A breakthrough in semiconductor technology has led to the development of an innovative ‘intelligent sensor’ that operates at high speeds with low power consumption, mimicking the visual neural system of insects. (Image courtesy of KAIST)
DAEJEON, Feb. 20 (Korea Bizwire) – A breakthrough in semiconductor technology has led to the development of an innovative ‘intelligent sensor’ that operates at high speeds with low power consumption, mimicking the visual neural system of insects.
The research team behind the discovery, led by professor Kim Kyung-min from the Department of Materials Science and Engineering at KAIST, successfully integrated various memristor devices to emulate the visual intelligence of insect optic nerves.
Memristors, a portmanteau of ‘memory’ and ‘resistor,’ are electronic components that alter their resistance state in response to input signals, making them ideal for tasks that require adaptability and memory retention.
In the age of advancing artificial intelligence (AI), vision systems utilizing AI play a crucial role in image recognition, object detection, and motion analysis. However, traditional vision systems rely on complex algorithms to process signals received from image sensors, recognizing objects and their movements.
This approach generates significant data traffic and consumes a substantial amount of power, posing challenges for implementation in mobile and Internet of Things (IoT) devices.
The KAIST research team overcame these limitations by integrating memristor devices with varying functionalities to directly replicate the optic nerve system of insects, known for efficiently processing visual information to detect and recognize objects and their movements.
By leveraging the developed motion recognition device, the team designed a neuromorphic computing system capable of predicting vehicle trajectories. The system demonstrated a significant reduction in energy consumption by 92.9% compared to existing technologies, enhancing its ability to accurately predict object movements.
The potential applications for this innovative technology are vast, ranging from autonomous vehicles and transportation systems to robotics and machine vision.
Professor Kim emphasized the significance of the research, stating, “Insects utilize simple visual intelligence to remarkably perceive object movements with agility. This study holds great importance as it demonstrates the capability to replicate neural functions using memristor devices.”
The findings of this groundbreaking research were recently published in the online edition of the international journal ‘Advanced Materials,’ showcasing the potential of memristor-based technologies to revolutionize the field of intelligent sensors and computing systems.
Kevin Lee (kevinlee@koreabizwire.com)
