MSFANC: A Case Study of Memory-Based Computing for Wearable Device Applications

Weibang DAI; Xiaogang CHEN; Houpeng CHEN; Sannian SONG; Yichen SONG; Shunfen LI; Tao HONG; Zhitang SONG

doi:10.1587/transfun.2024EAL2074

Weibang DAI, Xiaogang CHEN, Houpeng CHEN, Sannian SONG, Yichen SONG, Shunfen LI, Tao HONG, Zhitang SONG

Author information

Weibang DAI
Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences
Xiaogang CHEN
Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences
Shanghai Xinchu Integrated Circuits Co., Ltd.
Houpeng CHEN
Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences
Sannian SONG
Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences
Yichen SONG
Shanghai Xinchu Integrated Circuits Co., Ltd.
Shunfen LI
Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences
Tao HONG
Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences
Zhitang SONG
Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences

Keywords: memory-based computing, lookup table, wearable devices, active noise control, emerging NVM

JOURNAL FREE ACCESS Advance online publication

Article ID: 2024EAL2074

DOI https://doi.org/10.1587/transfun.2024EAL2074

Details

Abstract

This letter emphasizes memory-based computing's potential in enhancing efficiency for wearable devices with limited resources, focusing on active noise control (ANC) in wireless earbuds. Our proposed memory-based selective fixed-filter ANC (MSFANC) scheme reduces power consumption and delay through lookup table (LUT) computation and simplified filter matching. Theoretical analysis and simulations demonstrate MSFANC's effectiveness, offering a new paradigm for resource-constrained wearable systems with non-volatile memories' (NVM) support.

Corresponding author

Register with J-STAGE for free!