Impact of emotional states on the effective range of electric vehicles

Over the last decade, a large interest in reducing transportation dependence on fossil fuels as well as the cost reduction in battery technologies, have driven the electric cars market uptake. However, information is scarce about factors that affect the driving range. Besides the battery’s capacity, other factors may affect the overall vehicle’s range, for instance: driving behavior, fluctuations in temperature, number of battery cycles, etc. Accordingly, this paper proposes an approach to evaluate the impact of emotions and driving behavior on the range of electric cars using physiological signals and vehicle performance features. This work was developed in three stages. During the first stage, the heart rate and galvanic skin response of 20 volunteers were recorded from biosensors. The vehicle’s data was obtained from a driving simulator. Afterward, the driving profile was used as an input source to simulate an object-oriented electric vehicle model to estimate the driving range. Finally, during the third stage, feature selection techniques and subject-dependent classifiers were evaluated using metrics such as the accuracy and the area under the curve. Support-vector machines with radial kernel and tree-bagged models provided the best global performance with the bio-signals and driving performance subsets to discriminate between calm and aggressive driving. Results showed that driving behavior could be evaluated from physiological and vehicle features. Furthermore, the subjects’ statements showed that users’ beliefs, thoughts, and prior social contexts influence the way they perceive driving behavior. Reductions in the range of up to 68% when driving aggressively compared to a calm manner were found.