How Self-Driving Cars Detect and Respond to Road Emergencies

Waymo's patent describes how an autonomous vehicle uses sensors to spot unexpected dangers like people in the road or unusual traffic signs, then takes action to keep everyone safe.

What it covers

The patent describes a self-driving vehicle designed to handle unexpected dangers on the road. It uses a "perception system" with sensors to detect objects like people, other vehicles, or traffic signs around it (Claim 1). A "control system" then looks at this sensor data to figure out if there's an "emergency situation" (Claim 1). This check might involve seeing if a person is in an "unexpected location," like a traffic lane (Claim 2), or if another vehicle is blocking the road (Claim 6). If an emergency is detected, the vehicle identifies a safe action, such as maneuvering away from the danger (Claim 13) or rerouting to avoid it (Claim 14). For example, if the vehicle's sensors detect a broken-down car blocking a lane ahead, the control system could decide to change lanes or take a different route.

What it doesn't cover

• —Does not cover detecting internal vehicle malfunctions, like a flat tire or engine trouble, that are not related to external objects.
• —Does not cover situations where a human driver is actively controlling the vehicle and responding to hazards.
• —Does not cover detecting potential theft of the vehicle itself, as the claims focus on external "emergency situations" rather than internal security breaches.
• —Does not cover general traffic monitoring without specifically identifying an "emergency situation" based on unexpected positioning or condition of objects.
• —Does not cover simply recording sensor data without evaluating it for an emergency and performing a vehicle operation.

The clever bit

The novelty lies in precisely defining how an autonomous vehicle's control system evaluates specific factors, like the unexpected positioning or atypical behavior of external objects, to identify an "emergency situation" and then automatically select and perform a corrective driving action. This moves beyond simple obstacle avoidance to contextual understanding of threats.

Why it matters

This patent is important for the safety and reliability of autonomous vehicles. By defining how self-driving cars identify and react to unexpected dangers, it helps build trust in the technology. This type of system is critical for ensuring that autonomous vehicles can operate safely alongside human drivers and pedestrians in complex real-world environments, reducing the risk of accidents.

Real-world examples

• 1.Waymo's autonomous vehicles
• 2.Cruise LLC's self-driving cars
• 3.Tesla's Full Self-Driving (FSD) system
• 4.Mobileye's advanced driver-assistance systems (ADAS)