New Google Patents · Filed Feb 23, 2026 · Published Jul 2, 2026 · verified — real USPTO data

Waymo Patents a Way for Remote Operators to Inch Robotaxis Past Blocked Views

When a self-driving car can't see around a corner, it doesn't have to just sit there anymore. Waymo's newest patent describes a system where a human operator, watching from a control center, can authorize the car to nudge forward a few inches at a time until the cameras get a clear view.

Waymo Patent: Remote Operator Controls Blocked Sensor View — figure from US 2026/0186482 A1
FIG. 1A — rendered from the official USPTO publication PDF.
Publication number US 2026/0186482 A1
Applicant Waymo LLC
Filing date Feb 23, 2026
Publication date Jul 2, 2026
Inventors Ioan-Alexandru Sucan, Collin Winter, Chien-Yu Ko, Vishay Nihalani
CPC classification 701/2
Grant likelihood Medium
Examiner CENTRAL, DOCKET (Art Unit OPAP)
Status Docketed New Case - Ready for Examination (Mar 27, 2026)
Parent application is a Continuation of 18342187 (filed 2023-06-27)
Document 20 claims

What Waymo's blocked-sensor workaround actually does

Imagine you're in a Waymo robotaxi and it stops at an intersection where a delivery truck is blocking the view. The car can't see oncoming traffic, so it refuses to move. Right now, that can mean a long, awkward wait.

This patent describes a fix: the car sends a help request to a remote operator, who sees a live dashboard of what the car's sensors can detect. The operator can then press a button authorizing the car to creep forward a short, predefined distance, just enough to peek around the obstruction.

Here's the important part: the car doesn't blindly follow that instruction. Even after the operator gives the go-ahead, the car keeps watching its surroundings and will brake on its own the moment it detects anything unexpected, like a pedestrian stepping off the curb. The human gives permission; the car still does the safety work.

How the remote creep-forward instruction loop works

The patent describes a two-part system: an onboard decision layer inside the car, and a remote operator interface.

On the car's side, the system maintains a predetermined list of navigation situations that trigger a help request. A partially occluded sensor view (meaning the cameras or lidar can't see clearly because something is physically in the way) is one of those triggers. When the car identifies a match, it packages up its current sensor data and transmits it to a remote operator.

On the operator's side, a graphical user interface (GUI) shows the car's current state, including what the sensors can and cannot see. A selectable button lets the operator authorize a creep-forward maneuver, moving the car a small, fixed distance at a speed below a set threshold.

Once that instruction reaches the car, the vehicle executes it while autonomously monitoring for environmental changes (new objects, people, or vehicles entering the scene). If the car detects any change that warrants stopping, it halts before completing the full authorized distance. The key design choice is that the human approves the move, but the car retains full authority to abort it.

What this means for robotaxi reliability in real cities

Blocked sensor views are one of the most common reasons self-driving cars stall in dense urban environments. Trucks double-parked, construction scaffolding, overgrown hedges at intersections: all of these create situations where a fully autonomous system has no safe way to proceed without more information. A remote-assist system that lets a human authorize a controlled micro-movement could meaningfully cut the number of times a robotaxi needs a physical rescue vehicle.

For Waymo, which is actively expanding its commercial service in cities like San Francisco and Phoenix, the practical ceiling on scale is partly determined by how often cars need hands-on human help. A patent like this is less about flashy autonomy and more about the operational plumbing that makes a large robotaxi fleet actually run.

Editorial take

This is a genuinely practical piece of engineering that addresses a real, recurring problem in robotaxi operations. It's not autonomous driving getting more autonomous, it's autonomous driving getting better at knowing when to ask for help and how to act on that help safely. That distinction matters, and Waymo is one of the few companies with enough real-world fleet data to know exactly which situations need this kind of fix.

Which company should we read for you?

We track 17 companies here. Pro is the same weekly breakdown for any company you choose, delivered privately. Type a name and we'll scope it and send you a quote.

Get one Big Tech patent every Sunday

Plain English, intelligent commentary, no hype. Free.

Source. Full patent text and figures from the official USPTO publication PDF.

Editorial commentary on a publicly published patent application. Not legal advice.