Driver State Monitoring through Driving Style Estimation

Human factor is a major component in jeopardization of road safety. Driver State Monitoring (DSM) systems are solutions in charge of detecting impaired driving situations in order to let a vehicle raise an alert and take counteractions. The current paper presents the steps that we are taking towards the development of an AI-powered indirect DSM system based on Driving Style Estimation (DSE) techniques. The paper demonstrates the feasibility of detecting the driver state by analyzing his or her driving behavior. We validated the proposed approach on the detection of aggressive driving situations. This DSM system proved to be effective considering input data spanning a time window of just one second.