Evrim Yazgin
Cosmos science journalist
Robots, autonomous vehicles and other next-generation technologies might have a human touch thanks to a new device which processes information like a brain.
The device was designed by engineers at RMIT University in Melbourne, Australia. It is described in a paper published in the journal Advanced Materials Technology.
The tiny gadget is “neuromorphic”, meaning it mimics the structure and function of neurons in the human brain. Results published in the journal reveal that the RMIT design is able to detect hand movement and store memories without the need for an external computer.
A key focus of neuromorphic technology is developing artificial vision systems for machines, like autonomous vehicles.
“Neuromorphic vision systems are designed to use similar analogue processing to our brains, which can greatly reduce the amount of energy needed to perform complex visual tasks compared with digital technologies used today,” says corresponding author Sumeet Walia, director of RMIT’s Centre for Opto-electronic Materials and Sensors (COMAS).
The device is based on the compound molybdenum disulphide. The researchers showed how atomic-scale defects in the material could be used to capture light and process the information as electrical signals.
“This proof-of-concept device mimics the human eye’s ability to capture light and the brain’s ability to process that visual information, enabling it to sense a change in the environment instantly and make memories without the need for using huge amounts of data and energy,” Walia says.
Experiments showed that the device detected a waving hand’s movement. The “memory” of those movements was then stored within the structure of the device itself.
Previous work, including by the RMIT team, had only involved processing of still images, not movement.
The team says their research is a step toward autonomous devices being able to navigate potentially dangerous or unpredictable environments that are unsafe for humans.
“Neuromorphic vision in these applications, which is still many years away, could detect changes in a scene almost instantly, without the need to process lots of data, enabling a much faster response that could save lives,” Walia says.
“For robots working closely with humans in manufacturing or as a personal assistant, neuromorphic technology could enable more natural interactions by recognising and reacting to human behaviour with minimal delay,” says corresponding author Akram Al-Hourani, deputy director of COMAS.
The current proof-of-concept molybdenum disulphide device has just 1 detecting pixel. The work of the COMAS team will be to scale it up to a larger pixel array.
“While our system mimics certain aspects of the brain’s neural processing, particularly in vision, it’s still a simplified model,” Walia says. “We will optimise the devices to perform specific real-world applications with more complex vision tasks and further reduce power consumption.”
Walia adds that the neuromorphic system will work with traditional computer-based artificial vision systems, rather than a replacement.
“Conventional systems excel at many tasks, while our neuromorphic technology offers advantages for visual processing where energy efficiency and real-time operation are critical.”
