Adaptive Time of Arrival (ETA) Prediction in Light Rail Transport Systems Using Regression and Nodal Segmentation: CDMX Light Rail Case Study

Uncertainty in waiting times is one of the main factors affecting the efficiency of mass public transport. We present a robust model for estimating time of arrival (ETA) on the Mexico City Light Rail (Line 1). Unlike conventional transit models, our proposal employs a nodal segmentation methodology that divides the 13 km route into 17 segments, integrating time series variables and the impact of at-grade road intersections. The results demonstrate an 83% accuracy in predicting arrivals, surpassing the metrics of simple linear regression models when considering time dependence and dwell time at stations. It is important to clarify that this study focuses specifically on the Mexico City Light Rail system and does not address the Mexico City Metro network. Line 1 of the light rail system was selected as the case study because its operational characteristics and available data provide an appropriate environment for evaluating the proposed travel time prediction approach. Consequently, the results presented in this study should be interpreted within the context of the light rail operating conditions considered rather than as a direct representation of the entire urban rail transportation network of Mexico City.