Adaptive learning for efficient driving in urban public transport

Concerns about global warming and energy costs have induced transport companies to take measures to reduce fuel consumption. Among the different options available, efficient driving is widely used, allowing a reduction in fuel consumption of around 10%. However, changing the driver's behavior is not exempt of problems. The success of efficient driving techniques in the long term is related to the motivation of the driver and, for that reason, an adaptive training system according to the driver's needs can prove much more successful than giving general instructions that do not solve their inefficiencies. Therefore, a properly description of the driver's behavior and the adaptation of the evaluation analysis to the context is a key factor for the learning process in efficient driving. In this paper we propose an adaptive learning system for efficient driving, which allows the evaluation of professional drivers of urban public transport in their work environment. With the proposed system, we can identify failure points relating to the context, making a focused evaluation. The first evaluation results show that the set of patterns designed to evaluate the application of efficient driving techniques can identify the incorrect actions of the drivers. Based on the results, it is possible to make personalized recommendations to improve driver performance.