Scaling Methods To Estimate Macroscopic Fundamental Diagrams in Urban Networks with Sparse Sensor Coverage

Abstract

Accurately estimating traffic variables across unequipped portions of a network remains a significant challenge due to the limited coverage of sensor-equipped links, such as loop detectors and probe vehicles. A common approach is to apply uniform scaling, treating unequipped links as equivalent to equipped ones. This study introduces a novel framework to improve traffic variable estimation by integrating statistical scaling methods with geospatial imputation techniques. Two main approaches are proposed: (1) Statistical Scaling, which includes hierarchical and non-hierarchical network approaches, and (2) Geospatial Imputation, based on variogram modeling. The hierarchical scaling method categorizes the network into several levels according to spatial and functional characteristics, applying tailored scaling factors to each category. In contrast, the non-hierarchical method uses a uniform scaling factor across all links, ignoring network heterogeneity. The variogram-based geospatial imputation leverages spatial correlations to estimate traffic variables for unequipped links, capturing spatial dependencies in urban road networks. Validation results indicate that the hierarchical scaling approach provides the most accurate estimates, achieving reliable performance even with as low as 5% uniform detector coverage. Although the variogram-based method yields strong results, it is slightly less effective than the hierarchical scaling approach but outperforms the non-hierarchical method.