Towards Vehicle Emission Estimation from Smartphone Sensors

摘要

CO emissions from transport constitutes a large, and growing, part of the total carbon emissions. We present a model to estimate CO emissions from passenger cars on basis of GPS and accelerometer data gathered from the driver's or passengers' smartphone apps. As part of an experiment to establish ground truth, a method for measuring fuel consumption without instrumenting vehicles is presented. As part of this estimation model, a method for discerning between different driving modes (idle, accelerating, cruising and turning) is presented, using the K-means clustering method. This method will increase the accuracy of the emission model by estimating the fuel consumption while accelerating and braking, as well as estimate idle consumption. The model will enable a more detailed emission inventory, both in terms of location and time. Apart from emission of CO, the method can also be used for estimation of other transport related emissions. The detailed localised emissions can be used to monitor air quality in an area in near real-time as well as help in creating precise emission data for green accounting.