Member-only story
DeepPiCar Series
DeepPiCar — Part 5: Autonomous Lane Navigation via Deep Learning
Use Nvidia’s End-to-End Deep Learning approach to teach our PiCar to navigate lanes autonomously.
Executive Summary
Welcome back! If you have read through DeepPiCar Part 4, you should have a self-driving car that can navigate itself pretty smoothly within a lane. In this article, we will use a deep-learning approach to teach our PiCar to do the same, turning it into a DeepPiCar. This is analogous to how you and I learned to drive, by observing how good drivers (such as our parents or driving school coaches) drive and then start to drive by ourselves and learn from our own mistakes along the way. Note that for most of this article, you don’t need to have the DeepPiCar to follow along, as we will do the deep-learning on Google’s Colab, which is free.
Introduction
Recall in Part 4, we hand engineered all the steps required to navigate the car, i.e. color isolation, edge detection, line segment detection, steering angle computation, and steering stabilization. Moreover, there were quite a few parameters to hand tune, such as upper and lower bounds of the color blue, many parameters to detect line segments via Hough Transform…
