Researchers claim that the UCF-designed device can be made faster by combining the three processes. The technology is compact, with hundreds of devices fitting onto a chip measuring one inch in width.
Tania Roy (principal study investigator and assistant professor at UCF’s Department of Materials Science and Engineering and NanoScience Technology Center) says that “it will change how artificial intelligence is realized today.” “Today, everything runs on traditional hardware and comprises discrete components. We can perform in-sensor computing with one device and one platform.
This technology builds on previous research by the research team, which created brain-like devices that allow AI to operate in remote areas and spaces.
Roy says that while the devices behaved the same way as the brain’s synapses, they were still not feeding the image directly. We now have devices with image sensing capability that can sense, process, recognize and recognize images.
The device’s versatility will make it safer to drive in self-driving cars, even at night, according to Molla Manjurul Islam ’17MS. She is the study’s lead author and a doctoral student at UCF’s Department of Physics.
Islam said that if the vehicle’s imaging system operates at one wavelength (e.g. visible wavelength), it will not be able to see what is ahead of it at night. It can see the whole condition with our device.
He says that no device comparable to this one can simultaneously operate in the ultraviolet range, visible wavelength, and infrared wavelength. This is the best selling point of this device.”
This technology is key to engineering the nanoscale surfaces of molybdenum disulfide (and platinum ditelluride). This allows for multi-wavelength memory and senses. This collaboration was done in close cooperation with YeonWoongjung Jung, an assistant professor who holds joint appointments at UCF’s NanoScience Technology Center and the Department of Materials Science and Engineering.
Researchers tested the device’s accuracy by having it recognize mixed wavelength images — an ultraviolet number “3” and an infrared portion that is the mirror image for the digit. They showed that the technology could distinguish the patterns and identify them as “3” in ultraviolet and “8” (infrared).
Adithi Krishnaprasad, a doctoral student at UCF’s Department of Electrical and Computer Engineering, said, “We got 70-80% accuracy which means that they have very high chances of being realized in hardware.”