DYNAMIC SENSOR SELECTION FOR VISUAL INERTIAL ODOMETRY SYSTEMS

Invented by Borys; Olha, Halmetschlager-Funek; Georg, Kalkgruber; Matthias, Wolf; Daniel, Zillner; Jakob
Smart eyewear is changing the way we see the world, right in front of our eyes. But how does it know where you are, and how can it do all this without running out of battery? Let’s explore a patent that shows a new way for eyewear to track itself more smartly and efficiently.
Background and Market Context
The dream of blending digital content with the real world has pushed companies and inventors to create wearable technology that is lightweight, easy to use, and always aware of its place in space. Smart glasses and AR (augmented reality) eyewear are becoming more popular, with big tech names racing to lead this market. People want glasses that can help them find their way, give them information, or play games, all while being comfortable and lasting all day on a single charge.
But there is a big challenge. For these glasses to work, they need to know exactly where they are and how they are moving at every moment. This is called “tracking.” Most modern eyewear does this by using lots of sensors. Cameras take pictures of what’s around you. Inertial sensors feel when you move your head or walk. Sometimes, GPS or even radar is used. All these sensors together help the glasses know their position and orientation, so the digital images stay glued to the real world.
However, running all these sensors all the time uses a lot of power. If you want lightweight glasses, you can’t use huge batteries. If you use small batteries, you need to save power. If the sensors are turned off too much, the tracking gets worse. This is the balancing act at the heart of every smart eyewear device.
The market expects these glasses to be sleek, last long, and give a smooth experience. Users want to trust that the digital content will stay in place, even if they move quickly or go outside into a new setting. The solution must be clever enough to use only what’s needed and nothing more, so the battery lasts longer without losing the magic of AR.
Scientific Rationale and Prior Art
In the past, tracking for wearable devices has mostly relied on “visual-inertial odometry” systems. This simply means using both cameras (visual) and movement sensors (inertial) together to figure out where the device is. These systems use algorithms to combine what the camera sees with the movement data from sensors like accelerometers and gyroscopes. By doing this, the device can tell if you are turning your head, walking, or standing still.
Older solutions often kept all the sensors on at full power to make sure the tracking never failed. This worked well for accuracy, but it drained the battery fast. If the glasses had to last for many hours, this was a big problem. Some tried to use only one sensor at a time, but this made the tracking less reliable. If the camera was off, the system might not notice that you quickly turned your head. If the movement sensors were off, the system couldn’t tell if you had started walking.
Other systems tried to save power by reducing the quality of the images or lowering the number of times per second the sensors were checked. But this was not always smart. Sometimes, when the glasses were in a new place or moving quickly, they needed high-quality data to keep the tracking sharp. There was no good way to know when to turn things down and when to turn them up.
There were even some systems that used GPS or Wi-Fi to help tracking. But GPS is not always available indoors, and Wi-Fi signals are not always strong everywhere. So, these methods alone were not enough.
In short, previous approaches were either too simple—turning sensors on or off without thinking about what was happening around the user—or too wasteful—using all sensors all the time. The market needed something that could make smart decisions, adjusting itself in real time to use just the right amount of power without sacrificing the quality of the tracking. That’s where this new invention comes in.
Invention Description and Key Innovations
This patent introduces a smart way for eyewear to track itself by monitoring its sensors and adjusting them in real time, based on what’s happening around the user and how the glasses are moving. Let’s break down what makes it special, and how it works.
At the heart of the system is a group of sensors—cameras, movement sensors (IMU), maybe a GPS, and more. The glasses don’t just run all these at full blast. Instead, they have a brain (the processor) that watches the sensors and decides what needs to be on, how often they need to check, and at what quality.
Here’s how it works in a simple way:
The glasses are always checking their sensors. But instead of just collecting data, they are looking for clues about their current situation. Are you moving fast or slow? Are you inside or outside? Is this a place the glasses have seen before, or somewhere new? Has something in the environment changed—like new objects or different lighting?
The system creates a “status” of the tracking system, kind of like a mood ring. If everything is calm, you’re in a known place, and you’re not moving much, the system knows it can relax. It turns off some sensors, lowers the quality or sampling rate of others, and saves power. If you suddenly start moving quickly, or if the glasses see new things in the environment, the status changes. The system wakes up more sensors, increases their sampling rate, or raises their quality to make sure the tracking stays sharp.
Let’s say you’re sitting in your living room, not moving much. The glasses recognize this as a safe, low-motion environment. Maybe just one camera and the movement sensor (IMU) are on, running slowly. If you get up and walk outside into a busy street, the glasses see all the new things in the environment and sense your higher speed. The system quickly turns on extra sensors, increases the quality and rate, and makes sure nothing is missed.
This approach is very flexible. The glasses can decide to use only a few sensors, or all of them, depending on what’s needed. They can select different sampling rates for each sensor. They can even choose to reduce the image quality or frame rate to save power, or boost them if the situation demands it.
The system also uses what it sees with the cameras to identify the environment. If you walk into a room the glasses have seen before, they can use stored information to help with tracking. If you go somewhere new, the glasses react by using more sensors until they understand the new place.
This smart adjustment is not random. It is based on clear rules and mapping between “status” levels and sensor settings. For example, the system might have a table that says if motion is low and uncertainty is low, use low power. If either is high, bump up the power and turn on more sensors. The system even measures “uncertainty”—a way of checking how sure it is about its position. If it feels unsure, it increases the amount of data it collects to get back on track.
What does all this mean for the user? The glasses can last longer on a single charge, because they are not wasting power. At the same time, the tracking stays smooth and reliable, because the system is always matching the sensors to what is going on. The digital images and information stay glued to the real world, even if you move quickly or go somewhere new. The user doesn’t have to do anything—the system is fully automatic.
On top of all this, the invention covers the method (the way the glasses decide and adjust), the device (the glasses themselves), and the software (the program that makes it all work). This means the idea is protected in every form, from the hardware to the code that runs it.
The key innovations here are:
1. Real-time monitoring and smart adjustment of all sensors, not just on or off, but at different rates and qualities.
2. Using the environment and motion data to decide which sensors need to be on, and at what level.
3. Mapping “status” levels to sensor configurations, so the system is always using the right amount of power and data.
4. Making this work automatically, so the user doesn’t have to worry about settings or modes.
5. Applying the idea to the device, the method, and the software, so the whole system is protected.
Conclusion
This patent shows a leap forward for smart eyewear. Instead of always running at full speed or guessing when to save power, these glasses can think for themselves. They know when to relax and when to work hard, always matching their behavior to what’s happening around the user. This means longer battery life, better tracking, and a smoother AR experience. As smart glasses become more common, inventions like this will set the standard for how wearable technology should perform—smart, efficient, and always ready for what’s next.
Click here https://ppubs.uspto.gov/pubwebapp/ and search 20250216679.