Google Patent Uses Laser Reflections to Help Self-Driving Cars Detect Weather
When a self-driving car rolls over a patch of black ice, it needs to know before it starts sliding. Waymo has filed a patent for a system that uses the laser sensors already mounted on its vehicles to detect wet or icy road surfaces without any extra hardware.
How Waymo's lidar spots ice, rain, and wet pavement
Ever wondered how a self-driving car knows to slow down when roads get slippery? Right now, most autonomous vehicles rely on maps, weather forecasts, or cameras, none of which can tell you exactly how the pavement beneath your wheels feels at that moment.
Waymo's approach is to repurpose the lidar sensors the car already uses to "see" the world around it. Lidar works by firing laser pulses and measuring what bounces back. A dry road reflects those pulses differently than a wet or icy one does, and this system watches for those differences in real time.
When the laser return values stop matching what a clean, dry surface should look like, the system flags a weather-related condition and can adjust how the car drives. It also picks up on floating laser points that don't belong to any recognized object, like water droplets in the air, as a secondary clue that conditions have changed.
… comparing, by the computing device, the determined intensity values to expected intensity values for the surface; based on the comparison, identifying a weather-related condition of the surface based on a difference between the determined intensity values and the expected intensity values …
Translation: The car spots bad weather by checking if laser reflections from the road match what is normally expected.
How the system compares laser bounces to a dry-road baseline
The patent describes a two-pronged detection method, both relying on the lidar unit already standard on Waymo's vehicles.
The first approach focuses on the road surface itself. The system continuously measures the intensity values of laser pulses that bounce back from the ground (intensity here means how much laser light returns compared to how much was sent). It then compares those readings to stored expected values for a normal, dry surface. A significant difference signals something has changed, like water, ice, or snow covering the road, and the vehicle adjusts its behavior accordingly.
The second approach looks at laser points that don't attach to any recognized object in the scene. In standard lidar processing, returns are sorted and grouped into known things like cars, pedestrians, and curbs. Points left over, called untracked objects in the patent, may represent rain, fog, or airborne debris. A cluster of such stray returns is treated as evidence of a weather condition.
- Lidar intensity drops or spikes on road surface detected
- Comparison made against a baseline for dry conditions
- Leftover unassociated laser returns flagged as potential precipitation or fog
- Vehicle control adjusted based on the identified condition
… determining given laser data points of the plurality of laser data points that are unassociated with the one or more objects in the environment as being representative of an untracked object.
Translation: Random laser hits that do not match any solid objects are treated as potential rain or snow.
What real-time road-surface sensing means for self-driving safety
Bad weather is one of the hardest problems for self-driving vehicles. Cameras wash out in heavy rain, GPS tells you nothing about ice, and a human driver's instinct to ease off the accelerator on a slick surface is genuinely difficult to replicate in software. A system that uses sensors the car already carries, without requiring new hardware, is a practical answer to a real and expensive safety gap.
For passengers, this could mean a Waymo robotaxi that responds to a sudden rain shower or a frost patch the way a cautious driver would, automatically and without waiting for a map update. Waymo's run of weather-sensing filings suggests the company is treating adverse conditions as a core solvable problem, not an edge case to be avoided.
Google's 53rd filing we've tracked since May in our self-driving sensing watchlist follows predicted driver reaction times and LiDAR-synced low-light cameras.
The problem this patent attacks is not minor. Weather-related crashes account for hundreds of thousands of injuries in the United States every year, and slippery roads are a leading factor. For a self-driving company, failing to handle a wet road isn't an edge case, it's a liability that determines whether the technology can ever work outside of sunny, well-mapped cities.
What makes this filing credible is the pragmatism of the approach. Waymo isn't proposing new sensors or external data feeds. It's asking whether the laser hardware already on the roof can do double duty, and the physics actually support that: ice and water really do change how laser light reflects. The baseline-comparison method is the kind of thing you'd expect a careful engineering team to reach for.
The honest caveat is that lidar intensity can vary for lots of reasons beyond weather, road paint, debris, sensor aging, and the patent would need to handle all of those false positives to be genuinely useful. That's the hard part, and this document doesn't fully resolve it. Still, the problem is real enough and costly enough that even a partial solution earns serious attention.
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
7 drawing sheets from US 2026/0259347 A1 · click any drawing to enlarge
Want this weekly breakdown for a company we don't cover? Patentlyze Pro →