Forecasting carbon monoxide concentrations near a sheltered intersection using video traffic surveillance and neural networks

In this preliminary study the authors investigated the relationships between traffic and carbon monoxide (CO) concentrations measured near an intersection which is sheltered from the wind by multi-story buildings. Detailed information on traffic parameters was obtained by video camera technology during a single, 4-h period (2-6p.m. local time) of calm winds (¸<1m/s). A neural network was trained with both lane specific traffic information as well as on-site wind parameters and used with an independent test set to predict 1-min average CO concentrations with reasonable accuracy (R²=0.69). Standard linear regression models as well as two dispersion models could not reliably predict CO leavers from the same data set.