Real-Time Parameter Estimation for Adaptive Control of High-Q Micromachined Accelerometers

We are presenting the hardware and software components needed to estimate the operating parameters of a High-Q micromachined accelerometers offer superior sensitivity and resolution but pose challenges due to increased transients and parameter drift over time. This paper presents a real-time parameter estimation framework that enables adaptive control by continuously monitoring and adjusting system parameters. The developed embedded system employs a direct digital synthesizer for excitation and a digital synchronous demodulator for precise response measurement. Various system identification methods are evaluated, including rank regression, recursive rank regression, and state-space identification, to estimate critical parameters such as resonance frequency, quality factor, and sensitivity. Experimental validation is performed using a high-Q capacitive accelerometer under varying environmental conditions, demonstrating the robustness of the proposed approach. The results show that state-space identification provides the most accurate parameter estimation, ensuring reliable long-term operation through self-calibration and adaptive control. By integrating computationally efficient techniques on an embedded platform, this method enables real-time compensation for fabrication and environmental variations, improving the stability and performance of high-Q MEMS sensors.