A Hybrid Method for Predicting a Potential Next Rest Stop of Commercial Vehicles

Long-distance trips such as freight and passenger transports over cities can create driver fatigue, so drivers prefer to get a rest for a while during their long-time driving. In Thailand, there are rest stops along main roads between cities, such as petrol stations, travel plazas, wayside parks, and scenic areas. In order to provide a better service to customers, the rest stops must have a good management, so the prediction of the number of potential vehicles in a period of time is primarily needed. One important task is to predict the next rest stop of every car at a period of time. Due to this requirement, this paper aims to introduce a prediction model for predicting the next rest stop of a vehicle by analyzing the global positioning system (GPS) tracking data of all commercial vehicles in Thailand. The proposed prediction model is a hybrid model that comprises of three scoring functions depended on the frequent pattern of connected rest stops, the direction of connected rest stops in a route, and the popularity of the rest stops. The experimental result shows that the proposed prediction model gives high accurate result in terms of the area under the receiver-operating-characteristic curve (AUC). This predicted result is also useful for a government department and rest stops' owner to improve transportation, road safety, and other service.