| With the rapid growth of social economy,the number of motor vehicles also increases,which leads to urban traffic congestion problem is increasingly serious.How to effectively and quickly solve the problem of traffic congestion on the limited road is a difficult problem in the field of intelligent transportation.With the Development of big data,the traffic flow prediction in the urban traffic network can be realized by using massive data,and the problem of road congestion can be alleviated by guiding vehicles.Traffic flow prediction and dynamic path guidance system are key technologies in the field of intelligent transportation.This paper conducts in-depth research on traffic flow prediction and path guidance algorithm,constructs a short-term traffic flow prediction model based on deep learning method,designs a dynamic path guidance algorithm based on improved ant colony algorithm,and develops a dynamic path guidance system based on short-term traffic flow prediction.The main work contents are as follows:Firstly,a short-time traffic flow prediction model based on spatial correlation of road network is proposed,aiming at the existing methods which seldom consider the influence of different correlation sections on the predicted sections.The Pearson correlation coefficient was used to analyze the spatial correlation of road network sections,and the upstream and downstream sections with strong correlation with the predicted sections were selected.The traffic flow of these sections was used to construct the spatial-temporal feature matrix as the input of the prediction model.This method can reduce the interference of redundant section information.In order to extract the spatial and temporal features of traffic flows more effectively,a combined prediction model based on CNN and Bi GRU was built,and the attention mechanism was added to the model to capture the traffic flows with significant features in different time series.Finally,multiple features were fused to obtain the final output.The experimental results show that compared with other latest models,this model can comprehensively deal with the complex characteristics of traffic flow and has a better prediction effect.Secondly,aiming at the defects of traditional ant colony algorithm and the characteristics of traffic flow dynamic change,a dynamic path induction algorithm based on improved ant colony algorithm is proposed.Based on the traditional ant colony algorithm,a new feedback pheromone is introduced,which includes the length of the actual road,traffic flow and other factors,so that the optimal path is more in line with the actual road conditions.In addition,a permutation factor is introduced into the ant colony algorithm,and the α and β parameters and pheromone recording rules of the traditional ant colony algorithm are designed and optimized to avoid the local optimal problem of the ant colony algorithm.Through MATLAB simulation,the experimental results show that the proposed algorithm can accelerate the convergence rate,enhance the optimization ability,solve the optimal path faster,and achieve real-time and accurate dynamic path induction.Finally,in order to verify the effectiveness of the dynamic path guidance algorithm based on traffic flow prediction,a dynamic path guidance system based on MATLAB-VISSIM was built.Through VISSIM simulation of the actual urban road network system,combined with MATLAB to complete the short-term traffic flow prediction and improved ant colony algorithm of dynamic path induction,through the COM interface for information transmission between the two platforms,the predicted traffic flow feedback to the path induction algorithm to obtain the optimal path,and finally the optimal path provided to the vehicle and implementation of induction.The system can predict the traffic information of the next moment according to the real-time traffic information of each section of the road network,so as to determine whether the vehicles need to re-plan the optimal path and optimize the shortest travel time.Through the simulation experiment,it is verified that the dynamic path guidance system based on traffic flow prediction constructed in this paper can make vehicles avoid the congested sections,effectively shorten the travel time and improve the road network traffic efficiency. |
