Mapping to cells: a map-independent approach for traffic congestion detection

Traffic congestion has become a very serious urban problem, which increases travel costs and fuel consumption and reduces the efficiency of the transportation system. However, most of the existing traffic congestion identification efforts rely on high-precision electronic maps and consider traffic congestion as an instantaneous state, which ignores the cumulative effect of vehicle speed fluctuations on traffic state transition. To address the above challenges, this paper proposes a map-independent method for urban traffic congestion detection, which consists of three parts: meter-scale cellbased urban road network reconstruction, cell congestion modeling, and cellular congestion identification. Considering map-related issues, we divided the study area into meter-scale cells and then used floating cab data (FTD) to reduce the road network. The results show that customized congestion metrics are able to monitor changes in traffic speeds within cells and quantify traffic congestion. In addition, the methodology may inform specialized map inferences.