Get the newsletter
Waymo (Google) · Automotive · Filed Apr 14, 2026 · USPTO published Aug 20, 2026 · Aug 20 release

Waymo Files Patent for Autonomous Cars That Navigate Around Sensor Blind Spots

Rain, fog, and glare don't just obscure your windshield view, they degrade the sensors a self-driving car depends on. Waymo's latest patent describes a system that maps exactly how much each sensor's vision is impaired and then changes how the car behaves accordingly.

Autonomous vehicle navigating a street intersection while mapping out its sensor coverage zones and blind spots around surrounding buildings. Drawing from patent filing US 2026/0243892 A1.
Autonomous vehicle navigating a street intersection while mapping out its sensor coverage zones and blind spots around surrounding buildings.All 15 drawings ↓
Publication number
US 2026/0243892 A1(US20260243892A1)
Applicant
Waymo LLC
Filed · Published
Apr 14, 2026 · Aug 20, 2026
Inventors
Dmitri A. Dolgov, Christopher Paul Urmson
Status when we published
Waiting for an examiner
Document
20 claims
US classification
701/28
Parent application
Continuation of 18/346,486 (filed Jul 3, 2023)(18346486)

What Waymo's weather-aware sensor model actually does

Imagine you're in a Waymo robotaxi and a heavy rainstorm rolls in. The car's cameras and lasers can't see as far or as clearly as they normally do, but how does the car know that, and what should it do about it?

This patent describes a system that builds a 3D picture of what each sensor on the vehicle should be able to see under ideal conditions. When weather comes into play, the system adjusts that picture to reflect how much the rain, fog, or snow is cutting into each sensor's range. The car then combines all those adjusted sensor pictures into one big map of what's knowable and what's hidden. From there, it changes how it drives, slowing down, keeping more distance, or being more cautious around areas it simply can't see into well.

The key idea is that the car isn't just reacting to what it does see. It's also reasoning about what it can't see and making decisions based on that uncertainty.

From the filing · CLAIM 1
… generating, by the computer, a model of a field of view of at least one sensor coupled to the vehicle, wherein the at least one sensor is configured to detect objects in the environment of the vehicle …

Translation: The car creates a digital map showing what its sensors are supposed to see.

How the 3D field-of-view model accounts for rain and fog

The patent describes a multi-step process for building a real-time model of a self-driving vehicle's perceptual limits.

  • Per-sensor 3D field-of-view models: For every sensor on the car (lidar, cameras, radar, etc.), the system generates an individual 3D model of what that sensor could detect if nothing were blocking it. Think of it as the sensor's theoretical "cone of vision."
  • Weather adjustment: The system receives information about current weather conditions and uses it to shrink or distort each sensor's model. Heavy rain scatters lidar pulses; dense fog cuts camera range. The system applies those degradation factors mathematically to each model.
  • Aggregation into a comprehensive model: All the adjusted per-sensor models are merged into one master 3D model of the vehicle's collective awareness, accounting for overlapping sensor coverage.
  • Map integration: The comprehensive model is then combined with detailed map data that includes probability scores for detecting objects at specific locations, for example, a busy pedestrian crossing that statistically has a high chance of having someone in it.

The final output is a model of the vehicle's environment that tells the car not just what it currently sees, but how confident it can be about any given area around it. The vehicle's driving behavior is then modified based on that confidence map, becoming more conservative in zones where sensor coverage is poor.

From the filing · THE ABSTRACT
This view need not include what objects or features the vehicle is actually seeing, but rather those areas that the vehicle is able to observe using its sensors if the sensors were completely un-occluded.

Translation: Instead of tracking what is currently visible, the system maps areas that should be visible without blockages.

What this means for self-driving safety in bad weather

Bad weather is one of the hardest unsolved problems in autonomous driving. Most sensor systems can tell you what objects they detect; far fewer can tell you how much the weather has degraded their ability to detect objects in the first place. A car that doesn't know its sensors are impaired can't compensate for that impairment, which is precisely the scenario that leads to dangerous outcomes.

This patent tries to close that gap by making sensor limitation a first-class input into driving decisions, not an afterthought. For anyone riding in or near an autonomous vehicle on a rainy highway, that distinction matters. The approach is also notable for tying probabilistic map data into the model, so the car can weigh its degraded sensor coverage against prior knowledge about where people or other cars are likely to be. Autonomous vehicle safety is one of the more closely watched areas of Big Tech patent news, and this filing reflects how much engineering effort goes into the edge cases most human drivers handle instinctively.

Google's 49th filing we've tracked since May in our self-driving sensor competition follows work on backup braking systems and virtual street-view angles.

Source: US 2026/0243892 A1 at the USPTO (PDF). Full patent text and figures are in the official publication.

Get our take in your Top Stories

Liked this breakdown? Add Patentlyze as a preferred source on Google, and our plain-English take shows up more often in your Top Stories the next time Google patent news breaks.

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 sheets

15 drawing sheets from US 2026/0243892 A1 · click any drawing to enlarge

Block diagram of autonomous vehicle computer systems, drawing sheet 1 from US 2026/0243892 A1
Sheet 1 Block diagram of autonomous vehicle computer systems
Interior view of autonomous vehicle dashboard, drawing sheet 2 from US 2026/0243892 A1
Sheet 2 Interior view of autonomous vehicle dashboard
Side view of autonomous vehicle sensors, drawing sheet 3 from US 2026/0243892 A1
Sheet 3 Side view of autonomous vehicle sensors
Overhead view of vehicle sensor ranges, drawing sheet 4 from US 2026/0243892 A1
Sheet 4 Overhead view of vehicle sensor ranges
Front and rear sensor range diagrams, drawing sheet 5 from US 2026/0243892 A1
Sheet 5 Front and rear sensor range diagrams
Side-facing sensor range diagrams, drawing sheet 6 from US 2026/0243892 A1
Sheet 6 Side-facing sensor range diagrams
Overhead map of a street intersection, drawing sheet 7 from US 2026/0243892 A1
Sheet 7 Overhead map of a street intersection
Overhead map with a building obstruction, drawing sheet 8 from US 2026/0243892 A1
Sheet 8 Overhead map with a building obstruction
Show the other 7 sheets
Waymo Files Patent for Autonomous Cars That Navigate Around Sensor Blind Spots patent drawing, sheet 9
Sheet 9
Waymo Files Patent for Autonomous Cars That Navigate Around Sensor Blind Spots patent drawing, sheet 10
Sheet 10
Waymo Files Patent for Autonomous Cars That Navigate Around Sensor Blind Spots patent drawing, sheet 11
Sheet 11
Waymo Files Patent for Autonomous Cars That Navigate Around Sensor Blind Spots patent drawing, sheet 12
Sheet 12
Waymo Files Patent for Autonomous Cars That Navigate Around Sensor Blind Spots patent drawing, sheet 13
Sheet 13
Waymo Files Patent for Autonomous Cars That Navigate Around Sensor Blind Spots patent drawing, sheet 14
Sheet 14
Waymo Files Patent for Autonomous Cars That Navigate Around Sensor Blind Spots patent drawing, sheet 15
Sheet 15
Patent filing page

Watchlists this filing belongs to

All watchlists →