Related Experiment Video
Updated: Aug 5, 2026
06:12
Recording Horizontal Saccade Performances Accurately in Neurological Patients Using Electro-oculogram
Published on: March 13, 2018
In-Material Self-Adaptive Dynamic Encoding in FeOx Optomemory Device for Real-Time Analog Signal Processing
Jie Li1, Guangdong Zhou1, Haifeng Ling2
1College of Artificial Intelligence, Chongqing Key Laboratory of Brain-Inspired Computing and Intelligent Chips, Key Laboratory of Luminescence Analysis and Molecular Sensors (Ministry of Education), Southwest University, Chongqing, China.
Advanced Materials (Deerfield Beach, Fla.)
|August 4, 2026
Summary
We developed a novel FeOₓ optomemory for in-sensor computing. This device self-adapts between negative and positive photoconductance memory states, enabling efficient optical processing for edge computing applications.
Area of Science:
- Materials Science
- Edge Computing
- Optoelectronics
Background:
- In-sensor vision computing systems are emerging edge computing platforms for dynamic information processing.
- Current systems require electrical resets, limiting their ability to handle complex signals.
- There is a need for self-adaptive memory devices for fully optical computing.
Purpose of the Study:
- To propose and demonstrate a FeOₓ optomemory device with automatic state conversion.
- To enable fully optical computing by eliminating the need for electrical resets.
- To develop a self-adaptive reservoir computing system for enhanced signal processing.
Main Methods:
- Fabrication of FeOₓ optomemory devices.
- Investigation of the photoconductance memory effects (NPM and PPM) under varying light intensities.
- Analysis of the underlying physical mechanisms, including oxygen vacancy dynamics and charge carrier interactions.
- Implementation of the optomemory in a self-adaptive reservoir computing system.
Main Results:
- The FeOₓ optomemory successfully demonstrated automatic conversion from negative photoconductance memory (NPM) to positive photoconductance memory (PPM) at a threshold illumination of 0.72 µJ/µm².
- The state conversion mechanism involves the modulation of neutral and charged oxygen vacancies through photo-induced and Joule-heating assisted processes.
- The self-adaptive optomemory enabled fully optical computing, achieving 97.93% recognition accuracy in a reservoir computing system.
Conclusions:
- The FeOₓ optomemory offers a self-adaptive in-material encoding mechanism for in-sensor edge computing.
- This technology overcomes the limitations of electrical resets, paving the way for more efficient optical computing systems.
- The developed device shows significant potential for advanced dynamic information processing and artificial intelligence applications at the edge.
