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
This study introduces FeOx optomemory for in-sensor computing, enabling automatic state conversion for optical processing. This self-adaptive system achieves high accuracy in complex signal recognition.
Area of Science:
- Materials Science
- Edge Computing
- Optoelectronics
Background:
- In-sensor vision computing systems are limited by the need for electrical resets.
- Processing complex dynamic signals requires advanced memory capabilities.
Purpose of the Study:
- To develop an optomemory device for fully optical computing in edge systems.
- To enable self-adaptive state conversion for enhanced signal processing.
Main Methods:
- Fabrication of FeOx optomemory devices.
- Investigation of negative photoconductance memory (NPM) and positive photoconductance memory (PPM) effects.
- Demonstration of self-adaptive conversion between NPM and PPM states based on light dosage.
- Implementation of a self-adaptive reservoir computing (RC) system.
Main Results:
- FeOx optomemory exhibits automatic state conversion from NPM to PPM at a threshold light dosage (0.72 µJ/µm²).
- The device leverages oxygen vacancy dynamics and iron redox states for memory effects.
- The self-adaptive RC system achieved a recognition accuracy of 97.93% for dynamic feature encoding.
Conclusions:
- FeOx optomemory offers a novel in-material encoding mechanism for in-sensor edge computing.
- The self-adaptive optical computing capability overcomes limitations of traditional electronic resets.
- This technology paves the way for more efficient and capable edge AI systems.
