Pooled development fund Strategic Elements (ASX: SOR) has revealed proof-of-concept work that highlights its printable neuromorphic technology’s potential for data processing and self-learning in soft robotics and other signal processing applications.
The printable neuromorphic hardware is being developed from the company’s Nanocube Memory Ink technology in the Nanoionics laboratory at the University of New South Wales (UNSW).
According to Strategic Elements, early stage results show that in the case of computer vision applications, the technology uses less power to operate than the human brain and is able to use multiple resistance states with the potential capacity to process multiple points of data.
While artificial neural networks are not uncommon, most synapse networks exist only as software. The UNSW researchers are in the early stage development of a neural network hardware designed to be printable (low cost), portable, ultra-low power, flexible and semi-transparent.
Strategic Elements said these features are ideally suited to robotics and computer vision applications.
“For example, the ability to place flexible neuromorphic hardware onto soft robotics in health or manufacturing sectors or devices requiring such low power that battery or energy harvesting technologies, like humidity, could potentially be used as a power source,” it stated.
According to the company, the new artificial synapse fabricated by the UNSW team has shown “significant advancement lowering power consumption, ability to continuously change conductance with voltage pulses (like a biological synapse), encouraging endurance and multilevel switching”.
These features are believed to make possible applications such as image processing and mart/intelligent sensors for e-skin and soft robotics.
Emulating the human brain
A memristor – an electronic memory device that mimics the information-transmitting synapses in the human brain to carry out complex computational tasks – was fabricated with Nanocube Ink and tested for endurance known as “long-term potentiation (learning) and depression (forgetting)”.
The artificial synapse was pulsed for 100,000 “spikes”, which mimic neurons firing in the brain, and no significant degradation was observed. The synapse also showed good potential endurance after repeated learning and forgetting cycles, and the ability to have multilevel switching of up to 10 resistance states per cell, Strategic Elements reported.
This is in contrast to less accurate or efficient memory devices that only use two resistance states (high and low).
Strategic Elements said it will assess a potential program of work between the computer vision and robotics team at its subsidiary Stealth Technologies and the UNSW team to develop a prototype application.
Future work will also be carried out to reduce the temperature required in the manufacturing process, fabricate on flexible substrates and increase the number of memristors to meet requirements of image recognition and tactile touch sensors in robotics.