Electrothermal Tunable MEMS Oscillators for MEMS-Based Reservoir Computing
Yuan-Chieh Lee, Liang-Kai Wang, Yu-Chi Chuang, Hao-Chiao Hong, Yi Chiu · IEEE Sensors Letters · 2024
Nonlinear microelectromechanical-system (MEMS) resonators have been demonstrated in physical reservoir computing (RC) with promising results. Most MEMS-based RC was implemented using amplitude modulation in a delay loop, resulting in complex systems. In this letter, a tunable oscillator was constructed based on a MEMS clamped–clamped beam resonator. The oscillation frequency was tuned by Joule heating that changed the structural stiffness via thermal stress. Linear and nonlinear tuning were demonstrated. A tuning sensitivity of 84 kHz/V was obtained in the linear tuning regime. When operated in the nonlinear tuning regime, the proposed tunable MEMS oscillator showed rich nonlinearity and can be applied to future frequency-modulation-based MEMS RC to simplify the system design.