion and a storage medium to solve the problem that the prior art lacks multi-source data fusion, making it difficult to comprehensively and accurately perceive traffic conditions, resulting in poor traffic induction effect.

[0006] To achieve the above purpose, the present application is implemented by the following technical solutions:

[0007] A traffic dynamic induction method based on multi-source data fusion, comprising the following steps:

[0008] S1, obtain traffic information from multi-source heterogeneous data, including real-time traffic information, navigation map data, road network topology and public transportation timetable;

[0009] S2, data fusion processing is carried out on the traffic information obtained in S1 to generate fused traffic state estimation;

[0010] S3, based on the user's travel history record and preference, a personalized travel guide model is constructed to provide optimal travel suggestions;

[0011] S4, real-time monitoring of traffic conditions, dynamic adjustment based on user feedback and sudden events, optimization of travel path;

[0012] S5, integrate the data and services of traffic operators, navigation map providers and intelligent travel service providers, build a unified data sharing standard, realize data interoperability and information interaction between different platforms and applications.

[0013] The data fusion in S2 is calculated using a spatio-temporal graph convolution network and a Bayesian data fusion model, wherein the spatio-temporal graph convolution network comprises the following steps:

[0014] Obtain traffic monitoring camera, vehicle GPS, navigation map data and road network topology data;

[0015] The above data is processed using a spatio-temporal graph convolution network to build spatial and temporal dependency relationships, and the traffic state estimation of the spatio-temporal graph convolution network is calculated by the following formula:

[0016]

[0017] wherein, denotes the feature of node at time , denotes the feature of node at time , denotes the neighborhood set of node , is the convolution weight, is the activation function, is the bias term.

[0018] The Bayesian data fusion model comprises the following calculation steps:

[0019] Joint analysis is carried out on traffic information from different data sources, and the Bayesian data fusion model is used to dynamically adjust the weight according to the reliability of each data source;

[0020] The fused traffic state is estimated by the following formula:

[0021]

[0022] wherein, is the traffic state, is the observation data from different data sources, is the prior probability, is the likelihood function.

[0023] The generation of the guidance model in the S3 comprises the following steps:

[0024] Based on the user's historical travel records and preferences, the user portrait is constructed through the adaptive model;

[0025] The future travel behavior of the user is predicted through the bidirectional long short-term memory network, and the prediction formula is:

[0026]

[0027] wherein, and are the weight matrices of the forward and backward networks respectively, and are the forward and backward hidden layer states respectively.

[0028] The optimal travel suggestion in the S3 is generated through the following steps:

[0029] Combined with the historical data of the user and the real-time traffic, the scheme evaluation is carried out through the multi-objective optimization algorithm, and the time, cost and travel comfort are considered;

[0030] The multi-dimensional optimal path scheme is generated through the Pareto optimal solution search algorithm, and the multi-objective optimization model is:

[0031]

[0032] wherein, represents the time cost, represents the cost, represents the travel comfort.

[0033] The dynamic adjustment in the S4 is carried out through the following steps:

[0034] The path selection strategy is dynamically adjusted according to the real-time traffic condition and the user feedback using the Q-learning reinforcement learning algorithm, and the update of the Q value is carried out through the following formula:

[0035]

[0036] wherein, represents the Q value of selecting action in state , representing the next state the maximum Q value of all possible actions, is a learning rate, is a discount factor, is a reward value.

[0037] The optimization of the travel path in S4 includes the following steps:

[0038] Real-time monitoring of external traffic conditions, including traffic flow, emergencies, weather changes;

[0039] Using a Bayesian reinforcement learning model to adjust the user's path selection according to real-time feedback, the weight update of the path adjustment is calculated by the following formula:

[0040]

[0041] wherein, is the updated weight, is a learning rate, is the current feedback reward, is the reward predicted by the model.

[0042] The travel path planning in the S4 step includes the following steps:

[0043] Global path planning to determine the user's main driving route;

[0044] Local path optimization, refine the specific driving path through a hierarchical dynamic programming model, and the optimization model of the global path planning is:

[0045]

[0046] wherein, represents the time cost, represents the cost.

[0047] A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement a traffic dynamic induction method based on multi-source data fusion.

[0048] A readable storage medium, the readable storage medium has a computer program stored thereon, and the computer program is executed by a processor to implement a traffic dynamic induction method based on multi-source data fusion.

[0049] Compared with the prior art, the present application has the following beneficial effects:

[0050] 1、The application integrates real-time traffic information from multiple data sources, navigation map data, road network topology and public transportation schedules to ensure comprehensive perception of road traffic conditions. The spatio-temporal graph convolution network and Bayesian data fusion model can simultaneously process spatial and temporal traffic changes and dynamically adjust weights based on data source reliability. Compared with a single data source, the system can more accurately reflect the actual traffic state, reducing the impact of data silos and information gaps.

[0051] 2、The application analyzes users' historical travel data and preferences, uses a bidirectional long short-term memory network to predict user behavior, and generates personalized travel guidance strategies. Each user has different travel needs, and the system can provide personalized travel recommendations based on individual historical behavior, current traffic conditions and preferences, meeting users' diverse needs for shortest time, optimal route, lowest cost, etc., and improving user experience.

[0052] 3、The application monitors real-time changes in road traffic, such as traffic congestion, accidents or road closures, and the system can adjust travel plans in a timely manner, and use a Bayesian reinforcement learning model to dynamically adjust path selection based on user feedback and real-time traffic data. The system can adaptively adjust travel strategies when dealing with unexpected traffic events, ensuring users get the best dynamic travel plan, greatly improving travel flexibility and accuracy.

[0053] 4、The application provides efficient traffic flow perception and prediction capabilities for traffic management departments, which can better plan signal timing, adjust traffic strategies and optimize road network design by integrating multi-source data and using machine learning and deep learning techniques for real-time analysis and prediction of traffic conditions. Traffic managers can use the traffic guidance data provided by the application to understand the road conditions in a timely manner and make appropriate management decisions.

[0054] 5、The application uses intelligent path planning and dynamic adjustment of real-time traffic conditions to effectively reduce users' detention time on congested roads, avoid unnecessary waiting and detours, and optimize travel paths. At the same time, the system combines multi-objective optimization algorithms to balance time, cost, comfort and other dimensions, helping users choose the most suitable travel plan, reducing unnecessary expenses. BRIEF DESCRIPTION OF DRAWINGS

[0055] Fig. 1 The flowchart of the traffic guidance of the multi-source data fusion of the application;

[0056] Fig. 2 The schematic diagram of the traffic dynamic guidance device of the application;

[0057] Fig. 3 The schematic diagram of the readable storage medium of the application. DETAILED DESCRIPTION

[0058] The technical solutions of the present application will be described clearly and completely below in combination with the drawings of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0059] Embodiment 1

[0060] Reference Figs. 1-3 In the first embodiment of the present application, the present application provides a traffic dynamic guidance method based on multi-source data fusion, comprising the following steps:

[0061] S1, acquiring traffic information from multi-source heterogeneous data, the data including real-time traffic information, navigation map data, road network topology structure and public transportation timetable;

[0062] The multi-source data fusion method of the present application relies on effective collection and cleaning of multiple types of data. In order to achieve this goal, data collection covers a variety of data sources such as real-time traffic flow, vehicle position, road condition information in navigation map, etc. On the basis of these data, standardization and cleaning operations are carried out to ensure that the data can smoothly enter the next stage of fusion processing.

[0063] In this embodiment, the collection process of multi-source data includes the following main steps:

[0064] Collection of real-time traffic information:

[0065] The system acquires real-time data about the current traffic conditions of the road through traffic monitoring cameras, road sensors (such as ground coils) and vehicle GPS devices, etc. These data include vehicle speed, traffic flow, accident events, vehicle position, etc. In order to ensure the timeliness of the data, the present application uses a stream processing framework (such as Apache Flink) to collect and process real-time data streams, so as to ensure low data delay and timely feedback.

[0066] Acquisition of navigation map data:

[0067] By calling public API interfaces such as Gaode Map or Baidu Map, road condition information in real-time navigation map is acquired. Such data includes road length, road classification (such as expressway, trunk road, etc.), road closure information, road congestion degree and speed limit, etc. This information not only provides real-time road traffic conditions, but also provides basic data support for future path planning.

[0068] Acquisition of road network topology structure:

[0069] The system also obtains the latest road network structure data from the city traffic management department, covering the detailed information of the main roads, secondary roads, branch roads and traffic infrastructure (such as bridges and tunnels) in the road network. These data are stored in the form of vector diagrams or topological structure diagrams and serve as an important basis for path planning and traffic guidance.

[0070] Collection of public transportation timetables:

[0071] Through the connection with the city public transportation system, the system obtains data including the departure time, station information, transfer information and line changes of buses and subways. These information supports multi-modal travel plans and helps users to choose a mixed travel plan of public transportation and private cars.

[0072] In this embodiment, the collected data is extensive and heterogeneous, so after the data collection is completed, the data from different sources needs to be standardized and preprocessed to ensure the uniformity of the data format and the alignment of the time stamp, so as to facilitate the subsequent data fusion processing.

[0073] Data standardization: the system standardizes the data of different formats, such as converting the traffic flow information obtained from different devices and different data sources into a unified unit (such as speed unit and time unit). Through this standardization operation, the different data sources can be effectively fused.

[0074] Time stamp alignment: since the information from different data sources has different collection frequencies and time delays, the system uses a sliding time window algorithm to align the time stamp of the data. In each time window, the collection time of different data sources is matched to ensure that synchronous data is used in the same time period. The time span of the sliding window is set by the specific business scenario, generally in seconds or minutes.

[0075] Data cleaning and anomaly processing: in actual data collection, noise data, missing values or abnormal data may be encountered. The system uses the three-sigma method to detect outliers, and if the data deviates from the average value by more than three times the standard deviation, it is determined as abnormal data. In addition, for missing data, the invention uses linear interpolation method for data completion. The interpolation formula is as follows:

[0076]

[0077] wherein, and are the adjacent known values, and are the corresponding time points, is the estimated value of the missing time point. Noise filtering: due to environmental interference, the sensor may be disturbed, resulting in noise in the collected traffic data.

[0078] To this end, the application uses a Kalman filter to filter out noise from the data during preprocessing. The Kalman filtering process is as follows:

[0079]

[0080] where, is the optimal estimate value at time , is the observation value, is the Kalman gain coefficient, is the measurement matrix. The Kalman filter filters out noise by continuously adjusting the state estimate, improving the accuracy of the data. In this embodiment, after data cleaning and standardization, the data will enter the subsequent multi-source data fusion step, and the specific fusion method will be described in detail in the subsequent steps.

[0081] Through the above steps, the system can ensure that the heterogeneous data obtained from different data sources has a unified format and synchronized timestamp, and after cleaning and noise processing, it has high accuracy and reliability.

[0082] S2, performing data fusion processing on the traffic information obtained in S1 to generate a fused traffic state estimate;

[0083] For step S2, the application performs fusion processing based on the multi-source data collected and preprocessed in the aforementioned step S1 to generate a comprehensive and accurate traffic state estimate. The purpose of data fusion is to effectively integrate heterogeneous data of different types and sources, so that the system can accurately perceive and predict traffic conditions. This process combines a spatio-temporal graph convolution network (ST-GCN) and a Bayesian data fusion model to process data from three dimensions of space, time, and reliability of different data sources, ensuring the accuracy and reliability of the traffic state estimate.

[0084] In this embodiment, the data fusion technology is the core step of the intelligent traffic guidance system, ensuring that multi-dimensional information from different data sources can be processed collaboratively to provide accurate basic data for subsequent personalized travel guidance and dynamic adjustment.

[0085] Detailed description

[0086] In this embodiment, the fusion processing of multi-source data adopts the following two core technologies: a spatio-temporal graph convolution network (ST-GCN) and a Bayesian data fusion model, which respectively process traffic flow analysis in the spatio-temporal dimension and multi-data source reliability integration.

[0087] Application of spatio-temporal graph convolution network:

[0088]

[0089] ​In this embodiment, the spatio-temporal graph convolutional network is used to process the spatio-temporal features of road traffic data. The urban road network is modeled as a directed graph, where road segments, intersections, and other transportation infrastructure are treated as nodes in the graph, and the connections between nodes represent the topological structure of the road. In this way, the system can capture both the spatial structure of the road and the temporal changes in traffic conditions.

[0090] Convolution operation: The system first extracts spatial features using the graph convolutional network, which propagates information in the local field on the road topology graph through convolution operations, allowing the traffic state of each road segment to be influenced by its adjacent road segments. Then, the system combines temporal convolution to extract temporal features of traffic states at different times, modeling the changes of traffic flow in the time dimension. The specific spatio-temporal convolution formula is as follows:

[0091]

[0092] where, represents the feature of node at time , represents the feature of node at time , represents the set of nodes adjacent to node , is the weight matrix of the convolution layer, is the activation function, is the bias term.

[0093] Through this operation, the system can generate spatio-temporal related traffic state estimates. Spatial neighborhood influence modeling: In the graph convolutional network, the feature of each node depends not only on its own traffic conditions, but also on the traffic conditions of its neighborhood. Therefore, the system considers the flow, speed, and other information of adjacent road segments during the convolution process, making a more comprehensive estimate of the traffic state of the node.

[0094] Application of Bayesian data fusion model:

[0095] In this embodiment, the Bayesian data fusion model is used to integrate traffic information from different data sources. The reliability, timeliness, and coverage of different data sources vary, so the influence weight of different data sources on the final traffic state estimate needs to be dynamically adjusted during the fusion process. Bayesian inference provides effective theoretical support for this.

[0096] Bayesian inference formula: The system uses Bayesian formula to jointly infer data from different data sources. Suppose we have obtained traffic state from multiple data sources Based on the observation, the Bayesian inference model calculates the fused traffic state estimation by the following formula:

[0097]

[0098] where, is the posterior probability of traffic state estimated based on observation data is the likelihood function of different data sources, is the prior probability of traffic state is the marginal probability of observation data, and the dynamic weight adjustment of data source reliability; in the Bayesian model, the reliability of the data source should be adjusted according to its contribution.

[0099] For more reliable or updated data sources, their weights are higher in the fusion process, while for historical data or low confidence data sources, their weights are lower. In this way, the system can automatically adjust the impact of data according to the changes of data sources, ensuring that the fused traffic state estimation has high reliability.

[0100] Perception fusion and multi-modal data integration:

[0101] To further improve the accuracy of traffic state estimation, the embodiment also integrates multi-modal data fusion to jointly process data from different types of sensors (such as cameras, inductive coils, and vehicle GPS). The multi-modal fusion technology combines video data with time series data to achieve comprehensive perception of complex traffic scenes.

[0102] Application of multi-modal fusion network: Through convolutional neural network (CNN), the data of traffic cameras is processed to extract visual features such as traffic flow and accidents in images. At the same time, long short-term memory network (LSTM) is used to process time series data (such as vehicle speed and position) to predict future traffic changes. Finally, through the fusion network, these different modal data are integrated to form a more accurate traffic state estimation.

[0103] S3, based on the user's travel history records and preferences, build a personalized travel guidance model to provide optimal travel recommendations;

[0104] For step S3, the personalized travel guidance strategy of the present application is based on the multi-source data fusion results in the aforementioned step S2, as well as the user's historical travel data and preferences. By building a personalized travel model, the system can generate optimal travel recommendations according to each user's travel behavior and preferences. To achieve personalized path recommendation, the present application uses adaptive personalized models and deep learning techniques, such as bidirectional long short-term memory network (BiLSTM) for user behavior prediction, and generates personalized optimal travel strategies through multi-objective optimization algorithms.

[0105] Detailed Description

[0106] In this embodiment, the process of personalized travel guidance strategy formulation includes user profiling, user travel behavior prediction, and multi-objective optimization of personalized recommendation. This process combines the user's historical travel records and preferences, and predicts future travel demand through deep learning models to ensure the accuracy of personalized recommendations.

[0107] User Profiling and Historical Travel Data Analysis:

[0108] In this embodiment, by analyzing the user's historical travel data such as common travel time, preferred route, and frequently visited destinations, the system generates a user profile for each user. User profiling is mainly based on historical data and user behavior habits. Through these profiles, the user's travel patterns can be captured, providing a reference for subsequent travel strategy formulation.

[0109] The system uses an adaptive personalized model to classify the user's travel patterns based on their preferences. For example, for commuters, the system identifies their frequently used routes and travel times; for occasional travelers, the system focuses on their past destination and travel time selection preferences. These classification information helps to provide personalized travel strategies for different users.

[0110] User Travel Behavior Prediction:

[0111] In this embodiment, to predict the user's future travel demand, the system uses a bidirectional long short-term memory network (BiLSTM). This network processes user historical travel records to capture forward and backward dependencies between travel time periods and travel patterns, thus predicting the user's future travel behavior.

[0112] The working principle of BiLSTM is to pass the input data forward and backward, and the model can capture complex temporal dependencies based on the context information. The specific formula is:

[0113]

[0114] where, and are the weight matrices of the forward and backward networks, and represent the hidden layer states of the forward and backward, is the activation function, represents the user's future travel behavior prediction at time Through this bidirectional LSTM network, the system can effectively predict the user's departure time, possible destination, and preferred route, providing data support for personalized travel guidance.

[0115] Personalized recommendation and multi-objective optimization:

[0116] In this embodiment, the formulation of personalized travel strategy not only relies on the prediction of user travel behavior, but also combines real-time traffic conditions. In order to meet the multiple travel needs of users (such as shortest time, least cost, higher comfort, etc.), the system optimizes the travel strategy through multi-objective optimization algorithm.

[0117] In multi-objective optimization, the system takes different user travel needs as different objectives for solving, and uses Pareto optimal solution search algorithm to generate multi-dimensional optimal path recommendations. The objective function of multi-objective optimization is:

[0118] The solution search algorithm generates multi-dimensional optimal path recommendations. The objective function of multi-objective optimization is:

[0119]

[0120] where, represents the travel time cost, represents the economic cost (such as fuel cost, road toll, etc.), represents the travel comfort (such as reducing red light stay time, avoiding peak road sections, etc.).

[0121] Decision-making process of personalized recommendation: In actual application, the system combines the user's preference weight to weight the different optimization objectives. For example, if the user values time more, the system will increase the weight of the time objective.

[0122] The system uses Pareto optimization algorithm to find a set of optimal solutions under multi-dimensional objectives, and selects the path scheme that best meets the user's needs from them.

[0123] Through the personalized travel guidance strategy of this embodiment, the system can provide travel recommendations that meet the individual needs of users, ensuring the best balance of travel time, cost, and comfort. In this process, the combination of deep learning model and multi-objective optimization algorithm ensures the intelligence and accuracy of personalized travel strategy.

[0124] S4, real-time monitoring of traffic conditions, dynamic adjustment based on user feedback and sudden events, optimization of travel path;

[0125] For step S4, the travel scheme evaluation and dynamic adjustment of the present application is based on the personalized travel guidance strategy generated in step S3, combined with real-time monitoring of traffic conditions and user feedback, dynamically adjusting the travel scheme. In order to optimize the user's travel experience, the system evaluates the current travel scheme in real time from multiple dimensions such as time, cost, comfort, and dynamically adjusts the travel scheme based on real-time data to ensure that users can obtain the optimal travel path.

[0126] The application introduces a Bayesian reinforcement learning model, combines real-time traffic data, user feedback and prediction, and updates and adjusts the travel strategy in real time, thereby ensuring the flexibility and dynamics of the scheme.

[0127] Detailed description

[0128] In this embodiment, the evaluation and dynamic adjustment of the travel scheme include multi-dimensional evaluation and weight adjustment of the travel scheme, real-time monitoring and feedback mechanism, and dynamic optimization of the path based on Bayesian reinforcement learning. The system dynamically adjusts the generated personalized scheme by real-time perception of traffic flow, accidents, weather and other external factors, and combines the actual travel situation of the user to optimize the path selection.

[0129] Travel scheme evaluation system:

[0130] In this embodiment, the system establishes a multi-dimensional travel scheme evaluation system, mainly evaluating from three aspects of time cost, cost and travel comfort. The needs of each user are different, so the weights of these dimensions are dynamically adjusted.

[0131] The specific model of travel scheme evaluation uses a weighted comprehensive evaluation method, and according to the personalized needs of the user, different weights are given to each evaluation dimension. The specific weighted scoring formula is as follows:

[0132]

[0133] Among them, is the score of each evaluation dimension (such as time, cost, comfort, etc.), is the corresponding weight. The system continuously optimizes these weights according to user feedback and historical data to ensure the personalization and accuracy of the evaluation results.

[0134] Detailed description of the evaluation dimension:

[0135] Time dimension: The system scores according to the difference between the actual travel time and the estimated time. The shorter the time, the less the delay, and the higher the score.

[0136] Cost dimension: The system scores based on economic indicators such as fuel cost and road toll. The lower the cost, the higher the score.

[0137] Comfort dimension: Comfort is scored by reducing the number of red light stops and avoiding congested sections.

[0138] Real-time monitoring and user feedback mechanism:

[0139] In this embodiment, the system monitors traffic flow, accident conditions, and weather changes, among other external factors, in real-time through various sensors, cameras, and traffic management platforms. When congestion, traffic accidents, or other emergencies are detected, the system automatically triggers a dynamic adjustment mechanism.

[0140] The system also receives real-time feedback from users through mobile APPs or in-car navigation systems. For example, users can provide feedback on the accuracy and comfort of the current route, which will serve as important evidence for subsequent scheme evaluation and dynamic adjustment.

[0141] Dynamic adjustment based on Bayesian reinforcement learning:

[0142] In this embodiment, to adjust the user's travel plan in real-time, the system uses a Bayesian reinforcement learning model for path optimization. In reinforcement learning, the system optimizes the user's path selection strategy by constantly interacting with the environment. The Bayesian reinforcement learning model can adaptively adjust the path selection weights when faced with uncertainty or new traffic information, thereby generating a new optimal solution.

[0143] The main steps of the reinforcement learning process include:

[0144] State representation: the system models the current traffic state as a state vector, including real-time traffic conditions, the user's current location, and destination information, etc.

[0145] Action selection: the system selects an action based on the current state , i.e., whether to maintain the current path or switch to an alternative path.

[0146] Action selection is performed through the Q-learning algorithm of Bayesian reinforcement learning.

[0147] Q-value update: the system updates the Q-value of the path based on user feedback and traffic conditions, with the update formula being:

[0148]

[0149] where represents the Q-value of action in state , represents the maximum Q-value of all possible actions in the next state , is the learning rate, is the discount factor, is the reward value.

[0150] Through this process, the system can learn the optimal decision to choose different paths under different traffic conditions. Application of Bayesian inference in path selection: To better handle the uncertainty in the traffic environment, the system dynamically adjusts the selection probability of different paths through Bayesian inference. The Bayesian inference process combines the user's historical feedback and real-time traffic conditions to update the posterior probability of path selection, ensuring that the system can still provide the optimal path selection scheme in complex traffic environments.

[0151] Application scenarios of scheme adjustment:

[0152] Congestion adjustment: When the system detects congestion on the user's current path, the Bayesian reinforcement learning model automatically adjusts the weight of the current path and preferentially recommends alternative paths with less congestion.

[0153] Accident detour: When a traffic accident occurs on the current path, the system updates the traffic status in real time and suggests the user to detour.

[0154] User preference adjustment: If the user feedbacks that the current path does not meet their preferences (such as too long or not comfortable enough), the system will adjust the weight and scheme of path selection according to the user feedback.

[0155] Through this embodiment, the system can flexibly adjust the travel scheme based on real-time traffic conditions and user feedback, ensuring that the user is always on the optimal path. At the same time, the combination of Bayesian reinforcement learning and real-time evaluation enables the system to adaptively cope with complex traffic environments and provide more personalized and intelligent travel schemes.

[0156] S5, integrate the data and services of traffic operators, navigation map providers, and intelligent travel service providers, and build a unified data sharing standard to realize data interoperability and information interaction between different platforms and applications.

[0157] For step S5, the present application integrates multiple traffic-related data providers such as traffic operators, navigation map providers, and intelligent travel service providers to achieve data openness and interoperability, thereby building a unified smart travel service ecosystem.

[0158] The goal of this step is to provide a unified service interface for different types of travel participants and promote the interconnection and interoperation of different platforms and applications through a unified data standard, ensuring smooth data transmission throughout the ecosystem.

[0159] The technical solution in this part is based on existing data sharing and platform interoperability technologies, which can achieve seamless integration of multi-party data and implementation of unified standards.

[0160] Embodiment 2

[0161] In a second embodiment of the present invention, based on the same inventive concept, the present invention proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a traffic dynamic guidance method based on multi-source data fusion as described in the above embodiment.

[0162] Example 3

[0163] In the third embodiment of the present invention, based on the same inventive concept, the present invention proposes a terminal, the terminal comprising: a processor and a memory; the processor and the memory communicate with each other; the memory is used to store instructions; the processor is used to execute the instructions in the memory to execute a multi-source data fusion traffic dynamic guidance method of the above embodiment.

[0164] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0165] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A traffic dynamic induction method of multi-source data fusion, characterized in that, The method comprises the following steps: S1, obtaining traffic information from multi-source heterogeneous data, including real-time traffic information, navigation map data, road network topology and public transportation timetable; S2, performing data fusion processing on the traffic information obtained in S1 to generate fused traffic state estimation; the data fusion processing is calculated using a spatio-temporal graph convolution network and a Bayesian data fusion model, wherein the spatio-temporal graph convolution network comprises the following steps: Obtain traffic monitoring camera, vehicle GPS, navigation map data and road network topology data; Process the above data using a spatio-temporal graph convolution network to build spatial and temporal dependency, and the traffic state estimation of the spatio-temporal graph convolution network is calculated by the following formula: ， wherein, representing a node at time characterized by, representing a node at time characterized by, representing a node of a neighborhood set, is a convolution weight, is an activation function, is a bias term; The Bayesian data fusion model comprises the following calculation steps: Jointly analyze traffic information from different data sources, and dynamically adjust the weight according to the reliability of each data source using a Bayesian data fusion model; Estimate the fused traffic state by the following formula: ， wherein, is the traffic state, is the observed data from different data sources, is the prior probability, is the likelihood function; S3, based on the user's travel history record and preference, build a personalized travel guide model to provide optimal travel suggestions; the generation of the guide model comprises the following steps: Based on the user's historical travel record and preference, build a user portrait through an adaptive model; Predict the user's future travel behavior through a bidirectional long short-term memory network, and the prediction formula is: ， wherein, and are the weight matrices of the forward and backward networks, respectively, and are the forward and backward hidden layer states, respectively. S4, real-time monitoring of traffic conditions, dynamic adjustment based on user feedback and sudden events, and optimization of travel path; S5, integrate the data and services of traffic operators, navigation map providers and intelligent travel service providers, build a unified data sharing standard, and realize data interoperability and information interaction between different platforms and applications. 2.The traffic dynamic guidance method of multi-source data fusion according to claim 1, wherein, The optimal travel suggestion in S3 is generated by the following steps: Combine the user's historical data and real-time traffic conditions, and evaluate the scheme through a multi-objective optimization algorithm, considering time, cost and travel comfort; Generate a multi-dimensional optimal path scheme through a Pareto optimal solution search algorithm, and the multi-objective optimization model is: ， wherein, represents the time cost, represents the cost, represents the travel comfort. 3.The traffic dynamic guidance method of multi-source data fusion according to claim 1, wherein, The dynamic adjustment in S4 comprises the following steps: Real-time monitoring of external traffic conditions, including traffic flow, sudden events and weather changes; Adjust the user's path selection according to real-time feedback using a Bayesian reinforcement learning model, and the weight update of path adjustment is calculated by the following formula: ， wherein, is the updated weight, is the learning rate, is the current feedback reward, is the model predicted reward.

4. The traffic dynamic induction method of claim 1, wherein, The dynamic adjustment in S4 is performed by the following steps: Use the Q-learning reinforcement learning algorithm to dynamically adjust the path selection strategy according to real-time traffic conditions and user feedback, and the update of Q value is performed by the following formula: ， in, Indicates the state Next action Q value, Indicates the next state The maximum Q value of all possible actions. For learning rate, As a discount factor, This is the reward value.

5. The traffic dynamic induction method of claim 1, wherein, The optimization of travel path in S4 comprises the following steps: Global path planning to determine the main driving route of the user; Local path optimization to refine the specific driving path through a hierarchical dynamic programming model, and the optimization model of global path planning is: ， wherein, represents a time cost, represents a cost of money.

6. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the traffic dynamic induction method of multi-source data fusion according to any one of claims 1 to 5.

7. A readable storage medium, characterized by, The readable storage medium stores a computer program, and the computer program is executed by the processor to realize the traffic dynamic induction method of multi-source data fusion according to any one of claims 1 to 5.

Images