Freeway Travel Time Estimation Based On Intelligent Algorithm

Now the traffic load on the freeway is increasing and the freeway traffic system will be faced with a very serious challenge,the freeway management department needs simple and intuitive information-assisted decision-making.With the increasing number of freeway data and the increasing types of parameters reflecting traffic conditions,it is difficult for managers to find the key information needed in these complex information.In order to provide better management countermeasures for increasingly congested freeways,this paper selects the most concerned factors in traffic induction and traffic control:travel time as the main research object,and the main research contents are as follows:To achieve more accurate and effective travel time estimation,it is necessary to obtain more comprehensive and accurate traffic data.Therefore,,this paper first analyzes the commonly used data fusion technology and fusion model,proposes the multi-source data collection and preprocessing method,and on the basis of changyi high-speed toll station data,interchange station section data and blocking data,makes an in-depth analysis of the time-space characteristics of travel time and the influencing factors.Artificial fish-swarm(AF)algorithm was used to optimize the parameters of the model.By comparing the optimized parameter SVM model with the original SVM model,the advantages of the optimized parameter model are verified.According to the data survey,it is known that the highway data owned by different highway management departments have certain errors.Therefore,this paper USES different highway data to estimate the travel time,which can provide reference for highway managers with different data.Finally,the G5513 freeway(Changsha west to ningxiang)was used as an example analysis.On the basis of the freeway toll station data,freeway section traffic data,freeway block data and other environmental data on G5513,the support vector machine model of the optimized parameters,the basic BP neural network model,the Kalman filter model,and the multi-core support vector will be used to estimate the future travel time.Compare the results and find the advantages of each model.Different models are used to estimate the needs of the freeway management,focusing on the characteristics of the managenent.