Most road accidents occur due to human fatigue, inattention, or drowsiness. Recently machine learning technology has been successfully applied to identify driving styles and to recognize unsafe behaviors starting from in-vehicle sensors signals such as vehicle and engine speed, throttle position, and engine load. In this work, we investigate the fusion of different external sensors, such as gyroscope and magnetometer, with in-vehicle sensors, to increase machine learning identification of unsafe driver's behavior. Starting from these signals, we compute a set of features capable to accurately describe the behavior of the driver. A support vector machine and an artificial neural network, have then been trained and tested using several features calculated over more than 200 kilometers of travel. The ground truth used to evaluate classification performances has been obtained by means of an objective methodology based on the relationship between speed, and lateral and longitudinal acceleration of the vehicle. The classification results show an average accuracy of about 88% using the SVM classifier and of about 90% using the neural network demonstrating the potential capability of the proposed methodology in identifying unsafe driver's behaviors.

Improving Machine Learning Identification of Unsafe Driver’s Behavior by means of Sensor Fusion

Emanuele Lattanzi
;
Valerio Freschi
2020

Abstract

Most road accidents occur due to human fatigue, inattention, or drowsiness. Recently machine learning technology has been successfully applied to identify driving styles and to recognize unsafe behaviors starting from in-vehicle sensors signals such as vehicle and engine speed, throttle position, and engine load. In this work, we investigate the fusion of different external sensors, such as gyroscope and magnetometer, with in-vehicle sensors, to increase machine learning identification of unsafe driver's behavior. Starting from these signals, we compute a set of features capable to accurately describe the behavior of the driver. A support vector machine and an artificial neural network, have then been trained and tested using several features calculated over more than 200 kilometers of travel. The ground truth used to evaluate classification performances has been obtained by means of an objective methodology based on the relationship between speed, and lateral and longitudinal acceleration of the vehicle. The classification results show an average accuracy of about 88% using the SVM classifier and of about 90% using the neural network demonstrating the potential capability of the proposed methodology in identifying unsafe driver's behaviors.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11576/2678797
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 13
  • ???jsp.display-item.citation.isi??? 8
social impact