A Robust Hybrid Model Based on Kalman-SVM for Bus Arrival Time Prediction

Abstract

Transport management is a system that allows travelers to plan, schedule and manage their journey from takeoff till arrival. Passengers involved in the journey either use private or public vehicles and they always estimate their arrival-time. As the total vehicles on the road increases, the possibility of traffic congestion would also increase which is becoming a major problem for traffic planning and scheduling. To address the issue of congestion, many scholars’ opted for solutions such as building a proper infrastructure and developing an accurate travel time prediction models such as historical datasets and statistical models. However, these solutions are not efficient from an economic perspective. Therefore, the key goal of this research paper is to propose a new robust model that could forecast the arrival-time of buses by using support vector machine and Kalman filtering algorithm. The experimental results showed that the prediction model produced an outstanding performance in terms of prediction accuracy over the available historical traffic data (achieved) models using Genetic Algorithm – Support Vector machine (GA-SVM) and Artificial - Neural - Networks (ANN). Based on the analysis, the performance of this prediction model is 87.63% better than ANN which was 83.5%. The proposed model has contributed to the field of study by introducing an integrated process for estimation bus arrival-time prediction in Artificial, data mining and Transportation Engineering field.