Intelligent Transportation System Based on Smart Soft-Sensors to Analyze Road Traffic and Assist Driver Behavior Applicable to Smart Cities

Abstract

Sensors are considered as a great invention that radically contributed to saving time and effort for humans while moving or transporting things. That's the reason why, researchers are constantly working on a smart vision sensor that can provide important data such as directions, temperature, speed, risks… to drivers. Nowadays, there is a new tendance of smart cities that recommend the equipment of vehicles with sensors that can facilitate communication between all categories of vehicles. In this paper, we proposed an in-depth study for developing an assistance system that assists controls the vehicle position, managed a vision sensor (camera) that provides a perception of the road environment in real-time. Then to calculate the Steering Angle (SA), we use the Smart Soft-Sensors that take into consideration the road edges in the images. Each road image is identified by the SA, these images will be trained in a GPU processor using Nvidia's End to End architecture. Then an adaptive control by applications of the proposed system, gives SA estimated to assist in controlling the actuator (Steering Wheel). The results obtained by this approach encourage integrating it into automotive embedded systems to make the right behavior on the road.