Identification and Prediction of Urban Traffic Congestion via Cyber-Physical Link Optimization

In this paper, to accurately predict and evaluate the traffic states in local urban areas, an effective as well as efficient algorithm combined with leakage integral echo state network (LiESN) and Pearson correlation analysis is presented. First, aiming at the parameter optimization of LiESN, the differential evolution algorithm is used to calibrate the key parameters. Second, the road weight allocation is carried out by Pearson correlation analysis considering the specific topology characteristics of local road nets. On this basis, the corresponding flow-frameworks are designed. Finally, the congestion delay data of local road nets in Chongqing Nan'an district, acquired from AutoNavi, are used to verify the validity and rationality of our method. Experimental results indicate that the proposed identification and prediction mechanism relates the congestion indicators to actual traffic operating mechanism in a more reasonable way.