Compressed Sensing of Sparse Spectrum Using Distributed Sound-to-Light Conversion Device Blinkies

In this study, we propose a method for estimating a sparse spectrum of sound using multiple sound-to-light conversion devices called Blinkies. A Blinky is a compact device that converts sound information into the light with a programmable operation. By distributing multiple Blinkies and monitoring them with a video camera, we can obtain acoustic information from a wider area without wired or wireless communication. However, due to the limited frame rate of the video camera, the bandwidth of the observed video signals is significantly narrower than that of the original sound. To restore the spectrum of an original sound signal, we design the entire signal acquisition process with multiple Blinkies and a video camera to be regarded as compressed sensing. We show that a sparse sound spectrum can be restored with the proposed method by numerical experiments.