Mems Reservoir Computing Using Frequency Modulated Accelerometer

We developed an intelligent MEMS sensor with data processing ability for data acquisition and conversion at low power consumption. It enables data processing at the sensing device for efficient distributed processing. We employed a MEMS reservoir which integrates sensing and computing functions in single device structure. In this research we designed a frequency modulated capacitive accelerometer and characterized its frequency response. The measured response was used for simulation of time series prediction and the machine learning task was successfully demonstrated.