Related Experiment Video
Updated: Aug 30, 2025
08:07
Assembly and Characterization of Biomolecular Memristors Consisting of Ion Channel-doped Lipid Membranes
Published on: March 9, 2019
7.9K
Diffusive Memristors with Uniform and Tunable Relaxation Time for Spike Generation in Event-Based Pattern Recognition
Fan Ye1, Fatemeh Kiani1, Yi Huang1
1Department of Electrical and Computer Engineering, University of Massachusetts Amherst, Amherst, MA, 01003, USA.
Advanced Materials (Deerfield Beach, Fla.)
|August 29, 2022
Summary
Researchers improved diffusive memristor uniformity for brain-inspired computing. This advancement enables more reliable neuromorphic hardware, enhancing performance in tasks like object recognition with reduced power needs.
Area of Science:
- Neuromorphic Engineering
- Materials Science
Background:
- Diffusive memristors show promise for brain-inspired computing.
- Device relaxation dynamics randomness hinders large-scale adoption.
Purpose of the Study:
- Engineer diffusive memristors for improved relaxation time uniformity.
- Tune memristor relaxation times over three orders of magnitude.
- Implement the Hierarchy of Time Surfaces (HOTS) algorithm for spike generation.
Main Methods:
- Engineered device stack for enhanced uniformity (σ reduced from ≈12 to ≈0.32 ms).
- Connected memristors with resistors or capacitors to tune relaxation time (1.13 µs to 1.25 ms).
- Implemented the HOTS algorithm for spike generation.
Main Results:
- Achieved significantly improved uniformity in memristor relaxation time.
- Demonstrated tunable relaxation times spanning three orders of magnitude.
- Attained 77.3% accuracy in recognizing moving objects on the N-MNIST dataset.
Conclusions:
- Engineered diffusive memristors offer a pathway to reliable neuromorphic hardware.
- Tunable and uniform relaxation behavior is key for spike generation.
- This work enables ultralow power neuromorphic computing systems.
