Prototyping

Dr. McKenna led the proof-of-concept technical solution development of a ground-breaking capability that uses short, recorded audio segments from a typical mobile device, such as a cellphone, to assess the health of machinery, in this case, US Military Tactical Quiet Generators. The approach, which was a mix of signal processing and deep learning (convolutional neural nets), yielded very encouraging results (e.g., 5-way classification accuracies near 80%) and paved the way for additional internal investment that led to, among other things, a patent award.