Online Recommendation System for Autonomous and Human-driven Ride-hailing Taxi Services

With the rapid development of autonomous vehicle technology, modern taxi services will potentially witness an important revolution, where some regular cabs will be substituted by self-driving taxis. In this study, we propose a hybrid taxi recommendation system where both autonomous and human-driven ride-hailing vehicles are guided in order to meet the needs of taxi customers as well as the expectation of human drivers. Customers desire to wait the minimum time before finding a taxi, while drivers aim to maximize their profits by speeding up their customer hunting. Autonomous taxis are used to back-up human-driven taxis in areas where customers' demand levels are low. Three major components compose our system: a taxi demand predictor, taxi-to-region assigner, and taxi routing optimizer. Our simulation model is applied on the borough of Manhattan, New York City (NYC), and is validated with realistic data. The selected results show that significant performance gains in terms of number of pickups, customer waiting time, and vacant traveled distance of human-drivers can be achieved compared to those of the traditional human-based taxi system.