Many embedded structural monitoring applications require extremely low-power sensors and signal processing capabilities. In particular, the continuous monitoring of vibrations, impacts, and acoustic noise within a structure typically requires significant power for not only the transducers themselves, but also the signal conditioning, analog to digital conversion, and digital signal processing necessary to extract useful information from the captured waveforms [1,2,3,4]. This paper presents a

more »

... sor methodology that performs spectral decomposition of acoustic and vibration waveforms using arrays of micromechanical resonant structures, mechanical filters, and embedded active films, with signal processing being performed in the mechanical domain. The use of embedded active films in these devices results in sensors that can be, in some instances, self-powered by the vibrations being detected. A MEMS-based acoustic emission spectral sensor employing an embedded electret film has been fabricated and tested in an impact environment. Sensor fabrication was performed using in situ polarization of the active film after completing the entire fabrication flow. This sensor consisted of an array of tuned micromechanical resonant elements with natural frequencies varying across the spectrum of interest. The sensor was incorporated into a ball drop apparatus in which controlled acoustic pulses could be delivered to the structure under test. Spectral output from the sensor was collected and could be used to distinguish between impacts involving different material sets and impact energies without the use of digital signal processing and its associated power consumption.