Human Voice Sensing through Radio-Frequency Technologies: A Comprehensive Review

Voiceprint is one of the promising solutions for biometric applications by distinctive patterns of certain voice characteristics of an individual. Due to the non-intrusive and high-accuracy sensing nature, many researchers have shown considerable attention to human voice sensing through radio frequency technologies (e.g., Wi-Fi and mmWave) thereby realizing a set of emerging applications. It has become a remarkable research field over the past decade. In this paper, we provide a comprehensive technological review of the current state-of-art of wireless technologies for human voice sensing. We perform a full analysis of voice print recognition connotations and applications and explore the principles of voice production. Human voice sensing technologies are divided into four modalities. In addition, we summarize the theory of wireless sensing and make an in-depth comparison of different radio frequency technologies for human voice sensing. An overview of human voice sensing is proposed, and this paper also highlights the voice signal acquisition, signal pre-processing, feature extraction as well as prediction and recognition. Furthermore, this paper presents a thorough survey of the current human voice sensing through radio frequency technologies and compares them in detail. We discuss the hardware system architecture and classify various applications of wireless voice sensing. At the end of the paper, we conclude the review with future directions and challenges associated with wireless human voice sensing.