FUSION OF REAL-TIME CONGESTION PATTERN TO LAND USE REGRESSION MODEL FOR AIR POLLUTION

Abstract

With unprecedented advancement in mobility technology, most of the vehicles in the world are still operating on fuels from natural resources. On burning these vehicles contribute to air pollution significantly affecting life of every individual with maximum concern for the older as well as to newest generations. Thus, it becomes extremely important to measure and model air pollution and take preventive actions as efficiently and quickly as possible. For modelling, traffic characteristics like volume, and density near fixed monitoring sites plays an important role. These flow characteristics can also be coupled with nearby land use to give a better spatio-temporally varied model for pollutant prediction. Real-time congestion data can provide a fast and accurate measure of various pollutants that a person can expect on a particular route. It can significantly help non-motorized transit users and active users to plan their route based on the greenest route available. In the present study, real-time congestion information is fused in a land-use regression model. The former is obtained from HERE maps Traffic Flow API (Application programming interface). To integrate land use information in the model, each raster pixel for the data inside the buffer region can be converted to a point, and a value is assigned to it, which is based on its distance from the monitoring station, land use, and traffic flow. Using these point data, regression analysis can be done to obtain a predictive model which can be used along any route to give a better-estimated value of pollutant concentration experienced by the user. These results can be integrated with map services to give a greener and safer route for active and non-motorized users leading to sustainable development.