Chapter 16. Simulating a Self-Driving Car Using End-to-End Deep Learning with Keras
From the Batmobile to Knightrider, from robotaxis to autonomous pizza delivery, self-driving cars have captured the imagination of both modern pop culture and the mainstream media. And why wouldn’t they? How often does it happen that a science-fiction concept is brought to life? More important, autonomous cars promise to address some key challenges urban societies are facing today: road accidents, pollution levels, cities becoming increasingly more congested, hours of productivity lost in traffic, and the list goes on.
It should come as no surprise that a full self-driving system is quite complex to build, and cannot be tackled in one book chapter. Much like an onion, any complex problem contains layers that can be peeled. In our case, we intend to tackle one fundamental problem here: how to steer. Even better, we don’t need a real car to do this. We will be training a neural network to drive our car autonomously within a simulated environment, using realistic physics and visuals.
But first, a brief bit of history.
Get Practical Deep Learning for Cloud, Mobile, and Edge now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
