신소재공학과

Research Highlight

A KAIST research team led by Professor Keon Jae Lee from the Department of Material Science and Engineering has developed a machine learning-based acoustic sensor for speaker recognition.  

Acoustic sensors were spotlighted as one of the most intuitive bilateral communication devices between humans and machines. However, conventional acoustic sensors use a condenser-type device for measuring capacitance between two conducting layers, resulting in low sensitivity, short recognition distance, and low speaker recognition rates. 

The team fabricated a flexible piezoelectric membrane by mimicking the basilar membrane in the human cochlear. Resonant frequencies vibrate corresponding regions of the trapezoidal piezoelectric membrane, which converts voice to electrical signal with a highly sensitive self-powered acoustic sensor. 

This multi-channel piezoelectric acoustic sensor exhibits sensitivity more than two times higher and allows for more abundant voice information compared to conventional acoustic sensors, which can detect minute sounds from farther distances. In addition, the acoustic sensor can achieve a 97.5% speaker recognition rate using a machine learning algorithm, reducing by 75% error rate  than the reference microphone. 

AI speaker recognition is the next big thing for future individual customized services. However, conventional technology attempts to improve recognition rates by using software upgrades, resulting in limited speaker recognition rates. The team enhanced the speaker recognition system by replacing the existing hardware with an innovative flexible piezoelectric acoustic sensor. Further software improvement of the piezoelectric acoustic sensor will significantly increase the speaker and voice recognition rate in diverse environments. 

Professor Lee said, “Highly sensitive self-powered acoustic sensors for speaker recognition can be used for personalized voice services such as smart home appliances, AI secretaries, always-on IoT, biometric authentication, and FinTech.” 

These research “Basilar Membrane-Inspired Self-Powered Acoustic Sensor” and “Machine Learning-based Acoustic Sensor for Speaker Recognition” were published in the September 2018 issue of Nano Energy. 

Flexible Piezoelectric Acoustic Sensors for Speaker Recognition 이미지1 
    Firgure 1: A flexible piezoelectric acoustic sensor mimicking the human cochlear.

 

Flexible Piezoelectric Acoustic Sensors for Speaker Recognition 이미지2 
Figure 2: Speaker recognition with a machine learning algorithm. 

No. Subject Author Date Views
» Flexible Piezoelectric Acoustic Sensors for Speaker Recognition ADMINI 2018.10.04 63
Notice Spray Coated Tactile Sensor on a 3-D Surface for Robotic Skin ADMINI 2018.10.04 73
Notice Levitating 2D Semiconductor for Better Performance ADMINI 2018.08.29 341
Notice Flexible Drug Delivery Microdevice to Advance Precision Medicine ADMINI 2018.08.14 396
Notice A New Efficient Oxide Coating Technology to Improve Fuel Cells ADMINI 2018.08.03 516
Notice A High-Performance and Cost Effective Hydrogen Sensor ADMINI 2018.08.03 423
Notice KAIST Team Develops Flexible Blue Vertical Micro LEDs ADMINI 2018.06.26 653
Notice Platinum Catalyst Has Price Lowed and Durability Doubled ADMINI 2018.06.26 709
39 Capillary Forces at Work for Lithium-Sulfur Batteries ADMINI 2018.06.26 577
38 New Material for Generating Energy-Efficient Spin Currents ADMINI 2018.06.26 506
37 Low-power, Flexible Memristor Circuit for Mobile and Wearable Devices ADMINI 2018.06.26 388
36 KAIST Develops Sodium Ion Batteries using Copper Sulfide ADMINI 2018.04.24 1109
35 Researchers develop flexible vertical micro LED file ADMINI 2018.02.07 1222
34 Lifespan of Fuel Cells Maximized using Small Amount of Metals file ADMINI 2018.01.21 1288
33 A New Spin Current Generating Material Developed file ADMINI 2017.12.21 1373
32 Professor Kim Sang Ouk of KAIST participated in editorship of GRAPHIN NEW MATERIALS file ADMINI 2017.10.23 1739
31 Pd nanowire array-based ultra-fast hydrogen gas sensor by using engineered nanofiltration of metal-organic framework file ADMINI 2017.09.27 1920
30 7 nanometer Semiconductor Manufacturing Technology with Camera Flash file ADMINI 2017.09.14 1972
29 Development of disease diagnjosis sensor using animal protein catalyst file ADMINI 2017.07.20 2118
28 Making Graphene Using Laser-induced Phase Separation file ADMINI 2016.12.07 4554