SOLID STATE PRESS
← Back to catalog
Self-Driving Cars: How Machines Learned to Drive cover
Buy on Amazon
US list price $2.99
(search link — ASIN auto-fills once available)
Artificial Intelligence

Self-Driving Cars: How Machines Learned to Drive

LIDAR, Neural Nets, and the Six Levels of Autonomy — A TLDR Primer

Your friend says their car has 'Full Self-Driving' and never touches the wheel. Your professor mentions LIDAR and you nod along without really knowing what it is. This primer clears up both, fast.

Most people can't explain how do self driving cars work beyond 'cameras and computers,' and marketing terms make it worse — Autopilot, FSD, and 'self-driving' get thrown around loosely even though engineers use a strict six-level scale to define them. This book starts with levels of autonomy explained clearly, using SAE's actual Level 0–5 framework, so you can say exactly what a car can and can't do on its own.

From there it walks through the sensor stack (cameras, radar, LIDAR, ultrasonic, GPS, and inertial sensors, and why no single one is trusted alone), the neural networks that turn raw pixels into labeled objects like pedestrians and lane lines, and the prediction-planning-control pipeline that turns 'I see a car' into a smooth, safe steering command. It closes with the parts nobody markets: crash statistics, why engineers roll their eyes at trolley-problem questions, who's liable when something goes wrong, and how far robotaxis really are from true general autonomy.

Written for high schoolers, early college students, and curious parents who want the real picture — no filler, no jargon left unexplained, no textbook-length detour required.

If you want to understand the technology behind the car before you get in one, start here.

What you'll learn
  • Distinguish the six SAE levels of driving automation and what each actually requires of the car and driver
  • Explain the sensor stack — cameras, radar, LIDAR, GPS, IMU — and why redundancy matters
  • Describe how neural networks turn raw sensor data into detected objects, lane lines, and predicted trajectories
  • Outline the perception–prediction–planning–control pipeline that runs many times per second inside an AV
  • Reason about the safety, ethical, and regulatory questions self-driving cars raise, including the trolley-problem framing and its limits
What's inside
  1. 1. What Counts as Self-Driving: The Six Levels of Autonomy
    Defines autonomous driving and walks through the SAE J3016 levels 0–5, clarifying what marketing terms like 'Autopilot' and 'Full Self-Driving' actually mean.
  2. 2. The Sensor Stack: How a Car Sees the World
    Explains the hardware — cameras, radar, LIDAR, ultrasonic, GPS, IMU — and why engineers fuse multiple sensor types instead of relying on one.
  3. 3. Perception: Neural Networks That Turn Pixels into Objects
    Covers how deep learning models detect cars, pedestrians, lane lines, and traffic signs, and how training data and labeling shape what the car can recognize.
  4. 4. Prediction, Planning, and Control: From Seeing to Steering
    Traces how the car forecasts what other agents will do, plans a safe path, and executes smooth steering, braking, and acceleration commands.
  5. 5. Safety, Ethics, and the Road Ahead
    Examines crash data, the trolley-problem framing (and why engineers dislike it), liability, regulation, and what remains between today's robotaxis and truly general autonomy.
Published by Solid State Press · September 2026
Self-Driving Cars: How Machines Learned to Drive cover
TLDR STUDY GUIDES

Self-Driving Cars: How Machines Learned to Drive

LIDAR, Neural Nets, and the Six Levels of Autonomy — A TLDR Primer
Solid State Press

Contents

  1. 1 What Counts as Self-Driving: The Six Levels of Autonomy
  2. 2 The Sensor Stack: How a Car Sees the World
  3. 3 Perception: Neural Networks That Turn Pixels into Objects
  4. 4 Prediction, Planning, and Control: From Seeing to Steering
  5. 5 Safety, Ethics, and the Road Ahead
Chapter 1

What Counts as Self-Driving: The Six Levels of Autonomy

A car company can call its product "Autopilot" or "Full Self-Driving" and still legally require you to keep your hands on the wheel. To cut through the marketing, engineers use a standard scale published by SAE International (originally the Society of Automotive Engineers) called J3016, which sorts every driving system into one of six levels, 0 through 5, based on one question: who is actually doing the driving, the human or the machine?

At Level 0, the human does everything. The car might beep if you drift out of your lane or warn you about a car in your blind spot, but it never steers, brakes, or accelerates for you. Most cars built before the mid-2010s are Level 0.

Level 1 adds a single automated function — either steering or speed control, not both. Standard cruise control, which holds a set speed but does nothing with the steering wheel, is Level 1.

Level 2 combines steering and speed control at the same time, so the car can center itself in a lane and maintain following distance simultaneously. This is where Tesla's Autopilot, GM's Super Cruise, and most systems marketed as "Full Self-Driving" actually sit. The critical thing to understand about Level 2 is that the human is still the driver in the legal and functional sense — the system assists, but you must watch the road at all times and be ready to take over instantly. This is called hands-off, eyes-on: your hands may leave the wheel, but your eyes and attention cannot leave the road.

A very common misconception is that "Full Self-Driving" means the car drives itself. As of this writing, Tesla's FSD is a Level 2 system — powerful, but the driver remains fully responsible for every mile.

About This Book

If you're a high school or college student curious about how do self driving cars work, taking an intro AI or robotics course, prepping for a class project on autonomous vehicles, or a parent trying to make sense of what your kid keeps talking about, this book is for you.

This is a self driving car study guide that covers the full pipeline: the six levels of autonomy explained in plain terms, a lidar vs radar vs camera cars comparison so you know what each sensor actually does, neural networks for self driving cars and how they turn raw pixels into detected objects, and how does autonomous vehicle ai work once it moves from seeing the road to predicting what happens next and steering accordingly. We also untangle Tesla Autopilot vs Full Self-Driving, a distinction most headlines get wrong. A concise overview with no filler.

Read it straight through first, then revisit the worked examples, and finish with the problem set to check what actually stuck.

Keep reading

You've read the first half of Chapter 1. The complete book covers 5 chapters — readable in one sitting.

Continue reading on Amazon