Driver status detection based on internet of things

Monitoring drivers' physiological conditions such as sleepiness, fatigue and stress has an important role in ensuring road safety. In this thesis, both offline and online systems are proposed to provide driver status detection. Physiological signals were used in order to detect the stress condition that occurs during driving in drivers. Physiological signals consist of electrodermal activity (EDA) and heart rate signals. In the proposed offline system, a dataset was constituted with physiological signals received in 20 sessions from 12 participants in the prepared experimental environments. Then, the dataset was classified with machine learning algorithms and offline driver status was detected. The classification accuracy obtained with multi-participant data using the entire dataset was 80.1%. In addition, the classification accuracy obtained with data from a single participant in 5 different sessions was found to be 90%. The proposed online system is based on the Internet of Things (IoT) technologies. With online driver status detection, drivers' physiological signals can be monitored remotely in real time and various alerts can be generated when necessary. In this thesis, with the proposed offline and online system, the stress conditions of the drivers can be detected. In this way, it contributes to road safety. Additionally, the IoT concept and its application areas, which have become popular in recent years, was examined in this thesis.