Abstract
This study developed a methodology to create detailed vehicle travel information through a vehicle reidentification system (REID) based on inductive vehicle signatures. A novel feature of this study is to utilize point and section measures, which are outputs of REID, for deriving individual vehicle speed profiles that can be further used to estimate vehicle emissions. The proposed methodology consists of three components. First, characteristics of vehicle maneuvers are identified thru clustering techniques. Second, speed profiles are constructed using a genetically optimized autoregressive model. Third, vehicle emissions are estimated using the Motor Vehicle Emission Simulator (MOVES) emissions model. Next Generation Simulation (NGSIM) data collected from the US 101 in Los Angeles, CA was used for model development and performance evaluation. Results revealed that less than 4% error of estimated emissions was achieved by the proposed method, which is promising for field implementation. It is expected that the outcome of this study will be valuable in developing more efficient and useful traffic surveillance systems for vehicle emissions monitoring.