Development of a Visual System to Improve Off Road Vehicles Safety

ADRIÁN ORELLANA; TOMÁS DURÁN; JUAN CARLOS CORREA; ANDRÉS LAGE

Neuquén

Congreso; JAR 2019 Jornadas Argentinas de Robótica; 2019

Universidad Nacional del Comahue

This work describes the advances made to improve the safety of huge off road vehicles as mining trucks. Using as a starting point the system developed in [1], drowsiness of the driver is intended to be detected. The captured infrared (IR) images are processed with the aid of artificial intelligence (AI) to detect blinking of the eyes, even when the operators are usingdark protection glasses, mandatory in daily conditions. Besides, processing the IR stereo camera information provided by the device, a 3D point cloud is obtained from the driver?s head and its position adds information to the system in order to determine drowsiness of the driver. The sensor has an RGB camera that will be used to get the steering wheel movements, which is also intended to add useful information to estimate drowsiness of the driver. The Developed system is prepared to associate with it GPS position of the truck, intended to interact with a local map and a central control, to evaluate distances to dangerous zones as precipices, buildings and proximity to other vehicles. The strategy of using non-contact sensors as Intel RealSense with IR stereo camera plus RGB device, was chosen because of the rejection most truck drivers show, when electrodes or special glasses with internal cameras are intended to be used with them to get their drowsiness state.