Performance of design options of automated ARIMA model construction for dynamic vehicle GPS location prediction

Abstract

Many applications in intelligent transportation systems are demanding an accurate vehicle Global Position System (GPS) location prediction. In this study, we satisfy this demand by designing an automated GPS location prediction system based on the well known traditional Auto-Regressive Integrated Moving Average (ARIMA). To increase the proposed model accuracy, make it dynamic, and reduce its execution time, the traditional ARIMA model has been modified extensively by using different combinations of design options of the model. To perform GPS location prediction, the proposed model depends the previous recorded vehicle locations, speed, and bearing reading to predict the vehicle future locations. To make the proposed model dynamic, it is designed to regenerate all its parameters periodically. To deal with such dynamic environment, only a specified window of the historical data is used. To reduce the regeneration of the model execution time, the model selection process is enhanced and several model selection approaches are proposed. The proposed model and the different design options are evaluated using a realistic vehicle dataset traces that are recorded using a GPS receiver embedded in a smart phone, as well as, using traces from a previous study called the INFATI Dataset. To deal with any imperfection in the data used in generating the model in this study, we propose a novel anomaly detection and smoothing technique. The results show that the proposed framework can generate ARIMA models that can predict the future GPS locations of a vehicle accurately and with a reasonable execution time. The results also show that the proposed model can predict the vehicle’s location for several future steps with an acceptable accuracy.