Brain-ADAS: A brain-operated Advanced Driver-Assistance System for Collision Avoidance

Abstract

According to the World Health Organization (WHO), 1.25 million deaths result from road accidents every year. Consequently, lots of technologies have been developed in order to reduce the accidents to save people’s lives. This project aims at developing an advanced driver- assistance system to increase the safety of the car by preventing collisions in emergency situations. It combines inputs from the surrounding environment with driver’s brain Electroencephalography (EEG) signals. By analyzing EEG activity, emergency braking can be initiated before the driver executes the physical action of braking. This could help in taking an appropriate action, whether by performing a brake and/or steer maneuvers, to prevent serious accidents. Temporal signatures in the recorded EEG that correspond to emergency braking are extracted from the recorded EEG. Such signatures are then used to train different machine learning algorithms that are subsequently used to decode the driver’s EEG activity to infer his/her intention. Implemented algorithms have been examined on data recorded throughout the project using the Emotiv EPOC neuroheadset. The proposed system comprises a computer vision component that analyzes the surrounding driving environment and augment the EEG-based decision. The proposed system has also been tailored to operate within the AUTOSAR framework.

Links

Documentation

Power Point Presentation

Demo Video

System Block Diagram