Research On Traffic Flow Balance Of Urban Road Network Based On Multi-source Data Fusion

As an important urban infrastructure,urban road network is closely related to people's daily travel.With the increase of urban population and the continuous growth of car ownership,the traffic flow pressure of urban road network is increasing.At the same time,the uneven distribution of road network traffic has become an important factor in the frequent occurrence of traffic congestion and traffic accidents.At present,the construction and development of digital city infrastructure provides a basic guarantee for real-time collection of big data such as road network operation status and real-time distribution of road network flow.Through the real-time collection,storage,analysis,processing and visualization of road network big data,we can comprehensively perceive the real-time state of urban road network operation and traffic accident risk level,which is of great significance to improve the operation efficiency and traffic safety of urban road network.Solving the problem of unbalanced traffic distribution on urban road network mainly depends on two aspects: one is the extraction and processing of traffic data in big data environment,and the other is the calculation effect of traffic allocation strategy and model in intelligent transportation system.The multi-source data fusion and processing of massive and multi-source traffic data can not only more accurately analyze the operation state of the road network,but also make the calculation results of the traffic allocation model more practical.Optimizing the intelligent transportation system and introducing a new traffic balance model and traffic flow prediction method will help to better allocate and optimize the resources on the road network and make residents get a better travel experience.Aiming at the two problems of data processing and model establishment,this paper first carries out the fusion and processing of multi-source data.On this basis,the vehicle dynamic path planning is studied,and the road network flow is predicted.Then,based on the traffic flow equilibrium theory,the traffic flow equilibrium models are established from the global perspective and the user perspective respectively to allocate the traffic flow and make the traffic flow on the road network reach a balanced state.At the same time,the real data is used to simulate the traffic balance,which verifies the effectiveness of the traffic allocation scheme.The main work and contributions of this paper are as follows:(1)In order to use multi-source and massive data to make the research results more accurate,this paper uses track data,mobile phone signaling data and other data from multiple different channels,and integrates multi-source data.Through the fusion of multi-source data,we can better analyze the real-time situation of road network and judge the state of road network.After multi-source data fusion is used as the input data of prediction model and traffic balance model,the prediction results and balance results closer to the actual situation can be obtained,which makes the research results more convincing and practical.(2)In order to better carry out path planning,this paper combines the advantages of static path planning and dynamic path planning,designs different optimization indicators according to different strategies,simulates the path planning process under static and dynamic urban traffic respectively,and plans the optimal travel path for all users on the current road network.At the user level,the travel of users is optimized.At the same time,combined with the two model structures of cyclic neural network and graph neural network,this paper models the spatial correlation of regional traffic and forecasts OD traffic.By predicting the upcoming flow trend in urban areas,the traffic can be dispatched and dredged to avoid low traffic operation capacity and potential safety hazards.(3)In order to balance the traffic flow and allocate and optimize the road network resources,this paper uses the Wardrop equilibrium theory as the theoretical basis,establishes a model based on equilibrium Markov chain from the overall perspective of the system to improve the traffic imbalance on the dynamic road network,and proposes a time-dependent congestion prediction and congestion mitigation algorithm to alleviate the congestion of the traffic network.The model reflects the real-time congestion through the transfer probability matrix,and carries out flow dredging and allocation.From the perspective of the overall system,the model allocates and transfers the traffic flow,optimizes the limited road network resources,realizes the dynamic control of traffic flow to a certain extent,and completes the organization and management of vehicles on the road.(4)In order to solve the problems of large computing load in the traditional centralized traffic flow scheduling methods,a distributed traffic flow scheduling model based on potential game theory is proposed from the perspective of users.The traffic flow system recommends a route to the user through calculation at the mobile terminal without relying on the central platform to schedule the user,so that the user is relatively evenly distributed on each route,so as to achieve a user satisfactory effect.The model completes the distribution and balance of traffic flow from the perspective of users,coordinates the resources on the road network,and improves the travel efficiency of users.To sum up,this paper integrates urban traffic multi-source big data and proposes a series of road network flow prediction and equilibrium models using multi-source data and artificial intelligence algorithms.All models are trained and verified through real data,which provides important solutions and technical support in coordinating traffic management,alleviating traffic congestion and improving travel efficiency.