| With the explosive growth of private cars in cities,the burden of urban traffic is becoming more and more serious.The traffic congestion caused by excessive car ownership in major cities has become one of the main reasons restricting the rapid development of urban transportation systems.With the rapid development of V2X(Vehicle-To-Everything)communication technology and artificial intelligence technology(AI),the management plan of intelligent transportation system vehicle-toeverything collaborative integrated management provides new solutions to solve the problem of urban traffic congestion.To alleviate urban traffic congestion,we propose a networked intelligent transportation path planning mechanism based on road network dynamic pricing.This thesis mainly studies two aspects,including real-time urban traffic flow information prediction based on deep learning and multi-vehicle dynamic path planning algorithms.The main research contents of this thesis are as follows:First,we propose a prediction model based on traffic flow deep learning.The driving of vehicles on urban roads is mainly affected by the topological structure of the road,speed limits,traffic jams,and other factors,etc.At the same time,the number of traffic on the road exhibits the characteristics of periodic changes.And this is the correlation between temporal and spatial of vehicle movement in cities.We use deep learning algorithms to extract the spatio-temporal correlation,and further predict the traffic data of the road network for a period of time in the future based on this feature.We analyze and compare a variety of common algorithms and simulation results show that the proposed algorithm based on deep learning has a better performance than traditional algorithm based on time series.Among them,the convolutional neural network(CNN)model has the highest prediction accuracy.Then,to alleviate the traffic conditions in the congested area of city,we propose a vehicle path planning algorithm based on dynamic pricing of the road network.Participants in the urban intelligent transportation system(ITS)are transport authority(TA)and vehicle users(VUEs).The goal of the TA is to alleviate the condition of the congested area in the city and balance the traffic flow of the entire city road network,while the goal of the VUE is to complete a journey from the departure point to the destination and minimize its travel costs.First,TA sets dynamic travel prices for different road sections based on the predicted real-time traffic density information,then for VUEs arriving at the intersection,they choose the optimal route at the next moment based on the surrounding road price information to reduce their travel costs.The simulation results show that the proposed algorithm can effectively alleviate the traffic congestion in urban areas and the traffic congestion will be transferred to the less congested area or uncongested area.Compared with the traditional shortest path algorithm such as the Dijkstra algorithm,the algorithm proposed in this thesis has a better performance.The number of congested areas in the road network is less than other algorithms.The average vehicle transit time of the road network is lower than that of other algorithms,throughput is also higher than the shortest path algorithm. |
