Waymo Patents a System That Plans Gear Shifts Before the Moment Arrives
A self-driving car that shifts gears at the wrong moment can lurch, hesitate, or slow down unpredictably. Waymo's new patent describes a system that looks ahead on the road and decides to shift gears before the tricky moment arrives, not during it.
How Waymo's self-driving cars pre-plan gear changes
Ever driven with someone who always waits too long to downshift before a hill? The car bogs down, the engine strains, and the whole thing feels clumsy. Autonomous vehicles face the same problem, and Waymo is filing a patent to address it.
The system reads sensor data to spot upcoming situations, think sharp turns, merge zones, or steep grades, and estimates how fast or how hard the car will need to accelerate through them. It then decides whether to tweak the car's speed or shift gears before reaching that point, rather than reacting in the moment.
The goal is a ride that feels composed and deliberate. Instead of your robotaxi hesitating mid-turn because the transmission is hunting for the right gear, the car has already sorted itself out by the time you get there.
… determining, by the computing system and based on the estimated future acceleration or speed, that adjusting a current acceleration or speed of the vehicle in advance of the predefined situation avoids a gear shift during navigation of the predefined situation …
Translation: The car tweaks its speed early to prevent unnecessary shifting.
How the system estimates speed needs before a situation
The patent describes a system built around what engineers call predictive gear control, using future trajectory data to make transmission decisions in advance.
Here is how it flows:
- The vehicle's sensors continuously map the environment, detecting upcoming road features that qualify as predefined situations (specific conditions the software is trained to recognize, like a tight corner, a freeway on-ramp, or a steep downgrade).
- The computing system estimates the future acceleration or speed the vehicle will need to handle that situation correctly.
- It compares that future demand against the car's current acceleration and determines whether a gear shift during the maneuver can be avoided by adjusting speed before arriving.
- If yes, it adjusts acceleration or speed proactively, so the car enters the situation already in the right gear at the right speed.
The key insight is that a gear shift mid-maneuver (mid-corner, mid-merge) is mechanically and dynamically disruptive. By doing the math early and modulating approach speed, the system can make the gear change happen on a straight, stable stretch of road instead.
… determine, based on the sensor data, the vehicle is approaching a predefined situation positioned along the path.
Translation: Sensors spot upcoming road events that require planning.
What smoother shifting means for Waymo riders
For anyone riding in a Waymo vehicle, this is about the feeling of confidence in the car. A robotaxi that lurches or hesitates mid-turn erodes trust quickly, even if nothing goes wrong mechanically. Pre-emptive shifting means the car handles upcoming road features the way an experienced human driver would: before the tricky moment, not scrambling through it.
There is also a mechanical argument. Frequent reactive gear changes under load wear transmissions faster. A system that reduces those mid-maneuver shifts could lower maintenance costs for a fleet operator running hundreds of vehicles. Waymo's interest in fine-grained vehicle control shows up across multiple technical layers, and this one sits at a level most passengers would never consciously notice but would absolutely feel.
Google's application is the 56th we've tracked since May in our self-driving sensor competition, following work like which LIDAR data to discard and predicting collisions from object paths.
Anyone who has ridden in a car that hesitates, lurches, or hunts for the right gear through a turn or up a hill knows the specific discomfort this patent addresses. The system described here lets the vehicle read the road ahead and select the right gear before the demanding moment arrives, rather than scrambling to catch up once it has.
For a paying passenger, that difference shows up as a smoother, quieter ride through situations that currently tend to feel uncertain or mechanical. They would notice it most on hills, curves, and stop-and-go streets, which happen to be the moments when trust in the vehicle is most fragile.
The prediction this system depends on is only as reliable as the vehicle's understanding of what lies ahead, so imperfect maps or unexpected road conditions could still produce the rough behavior it aims to prevent. That is the real limit on what passengers will feel, and it matters more than the gear logic itself.
There are more where this came from
We read every patent application Big Tech publishes and send you the ones worth knowing. Plain English, free, every week.
The drawings
15 drawing sheets from US 2026/0274273 A1 · click any drawing to enlarge
Want this weekly breakdown for a company we don't cover? Patentlyze Pro →
Be the first to weigh in