Three scientists at the Queensland University of Technology released their findings on Thursday, outlining a method to create a camera that processes images in ways similar to the human brain.
While there are further developments to unlock, they say neuromorphic computing could deliver a robotic revolution.
The latest discovery, published in the Science Robotics journal, uses a camera and computer processor the QUT team called LENS, which stands for "locational encoding with neuromorphic systems".
The system is inspired by the way the human brain works, author and QUT neuroscientist Adam Hines said, to save more than 90 per cent of power compared to a traditional robotic navigation system.
"The brain is so energy-efficient, it only uses about 20 watts of power to do everything from keeping us alert and awake and talking to constantly navigating and predicting where we're going next," Dr Hines told AAP.
"Traditional AI systems like ChatGPT use significantly more power than that so taking inspiration from the brain is a really great way to save on energy."
The LENS camera sensor and processor work by registering changes, such as light and movement, rather than recording images the entire time it operates.
The QUT research team, which included Michael Milford and Dr Tobias Fisher, tested the system on an eight kilometre journey and could make it work using 180 kilobytes or up to 300 times less storage than a traditional system.
Saving so much energy and storage could let robots operate and navigate by themselves in new areas or for significantly longer durations, Dr Hines said.
"The real use cases in mobile robotics... are search and rescue, underwater monitoring of places like the Great Barrier Reef, or even really extremely remote areas like space explorations," he said.
Neuromorphic computing has been a target for previous research but QUT Centre for Robotics director Professor Milford said it was vital to translate theory into practical applications.
"Impactful robotics and tech means both pioneering groundbreaking research but also doing all the translational work to ensure it meets end user expectations and requirements," he said.
5°C