Vehicle Speed Estimation Method for Urban Freeway Based on Empirical Mode Decomposition

Urban freeway traffic is a typical nonlinear, time-varying system. Aimed at the problem of estimation traffic flow parameters of urban freeway, a novel vehicle speed estimation method is proposed based on measured traffic flow data. Firstly, the measured traffic flow data are preprocessed. Then, the empirical mode decomposition and reconstruction are carried out to establish the training data set. Finally, based on the neural network algorithm, the non-analytical model between vehicle speed and density is established to realize the speed estimation and prediction. In order to verify the effectiveness of the proposed method and analyze the influence of data preprocessing and empirical modal decomposition on vehicle speed prediction results, and the measured traffic flow data of Beijing third Ring Road are used to test the algorithm. The results show that, traffic flow velocity parameter estimation root mean square error of the proposed method is 3.41, and the Pearson correlation coefficient is 0.87. Compared with BP neural network methods, the method proposed in this paper has higher accuracy for estimating road traffic flow speed.