Imagine a sensor so smart it can detect movement just like a bug’s eye, but works super fast and uses barely any power.
This isn’t a page from a science fiction book—it’s real, and it’s been developed by a team of scientists at KAIST, led by Professor Kyung Min Kim.
Their creation is set to change the game in everything from self-driving cars to home security systems.
On February 19, the team announced they had built a brand-new type of motion detector.
This isn’t just any motion detector; it’s designed to work like the optic nerve in insects’ eyes. Insects, as it turns out, are really good at spotting movement quickly and efficiently, thanks to their unique visual system.
The team’s study, which details how they did it, has been published in the journal Advanced Materials.
Nowadays, we use AI (Artificial Intelligence) for a lot of things, including making cameras smarter. These AI cameras can recognize faces, spot objects, and even understand what they’re seeing. But they need a lot of power to work, and they have to process tons of data. This makes them a poor fit for gadgets that need to be light on power use, like smartphones or devices connected to the Internet of Things (IoT).
The scientists looked at nature for inspiration. They noticed how insects process visual information super efficiently through something called the elementary motion detector in their optic nerves. This lets them see movements incredibly fast. But trying to recreate this with regular computer parts is pretty tough and hasn’t been very practical—until now.
Professor Kim and his team came up with a motion detector that’s not just efficient but also works at blazing speeds. And it’s surprisingly simple, made up of just two types of special components called memristors and a resistor. These memristors are the real stars, handling tasks like delaying signals and integrating and triggering signals, all of which are fancy ways of saying they help the sensor figure out when something is moving.
To show off what their sensor could do, the team set it up to predict where a vehicle would move next. The results were impressive: their device used 92.9% less energy than current technologies and was even more accurate in predicting motion.
What does all this mean for us? Professor Kim believes their work could be a big deal for making AI devices better. For instance, self-driving cars could become safer and more reliable. Robots and security cameras could get a lot smarter without needing so much power. And since these sensors are so energy efficient, they could be used in all sorts of mobile and IoT devices without draining the battery.
“Insects use their simple but highly effective visual systems to detect motion super quickly,” Professor Kim explained. “Our research is exciting because we’ve been able to replicate these functions with a memristor device.”
As we look to the future, this technology could pop up in autonomous vehicles, transportation systems, robotics, and anywhere else smart motion detection is needed. It’s a big step forward in making our technology more efficient and effective, inspired by the natural world.
Source: KAIST.
