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).

中文翻译：

---

mmPhone：使用具有自适应融合增强振动感应功能的毫米波无线电进行声音恢复


尽管驻极体电容麦克风（ECM）已广泛应用于声音采集，但声音采集技术仍然引起了越来越多的兴趣。为此，人们开发了激光麦克风和视觉麦克风等新兴技术，但它们存在许多基本问题。在本文中，提出了一种称为毫米波麦克风（mmPhone）的新颖概念，它可以通过毫米波振动感知来恢复高质量的声音信号。在mmPhone中，通过声学激励，可以感知多个轻物体的振动响应，但不同的物体具有不同的频率响应带宽。因此，我们提出了一种多目标自适应融合增强（MAFE）的声音恢复方法，允许重建真实的全频带声音信号。首先描述mmPhone系统和原理的概述。然后说明了用于实现 MAFE 方法的详细过程，包括对象选择、降噪和自适应融合增强。最后，给出了不同场景的实验结果来验证mmPhone的性能，表明所提出的方法对于物联网（IoT）中的声音采集很有前景。