A Combination of Feedback Control and Vision-Based Deep Learning Mechanism for Guiding Self-Driving Cars

Abstract

The purpose of this paper is to develop an agent that can imitate the behavior of humans driving a car. When human beings driving a car, he/she majorly uses vision system to recognize the states of the car, including the position, velocity, and the surrounding environments. In this paper, we implemented a self-driving car which can drive itself on the track of a simulator. The self-driving car uses deep neural network as a computational framework to "learn" what is the position of the car related to the road. While the car understands the position of itself related to the track, it can use the information as a basis for feedback control.