Personalization of user data to increase the accuracy of the driver assistance system based on machine learning

Pham Tuan Anh

This paper discusses a machine learning approach to personalizing user data to improve the accuracy of a driver assistance system that aims to adapt to driver preferences, driving styles, skills, and driving patterns. Driving sleep warning has become a mandatory feature in modern automotive systems. EAR methods for determining the state of eye openness are complex and inaccurate for people with small and narrow eyes, as, for example, in most Asians. Therefore, the author proposes a system for personalizing user data based on machine learning to improve the accuracy of determining the state of closed/open eyes of Asian faces. The system uses the front camera of the smartphone to collect user data and operate in real time.

Keywords:Driver assistance system, personalization of user data, drowsiness, system accuracy, machine learning.