MEMS reservoir computing using resonant accelerometer

Reservoir computing, which can utilize nonlinear physical dynamics as a computational resource, is expected to be one of the post-Moore computing technologies. In this research, we fabricated a resonant accelerometer with nonlinear MEMS resonators, and evaluated its training ability for acceleration signals such as human behavior.

Conference Presentation

Takahirio Mizumoto, Yoshikazu Hirai, Amit Banerjee, and Toshiyuki Tsuchiya
MEMS Reservoir Computing using Frequency Modulated Accelerometer
The 35th IEEE International Conference on Micro Electro Mechanical Systems (MEMS 2022), Tokyo, (January 9-13, 2022), pp. 487-490.
DOI: 10.1109/MEMS51670.2022.9699777

Takahiro Mizumoto, Amit Banerjee, Jun Hirotani, Toshiyuki Tsuchiya
Neuroaccelerometer using reservoir computing with parametric nonlinear response
The 48th International Conference on Micro & Nano Engineering (MNE) and Eurosensors 2022, Leuven, Belgium (Sep. 19-23, 2022), T3-P1-TuA-047.