Developing Adaptive Methods for Traffic Flow Modeling and Congestion Reduction

Urban congestion mitigation often assumes that traffic management systems have access to complete and reliable traffic observations. In data-limited urban environments, however, observations may be sparse, noisy, delayed, or partially missing. Adaptive signal control based only on point estimates can therefore make suboptimal decisions when the estimated traffic state is unreliable. This article proposes a methodological framework that links traffic-flow modeling, uncertainty assessment, and confidence-aware adaptive control for urban intersections with limited observations. The framework combines literature review, data preparation, model calibration, controlled missing-data experiments, and comparison against fixed-time and conventional adaptive signal-control baselines.