Abstract:
Literature proved the potential benefits of autonomous vehicles in terms of road safety, traffic congestion, and energy consumption. The autonomous vehicles must be supported by advanced sensors and technologies to build reliable awareness of the external environment. However, cars with different levels of automation entail different levels of human intervention during the driving tasks. In this context, the main issue is to determine the interaction between the human and the automated driving system which requires an exhaustive understanding of the driver behavior above all in critical situations. This paper presents a neural network-based classifier of EEG signals to identify the driver’s arm movements by his/her brain electrical activities, when he/she must steer to perform a right or a left turn on a curvilinear trajectory. The classifier based on a time delay neural network (TDNN) aims to classify the human’s EEG signals when the participant executes the action to move his/her arms gripping a real steering wheel while driving in a simulated environmental. The performances of the classifier related to the recognition of the driver’s arm movements by the brain signals demonstrated promising results that are worthwhile to be further explored.
Keywords:
EEG, Identification, Neural network, Autonomous vehicles, Safety.
_______________________________________________________________________
