US 2,012,083,960 · Filed 2011-10-03

Google's Crystal Ball for Self-Driving Cars: Predicting What Other Vehicles Will Do

Google patented a system that helps self-driving cars predict what other vehicles and pedestrians will do next—not by reading minds, but by analyzing their current movement and position. It's like the difference between just seeing a car ahead and actually guessing whether it's about to turn left, brake, or speed up.

The plain-English version

What it protects

The claim covers a computer system that observes detected objects (other cars, pedestrians, cyclists) on the road, analyzes their current behavior and trajectory, and predicts their future movements and actions. What's protected here is the specific combination of detection, motion analysis, and behavioral forecasting—a method that autonomous vehicles use to anticipate what nearby road users will do in the next few seconds.

Why it matters

For self-driving cars, prediction is survival. A car that only reacts to what it sees right now is dangerous; a car that can forecast a pedestrian stepping into the road or a vehicle changing lanes has time to brake or swerve. Google's patent covers foundational technology for making autonomous vehicles safer by giving them foresight—a critical competitive advantage in a field where split-second decisions determine safety outcomes.

Real-world use

When a self-driving car approaches an intersection and detects a cyclist ahead slowing down, this system predicts whether the cyclist will stop, turn, or continue—so the car can adjust its speed and path accordingly before anything happens.

Original USPTO abstract

Aspects of the invention relate generally to autonomous vehicles. Specifically, the features described may be used alone or in combination in order to improve the safety, use, driver experience, and performance of these vehicles.

Patent details

Publication number
US 2,012,083,960
Filing date
2011-10-03
Grant date
Application — not yet granted
Assignee
Google Inc.
Inventor(s)
ZHU JIAJUN, FERGUSON DAVID I., DOLGOV DMITRI A.
CPC class
B60T7/22

Want to file your own patent?

If you're designing safety features or autonomous systems for vehicles, check our free scanner to see if your prediction method steps on existing automotive patents.

Free patentability scan