Smartphone-Based Drivers Context Recognition

Various embedded sensors such as accelerometer and gyroscope have opened a new horizon in the scientific studies. One of the most prevailing areas of research is context recognition which can be adopted for smartphone-based parking, road condition detection and sports. To the best of our knowledge, the existing context recognition research covers human’s basic contexts such as walking, jogging and are position dependent that require tightening sensors in fixed position of the body. Furthermore, none of the work has seen to be more specific to detect the contexts of driver. Therefore, to be more specific, in this study, we have constructed a position-independent approach to recognize driver’s contexts that occurs while a driver parks car or leaves from parking place. The support vector machine, random forest and decision tree are employed and the accuracies of 83.38, 93.71 and 98.41% are obtained, respectively