mmPhone: Sound Recovery Using Millimeter-Wave Radios With Adaptive Fusion Enhanced Vibration Sensing

Although electret condenser microphone (ECM) has been widely used in sound acquisition, sound capture techniques still attract growing interests. To this end, the emerging techniques of laser microphone and visual microphone have been developed, but they suffer from numerous fundamental problems. In this article, a novel concept, called millimeter-wave microphone (mmPhone), is proposed, which can recover sound signals with high quality via millimeter-wave vibration perception. In mmPhone, with acoustic excitation, the vibration response of multiple light objects can be perceived, but different objects have different frequency response bandwidths. Thus, we propose a sound recovery method of multitarget adaptive fusion enhancement (MAFE), allowing reconstruction of the real full band sound signals. The overview of the mmPhone system and principle is first depicted. The detailed procedures including object selection, denoise, and adaptively fusion enhancement for implementing the MAFE method are then illustrated. Finally, experimental results with different scenarios are presented to validate the performance of mmPhone, showing that the proposed method is promising for sound acquisition in the Internet of Things (IoT).