Estimation of Transit OD Matrix Using Boarding and Alighting Data: A Case Study of Haridwar-Rishikesh Metro Corridor

Abstract

Transit OD matrix is fundamental to service planning and operation. Traditionally, transit OD matrices are generated through on-board personal interview surveys, albeit highly expensive, time-consuming, and labor-intensive. Therefore, utilizing the readily available data in the estimation process has been the subject of many research for the past few decades. This study focuses on the estimation of transit OD matrix using boarding and alighting passenger counts, which can further be extended to smart-card data. Four different methods from literature are explored, namely, IPF, Markov Model, Uncertainty Maximization, and Compressed Sensing. Furthermore, a new model for route-level transit OD estimation is introduced. Unlike compressed sensing, the proposed model employs $${\mathcal{\ell}}:{\infty }$$ norm regularizer instead of Euclidean norm and uses spot information about the critical cell in each direction. The proposed model is found to outperform the existing methods when tested for RMSE value and trip length distributions for true OD matrices. In the most favorable scenario, the RMSE value was found to be 18 compared to 145 and 95 for compressed sensing and IPF, respectively. Lastly, the trip matrix for the Haridwar-Rishikesh metro corridor was estimated using the above methods and the results were compared based on their similarities and differences.