Mode-Wise Corridor Level Travel-Time Estimation Using Machine Learning Models

Abstract

This research is oriented towards exploring mode-wise corridor level travel-time estimation using Machine learning techniques such as Artificial Neural Network (ANN) and Support Vector Machine (SVM). Authors have considered buses (equipped with in-vehicle GPS) as the probe vehicles and attempted to calculate the travel-time of other modes such as cars along a stretch of arterial roads. The proposed study considers various influential factors that affect travel time such as road geometry, traffic parameters, location information from the GPS receiver and other spatiotemporal parameters that affect the travel-time. The study used a segment modeling method for segregating the data based on identified bus stop locations. A k-fold cross-validation technique was used for determining the optimum model parameters to be used in the ANN and SVM models. The developed models were tested on a study corridor of 59.48 km stretch in Mumbai, India. The data for this study were collected for a period of five days (Monday-Friday) during the morning peak period (from 8.00 am to 11.00 am). Evaluation scores such as MAPE (mean absolute percentage error), MAD (mean absolute deviation) and RMSE (root mean square error) were used for testing the performance of the models. The MAPE values for ANN and SVM models are 11.65 and 10.78 respectively. The developed model is further statistically validated using the Kolmogorov-Smirnov test. The results obtained from these tests proved that the proposed model is statistically valid.

Keywords

• machine learning
• travel-time
• support vector machines
• artificial neural networks