Sony Files Patent for a Third-Sensor Method That Aligns Mixed Camera and Scanner Arrays
Getting two sensors to agree on where they are in space is one of the quiet hard problems in robotics and autonomous vehicles. Sony's new patent proposes a third sensor as a neutral referee to solve it.
What Sony's three-sensor calibration trick actually does
Today, when engineers bolt different types of sensors together on a robot or a car, they have to carefully hand-tune how those sensors understand their positions relative to each other. If a camera and a depth scanner don't agree on where they sit in space, the data they produce won't line up, and the system makes mistakes.
Sony's patent describes a way to fix that by bringing in a third sensor as a reference. That third sensor measures a known physical object from a position where both of the other sensors can also see it. The system then uses all three sets of readings together to calculate, automatically, exactly how each sensor is oriented relative to the others.
The upshot is that you don't need the two sensors to share the same sensing technology or even the same field of view. As long as they can both partially see the calibration object at the same time, the math works out. That makes it much easier to mix and match different sensor types without losing accuracy.
… calculates a relative positional relationship between a first sensor and a second sensor on a basis of third sensor data, first sensor data, and second sensor data …
Translation: It figures out how multiple cameras and scanners line up by combining data from all three sensors.
How the third sensor anchors the position math
The patent describes an information processing apparatus built around one core job: calculating the precise positional and rotational relationship between a first sensor and a second sensor.
The key ingredient is a third sensor that captures three-dimensional data about a physical calibration object, an object placed somewhere both the first and second sensors can at least partially see. The system then feeds all three data streams into a processing unit:
- Third sensor data: a 3D scan of the calibration object, giving the system a geometric ground truth
- First sensor data: that sensor's own reading of the same object or space
- Second sensor data: the other sensor's reading of the same object or space
By comparing how each sensor perceives the shared scene, the processing unit can calculate the relative positional relationship (meaning how far apart the sensors are, and at what angles they face each other) without requiring the sensors to use the same technology or cover the same range. The third sensor essentially provides a common coordinate anchor that bridges differences between the first two.
… enable calibration with high accuracy relating to relative positions/postures of a plurality of sensors that senses spatial information regardless of the type, sensing range, and the like of each sensor …
Translation: The system accurately calibrates mixed sensors regardless of their different types or ranges.
What this means for robots, cars, and mixed sensor rigs
In autonomous vehicles, industrial robots, and AR systems, mixing sensor types is common because no single sensor does everything well. A camera sees color and texture; a lidar scanner measures precise distances; a radar works in bad weather. Getting all of them to agree on a shared picture of the world depends entirely on accurate calibration.
Sony's approach, if it works as described, could reduce the manual setup time for these rigs and improve accuracy in systems where the sensors have very different fields of view or technologies. Sony's broader sensor patent activity has been building toward more automated sensor fusion for cameras and spatial computing hardware, which fits this filing neatly into that direction.
Sony's 23rd filing we've tracked since July in our self-driving sensing race watchlist follows one catching its own timing errors and one on signal processing.
Claim 1 covers any device that uses a third sensor's three-dimensional readings of a reference object to calculate the precise spatial relationship between two other sensors. The claim places no restrictions on what those sensors must be, what the reference object looks like, or which math produces the answer.
That breadth means any calibration system using a third sensor as a geometric bridge between two others would fall within Sony's reach, whether the product lives in a self-driving car, a robot, or a headset.
The practical consequence is significant: this filing targets the foundational step that makes every downstream sensing capability reliable. Before a car can see or a robot can navigate, someone has to solve exactly the problem Claim 1 describes, and this claim would give Sony a say in how that happens.
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
16 drawing sheets from US 2026/0259307 A1 · click any drawing to enlarge
Want this weekly breakdown for a company we don't cover? Patentlyze Pro →