Detección de nivel de cansancio de un conductor mediante DEEP LEARNING para reducir los índices de accidentes de tránsito.

Abstract

In this article we present the detection of fatigue level of a vehicle driver and according to this level the autonomous driving assistance is implemented, the detection of fatigue level is based on facial recognition using deep learning (Deep Learning) developed in the Matlab software, applying neural networks previously trained and designed, this detection sends us a metric that comprises four levels, according to the metric the position and velocity control of a simulated Car-Like vehicle in the Unity3D software is performed which presents a user friendly environment with the use of haptic devices, the development of the control algorithm is based on path correction which calculates the shortest distance to reenter the desired path.