Travel time prediction for school bus

School bus is a convenient and efficient transportation mode. Useful travel time information can help commuters make good choice of departure time as well as other travel decisions. However, providing real time travel time information for an en route school bus is a difficult task as travel time information is uncertain due to external interferences. Hence no consistent tendency can be easily observed. In this research, we develop a novel approach that combines grey theory (for travel time prediction) and empirical decomposition method (for data analysis) to predict the arrival time at each stop along the school bus route. The experiments showed that our prediction approach employing real time data collected from geographic positioning system as well as historic data outperforms the approach involving only historic data. In addition, sensitivity analysis is performed for different number of stops used in the prediction process.