Wearable Sensors for Evaluating Driver Drowsiness and High Stress

Abstract

High levels of stress, drowsiness and lack of concentration, are some of the main factors that affect the drivers, which can lead to traffic congestion and even accidents. One of the challenges that has caught the attention in the area of research of prevention of traffic accidents, it is the generation of mechanisms that contribute to monitoring and evaluating the driver behavior. This paper presents a prediction model based on Machine Learning techniques to detect cognitive states based on the monitoring of the electroencephalographic signals of drivers of vehicles acquired in various real driving scenarios. The proposed prediction model consists of three phases: 1. - Acquisition of electroencephalographic signals from drivers in real time. 2. - Select and extract the main characteristics of the signals. 3. - Develop the prediction model using the Support Vector Machine (SVM) algorithm. Finally, to evaluate the performance of the proposed model, it was compared with two Machine Learning Techniques: K-Nearest-Neighbors (KNN) and Logistic Regression (LR). The results obtained through the experiments demonstrate that the proposed model has the best performance in the evaluation and prediction of the cognitive workload of the physiological signals of the conductors, with a 93% accuracy in the classification of the information compared to other models that have an 84% and 54% accuracy in the classification.