Rice-sized AI chip mimics brain to extend small drone range, boost capability
AI requires a lot of processing power, and small, battery-powered drones have limited energy.
One day, small drones could be capable of self-piloting, object recognition, and making complex decisions.
But current AI technology drains their battery, slashing flight time drastically. AI requires a lot of processing power, and small, battery-powered drones have limited energy.
Researchers from Texas A&M University are developing a new solution: neuromorphic computing, or “neuron-like nano-devices.”
Inspired by the human brain, these systems mimic biological neurons, processing information efficiently.
These neurons communicate, learn, and make decisions, activating only when needed, reducing energy consumption.
Researchers are working to create a tiny, rice-sized AI chip for small drones.
Chip with artificial neurons
Drones require energy for various functions like flying, navigating, sensing, staying stable, and communicating.
Large drones have powerful engines that provide enough electricity to handle demanding AI computations. However, smaller, battery-powered drones face energy limitations.
Adding current AI chips to these drones drastically reduces their flight time. For instance, a flight normally lasting 46 minutes could be shortened to four minutes due to battery drainage.
A team led by Dr. Suin Yi, an assistant professor in the Department of Electrical and Computer Engineering, is working on the development of energy-efficient AI for drones. Researchers are exploring the potential of neuromorphic computing.
A key component of this approach involves developing artificial neurons by conducting thin polymer films.
The team states that these films can mimic biological neurons’ electrical signaling and information processing.
Moreover, researchers strive to reproduce the core functionalities of learning and decision-making within these artificial systems. This involves designing the conducting polymer thin films to operate to mirror the brain’s energy efficiency.
Specifically, the goal is to create artificial neurons that only activate when necessary, processing and transmitting information on demand rather than constantly consuming power.
Yi added,
Through this proposed project, unmanned aerial vehicles can become more intelligent with the integration of conducting polymer materials systems, material that is both mechanically flexible and can spike and oscillate like neurons in the brain,
“Eventually, together with artificial synapses, these flexible artificial neurons could comprise a complete neuromorphic computing system that can realize intelligent small drones.”
Enhancing drone capabilities
If all goes as planned, then this next-gen, tiny AI chip would enable the drones to perform complex tasks like making decisions, identifying objects, navigating autonomously, and recognizing surroundings. It will function based on the drone’s battery capabilities.
Yi explained in the press release,
Small drones don’t have an engine, so their energy budget is very small.
“That’s why it’s very important to come up with a game changer such as neuromorphic computing systems beyond digital computers that can allow battery-powered drones to keep the same fly-time with AI as those without,”
This research aims to expand the capabilities of small drones. Energy-efficient, AI-capable drones could be used for various complex tasks in diverse fields like surveillance, rescue operations, and environmental studies.
READ the latest news shaping the AI Chips market at AI Chips News
Rice-sized AI chip mimics brain to extend small drone range, boost capability, source
Add comment