Get the newsletter
Sony · Automotive · Filed Oct 31, 2025 · USPTO published Oct 1, 2026 · Oct 1 release

Sony Patents a System That Predicts How Dangerous Each Nearby Object Is to Your Car

Most car safety systems tell you when something is close. Sony's new patent describes a system that tells you how dangerous that thing is likely to be, including what would happen if a crash were already in motion.

A street scene with pedestrians and a car, showing visual indicators of predicted danger around objects. Drawing from patent filing US 2026/0296418 A1.
A street scene with pedestrians and a car, showing visual indicators of predicted danger around objects.All 17 drawings ↓
Publication number
US 2026/0296418 A1(US20260296418A1)
Applicant
SONY GROUP CORPORATION
Filed · Published
Oct 31, 2025 · Oct 1, 2026
Inventors
Seiji SUZUKI, Shinichi IRIYA, Katsuji MIYAZAWA, Satoshi AKAGAWA
Status when we published
Waiting for an examiner
Document
20 claims
US classification
340/425.5
Parent application
National Stage Entry of PCT/JP2024/015585 (filed Apr 19, 2024)(PCTJP2024015585)

What Sony's danger-prediction system actually does for drivers

You're driving through a busy intersection, and a cyclist to your left is moving unpredictably. Your car's sensors see the cyclist, but the real question is: should you be worried about that specific cyclist right now, or are they fine? Sony has filed a patent for a system that tries to answer that exact question.

The system does two things at once. First, it predicts where nearby objects (other cars, pedestrians, cyclists) are likely to go in the next few seconds. Second, it also models what would happen if a collision were already occurring, tracing the likely path an object would take during a crash. By comparing those two pictures, it calculates a danger score for each object around the vehicle.

That score then gets shown to the driver, so instead of a simple proximity alarm, you get something closer to a ranked threat list. The idea is to cut through the noise and surface the risks that actually matter before it's too late to react.

From the filing · CLAIM 1
a prediction unit which predicts movement information indicating a future movement of an object around a moving body and accident information indicating a movement of the object in a case where an accident occurs; …

Translation: The system forecasts both where nearby things are going and how they would move during a crash.

How the system models crashes that haven't happened yet

The patent describes three linked components working together inside a vehicle (or any moving object).

  • Prediction unit: This piece forecasts two separate things simultaneously: the normal expected movement of objects around the vehicle (where that pedestrian is probably heading) and a separate model of how those objects would move if a collision were already happening (think of it as a crash physics simulation running in parallel).
  • Calculation unit: It combines both forecasts into a single danger score for each nearby object. An object moving toward the vehicle on a normal path might score low, but one whose crash-trajectory overlaps heavily with the vehicle's path scores high.
  • Presentation control unit: This takes those scores and decides what to show the driver, presenting what the patent calls "danger prediction information." The format isn't fixed in the claim, leaving room for visual overlays, audio alerts, or other outputs.

The key distinction here is the dual-track prediction. Most proximity systems ask "how close is this thing?" This system asks "how bad would it be if things go wrong with this thing?" by essentially pre-simulating the accident and then weighing that against real movement data.

From the filing · THE ABSTRACT
… a calculation unit which calculates a degree of danger of the object on the basis of the movement information and the accident information; …

Translation: It figures out how threatening each surrounding object is by combining those movement and crash predictions.

What this means for driver-assist warnings in real traffic

Driver-assist systems today are often guilty of crying wolf: too many beeps, too many warnings, for things that pose no real threat. That alert fatigue is a documented problem, and it causes drivers to tune out the very systems designed to protect them. A danger-scoring approach that separates low-risk proximity from high-risk proximity could make warnings feel meaningful again, which is a bigger safety win than simply adding more sensors.

For Sony, this fits into its broader interest in automotive sensing and in-vehicle electronics. The system is described broadly enough to apply to any moving vehicle, not just passenger cars, which keeps the potential applications wide, covering buses, drones, or autonomous delivery vehicles. Whether it ever reaches production depends on far more than the patent itself.

Sony's 28th filing we've tracked since July in the self-driving sensing race adds to a run that includes one on keeping sensors aligned and one on picking warning sounds.

Source: US 2026/0296418 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 Sony 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 · 17 sheets

17 drawing sheets from US 2026/0296418 A1 · click any drawing to enlarge

Vehicle control system components., drawing sheet 1 from US 2026/0296418 A1
Sheet 1 Vehicle control system components.
Vehicle sensor detection ranges., drawing sheet 2 from US 2026/0296418 A1
Sheet 2 Vehicle sensor detection ranges.
Information processing unit for danger., drawing sheet 3 from US 2026/0296418 A1
Sheet 3 Information processing unit for danger.
Flowchart for danger prediction., drawing sheet 4 from US 2026/0296418 A1
Sheet 4 Flowchart for danger prediction.
Pedestrian danger visualization examples., drawing sheet 5 from US 2026/0296418 A1
Sheet 5 Pedestrian danger visualization examples.
Car danger visualization., drawing sheet 6 from US 2026/0296418 A1
Sheet 6 Car danger visualization.
Pedestrian danger visualization., drawing sheet 7 from US 2026/0296418 A1
Sheet 7 Pedestrian danger visualization.
Cyclist danger visualization., drawing sheet 8 from US 2026/0296418 A1
Sheet 8 Cyclist danger visualization.
Show the other 9 sheets
Sony Patents a System That Predicts How Dangerous Each Nearby Object Is to Your Car patent drawing, sheet 9
Sheet 9
Sony Patents a System That Predicts How Dangerous Each Nearby Object Is to Your Car patent drawing, sheet 10
Sheet 10
Sony Patents a System That Predicts How Dangerous Each Nearby Object Is to Your Car patent drawing, sheet 11
Sheet 11
Sony Patents a System That Predicts How Dangerous Each Nearby Object Is to Your Car patent drawing, sheet 12
Sheet 12
Sony Patents a System That Predicts How Dangerous Each Nearby Object Is to Your Car patent drawing, sheet 13
Sheet 13
Sony Patents a System That Predicts How Dangerous Each Nearby Object Is to Your Car patent drawing, sheet 14
Sheet 14
Sony Patents a System That Predicts How Dangerous Each Nearby Object Is to Your Car patent drawing, sheet 15
Sheet 15
Sony Patents a System That Predicts How Dangerous Each Nearby Object Is to Your Car patent drawing, sheet 16
Sheet 16
Sony Patents a System That Predicts How Dangerous Each Nearby Object Is to Your Car patent drawing, sheet 17
Sheet 17
Patent filing page

Watchlists this filing belongs to

All watchlists →
Reader comments

No comments yet

Start the discussion

Real name or a handle, either is fine. Comments are read by a person before they appear, so allow a little time. Keep it about the filing.