Research on Dynamic Traffic Flow Information Prediction Under Sud-den Congestion
[Purposes]This paper aims to establish a model that can accurately predict traffic flow under sudden congestion conditions,so as to better cope with sudden traffic congestion.[Methods]The mini-mum prediction period of traffic flow is derived from chaos theory,and based on that,the CDNN predic-tion model is proposed.Taking APE,MAPE and RMSE as the evaluation indexes,the prediction stability and accuracy of the CDNN,FCM,DLA and NN models are analyzed in comparison with the highway cases,and the characteristics and mechanisms of the various types of prediction models are summarized.[Findings]Dynamic traffic flow prediction period should be not less than 80s;CDNN model prediction performance is the best at the moment of emergencies,followed by no emergencies,and the duration of emergencies is relatively poor;CDNN model prediction performance is better than that of FCM,DLA and NN model.[Conclusions]The CDNN prediction model provides a new theoretical and practical way for traffic flow prediction under unexpected congestion conditions with better prediction performance,high-lighting its potential application value in coping with unexpected traffic situations,which is of great sig-nificance for future traffic management and planning.
NETL
NSTL
万方数据
