Prediction of Spatiotemporal Evolution of Urban Traffic Emissions Based on Taxi Trajectories

Abstract With the rapid increase of amount vehicles in urban areas, pollution vehicle emissions is becoming more and serious. Precise prediction spatiotemporal evolution traffic plays a great role planning policy making. Most existing methods usually focus on estimating at historical or current moments which cannot well meet demands future planning. Recent work has started to pay attention using multiple attributes related emissions, however, they are not effective efficient enough combination utilization different inputs. To address this issue, we propose joint framework predict based GPS trajectories taxis with multi-channel network motor emission simulator (MOVES) model. Specifically, first estimate spatial distribution matrices through map-matching algorithms. These can reflect status road networks such as volume, speed acceleration. Then, our used efficiently combine three key (volume, acceleration) feature sharing mechanism generate precise them period. Finally, adopt an MOVES model by integrating several factors including predicted states, statistical information vehicles. We evaluate Xi’an taxi dataset. Experiments show that proposed effectively temporal emissions.