Luna showcases the first software enabling any machine to learn like humans and animals do. IntuiCell
A European AI startup has introduced Luna, a robot dog equipped with a digital nervous system that mimics human and animal learning.
Developed by Swedish firm IntuiCell, Luna represents one of the first applications of physical agentic AI, allowing it to make decisions, adapt, and pursue goals autonomously, much like a real dog.
IntuiCell plans to hire a dog trainer to teach Luna to walk, mimicking how neurons process information, rather than relying on a generative AI model and large datasets, the company told an international news agency.
“We aim to give the world what AI’s always promised, but hasn’t come anywhere close to delivering. Systems that can think and learn the way we do. Machines that can genuinely make sense of any environment. That can actually understand what they’re doing. That learn by natural instinct, the way humans and animals do, instead of being fed with synthetic data,” reads the firm’s website.
Neuron-inspired robotics
The most profound intelligence is not artificial but has evolved through constant interaction with the natural world. Living creatures learn by processing endless streams of sensory inputs, making every experience a lesson in intelligence.
According to the firm, now, 600 million years after the emergence of biological intelligence, this concept has been translated into software. The result is the first-ever digital nervous system that learns autonomously, like humans and animals, by engaging with its environment.
Unlike traditional AI, it requires no pre-training or extensive simulations. IntuiCell, a 2020 spin-out from Lund University in Sweden is applying a learning approach that mimics human and animal behavior in robots. This marks a shift in robotics education, focusing on real-time learning rather than pre-programmed instructions, according to Reuters.
IntuiCell plans to hire a dog trainer to teach its robot dog, Luna, to walk using a method inspired by how neurons interact and process information.
The company’s innovative software enables machines to learn independently, eliminating the need for pre-training, offline simulations, or massive data centers. This digital nervous system allows robots to adapt and respond to their surroundings, offering significant advancements in the development of human-like machines.
AI explores frontiers
Beyond robotic dogs, IntuiCell’s technology has broader applications in humanoid and autonomous robotics. The adaptive capabilities of these machines make them particularly suitable for unpredictable environments, including space exploration, deep-sea missions, and disaster response scenarios, reports Reuters.
Intelligent machines equipped with this learning software could operate in remote locations, adjusting to unforeseen challenges without human intervention.
RECOMMENDED ARTICLES
A potential future application involves deploying robots on Mars to build human habitats. Without relying on pre-existing data, these machines would adapt and solve problems in real-time, responding directly to the unfamiliar Martian landscape.
According to the firm, the capability opens new possibilities for advancing space exploration and autonomous construction in extreme conditions.
“This is our first showcase of an off-the-shelf quadruped robot learning in real-time, in the real-world, from scratch. Utilizing the first ever digital nervous system, developed by IntuiCell. What Luna is able to do is just the beginning. The way machines learn just changed for ever,” said the firm in a LinkedIn post.
0COMMENT
ABOUT THE EDITOR
Jijo Malayil Jijo is an automotive and business journalist based in India. Armed with a BA in History (Honors) from St. Stephen's College, Delhi University, and a PG diploma in Journalism from the Indian Institute of Mass Communication, Delhi, he has worked for news agencies, national newspapers, and automotive magazines. In his spare time, he likes to go off-roading, engage in political discourse, travel, and teach languages.
