Zhenghua Zhang,∗ Jiafeng Zhang,∗∗ Rui Gao,∗ Chongxin Fang,∗∗∗ and Jin Qian∗∗∗∗
Traffic pollution and congestion, wireless sensor network, datapreprocessing, correlation coefficient, optimized BP network∗ Information Engineering College of Yangzhou Univer-sity, Yangzhou, Jiangsu Province, 225127, China; e-mail:[email protected], [email protected]∗∗ Information Engineering College of Yangzhou Univer-sity, Nantong, Jiangsu Province, 226371, China; e-mail:[email protected]∗∗∗ Information Engineering College of Yangzhou Univer-sity, Yangzhou, Jiangsu Province, 21
To study the relationship between traffic pollution and congestion, meanwhile, to encourage people to adopt environmentally friendly transportation, a back propagation (BP) neural network is proposed based on the correlation coefficient optimization. We can find some links between pollution and congestion, meanwhile predict road congestion. The monitoring of traffic data is carried out first. Next, the traffic data are subjected to data preprocessing. Two methods of error elimination and adaptive weighting are used. This makes complex multivariate data simple and accurate. The next step is to optimize the initialization weight by analysing the correlation coefficient between each pollutant and the congestion index. Finally, training traffic congestion and pollution data samples to get a prediction model. The experimental results show that the fusion scheme based on data preprocessing, correlation analysis and BP neural network modelling makes the error of prediction model smaller. The mean absolute error (MAE) and root-mean- square error (RMSE) are 0.086 and 0.1293, respectively. It can effectively reflect the correlation between pollution and congestion. The prediction model can obtain more accurate prediction values. It proves that the fusion scheme of traffic congestion and pollution data is effective.
Important Links:
Go Back
