SmartTaxiNet: Mobile Service of Driver-Passenger Matching and Taxicab Scheduling System
Taxi drivers always suffer from empty cruising while at the same time passengers sometimes have problem to hire a taxi on the road. The taxi booking systems and Social Networking Service based approaches are introduced to solve the problem, but they often have flaws in terms of efficiency and skipper detection.
We propose and implement a mobile computing system named SmartTaxiNet to solve the driver-passenger matching problem. It provides location-aware services to help taxicabs and passengers find each other quickly. In addition, strategies such as skipper detection scheme are designed to meet the practical needs as well as clients' contentment. The CabMatch algorithm is designed and applied in the system to perform driver-passenger matching within the complexity .
Based on the analysis of the accumulated matching data, the system can also estimate potential passenger locations and traffic patterns. Furthermore, by using real GPS trajectories of over 500 taxis, picking up 460,000 passengers over 30 days in San Francisco, we predict that if SmartTaxiNet were deployed, the number of active taxis could be reduced to only 50% of the current amount to satisfy citizens’ daily travelling requests.
SmartTaxiNet is fully implemented and tested on SJTU campus. Experimental results show that the average passengers' waiting time and the average fuel waste for empty cruising can be reduced by approximately 27% and 45%, respectively. Large scale simulation also indicated similar result.