Editor's Recommendation: Researchers from Chongqing University and other units reported a waterproof acoustic sensor (WAS) with a wearable translation interface for communicating with machines, with a range of compelling features for high-fidelity recording, 98% speech recognition accuracy, and wireless control of smart cars using speech recognition. Represents a milestone in the field of human-computer interaction in artificial intelligence and supports the evolution from touch-based devices to voice-operated electronic systems.
The rapid development of artificial intelligence (AI), the Internet of Things (IoT), and smart homes is updating our lifestyles in meaningful and fundamental ways, and the cornerstone of this change is the powerful and accurate communication between people and machines. The human-machine interface that connects the user and the machine is a structured system for this communication, which can include speech or gestures, the latter being the language currently widely adopted. Communication and interaction with machines is changing the way we live. However, developing an acoustic interface that is waterproof, wear-resistant, high-fidelity and high-precision human-computer interaction is still a huge challenge.
Scholars from Chongqing University reported a waterproof acoustic sensor (WAS) for a wearable translation interface for communicating with machines. Due to the sound response capability of the internal particles, IS has a distinct frequency response range of 0.1-20 kHz, covering almost the entire human hearing range. WAS is stable in perspiration, displays a full range of responses, and displays excellent frequency detection resolution of 0.0001 kHz. WAS has a range of compelling features that can be used as a wearable acoustic HMI and high-fidelity music recording listening platform. In addition, with the help of artificial intelligence algorithms, the WAS-based acoustic interface has a significant speech recognition accuracy rate of 98%. Finally, the WAS-based acoustic interface demonstrates speaker authentication and identity recognition for implementation in a highly secure biometric authentication system and wireless control of smart cars using speech recognition. This WAS-based acoustic interface represents a high-fidelity translation platform for human-computer interaction moving toward practical applications including the Internet of Things, assistive technologies, and intelligent identification systems. The article was published in Advanced Functional Materials under the title "A Personalized Acoustic Interfacefor Wearable Human–Machine Interaction."
Thesis Link:
https://doi.org/10.1002/adfm.202109430
Figure 1. WAS design, how it works, and frequency response performance. a) Schematic diagram of a sound-sensitive wristband for wearable human-computer interaction applications. b) The photo of the made wearable device is. c) Paste a photo of the made wearable device on the sweaty wrist for acoustic sensing. d) Schematic design of WAS. e) Potential well model for charge transfer during periodic contact and separation of electrical signal generation. f) Compare the highest response frequency of the work with other reported works.
Figure 2. Response characteristics to WAS Sounds. a) Frequency response measurements at 94dB of SPL. b) Output voltage below 1 kHz with SPL of 94dB; c) FFT-processed spectrum. d) Sound pressure level to output voltage amplitude, etc. e) voltage and f) spectrum of two monochromatic sound waves with frequencies of 1 and 1.0001 kHz simultaneously excited by fast Fourier transforms.
Figure 3: A demonstration of a WAS-based sound-sensitive wristband and its waterproofing as a high-fidelity listening platform for music recording. a) Recording of the famous classical music "Hungarian Dance No. 5" b) original and recorded sound wave information and c) corresponding to the spectrogram. d) Record the sentence "This is a wearable wireless acoustic human-machine interface" before running.
Figure 4: WAS-based speaker verification system and real-time wireless intelligent car control system. a) Wass-based speaker confirmation system schematic. b) A photo demonstrating the WASS-based Speaker Verification System. c) A short-term Fourier transform spectrogram of the voltage signal obtained from the authorized user and the corresponding phrase "open door". d) Schematic diagram of real-time wireless intelligent car control system. e) A photo demonstrating a wast-based intelligent car control system.
In this work, a wearable, waterproof acoustic human-machine interface is reported, which has a full range of response, wide working bandwidth, self-powered operation and low cost. Compared to previously reported advanced acoustic sensors such as capacitive, piezoresistive, piezoelectric, and triboelectric acoustic sensors, WAS is waterproof and exhibits significant enhancements in the operating frequency range of 0.1-20 kHz due to the acoustic responsiveness of its internal particles, covering almost the entire human hearing range. In addition, the closed structure design and waterproof PET film give the WAS excellent waterproofness and stable resistance to human perspiration. In addition, economically viable materials and simple manufacturing procedures are adapted to mass production of WASAs. (Text: SSC)
This article is from the WeChat public account "Materials Science and Engineering". Welcome to reprint please contact, without permission to reprint to other websites.