Delay Estimation for Shared Rides From GPS Data

Ride-sharing is one of the most innovative topics in the new era of intelligent transportation. RIDE2RAIL is a European initiative that aims to design a system that connects multiple shared rides to one another or to other modes of transportation. Accurate time planning and high user satisfac- tion are key factors for the success of ride sharing schemes, which entails that accurately estimating the actual duration and delays of rides is essential. This paper proposes a method to estimate delays in which shared rides incur during their execution. The proposed mechanism relies on link-based route time estimation: It first makes an initial estimation based on a set of parameters related to the city and the situation (e.g., time and day of the ride); then, it dynamically adjusts the estimation based on GPS data that is continuously received during the ride. The mechanism has been implemented and tested using data collected from rides taken in the cities of Milan and Tehran. The results of the evaluation, which has been performed using several error criteria, showed that the proposed approach has a good level of accuracy.