Conference Paper

Neural Network Modeling of Vehicle Gross Emitter Prediction Based on Remote Sensing Data

Coll. of Autom. Sci. & Eng., South China Univ. of Technol., Guangzhou
DOI: 10.1109/ICNSC.2006.1673275 Conference: Networking, Sensing and Control, 2006. ICNSC '06. Proceedings of the 2006 IEEE International Conference on
Source: IEEE Xplore


Vehicle emissions are a significant source of air pollution in cities. A neural network model for vehicle gross emitter prediction was established based on remote sensing data. The states of vehicle emission remote sensing system in China were described first, followed by a brief introduction to idle testing and remote sensing testing. After data collection, the choices in the algorithm and architecture, as well as original data were then analyzed and compared. The back-propagation (BP) neural network model with 7-20-1 architecture was also selected as the optimal approach with satisfied prediction. Compared with traditional model, the proposed approach has better accuracy and generality. The 81.63% correct results show the potentiality and validity of remote sensing for gross emitter prediction by using the neural network

1 Read
  • [Show abstract] [Hide abstract]
    ABSTRACT: Interest has focused on the analysis of vehicle emission based on the remote sensing data during the last two decades. This paper proposes an artificial neural network model for predicting taxi gross emitters using remote sensing data. Firstly, it introduces the field test in Guangzhou, and then analyzes the various factors from the emission data. Secondly, after doing principal components analysis and selecting algorithm and architecture, the back-propagation neural network model with 8-17-1 architecture was established as the optimal approach. It gives a percentage of hits of 93%. Finally, comparison among our former research results and aggression analysis results were presented. The results show the potentiality and validity of the proposed method in the prediction of taxi gross emitters
    No preview · Conference Paper · Jan 2007