A 2.5 <i>μ</i>W KWS Engine With Pruned LSTM and Embedded MFCC for IoT Applications

Page view(s)
28
Checked on Sep 10, 2023
A 2.5 <i>μ</i>W KWS Engine With Pruned LSTM and Embedded MFCC for IoT Applications
Title:
A 2.5 <i>μ</i>W KWS Engine With Pruned LSTM and Embedded MFCC for IoT Applications
Journal Title:
IEEE Transactions on Circuits and Systems II: Express Briefs
Publication Date:
20 September 2021
Citation:
Chong, Y. S., Goh, W. L., Nambiar, V. P., & Do, A. T. (2022). A 2.5 μW KWS Engine With Pruned LSTM and Embedded MFCC for IoT Applications. IEEE Transactions on Circuits and Systems II: Express Briefs, 69(3), 1662–1666. https://doi.org/10.1109/tcsii.2021.3113259
Abstract:
Always-on keyword spotting (KWS) hardware is gaining popularity in ultra-low power IoT applications where specific words are used to wake up and activate the power hungry downstream system. This work proposes a low power KWS engine with a power-optimized Mel-frequency cepstral coefficients (MFCC) feature extraction module and a memory-optimized long short term memory (LSTM) accelerator. Our LSTM model is pruned and compressed to reduce 89% of model size and 76% of computation. The LSTM accelerator adopts the weight stationary dataflow to reduce energy. Our simulation using a 40nm CMOS process achieves a power consumption of 2.51 μW , which is at least 2× lower than the state-of-the-art.
License type:
Publisher Copyright
Funding Info:
This research / project is supported by the Agency for Science, Technology and Research - Advanced Manufacturing and Engineering (AME) - Neuromorphic Processor Design - Neuron Circuits
Grant Reference no. : A1687b0033
Description:
© 2022 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
ISSN:
1558-3791
1549-7747
Files uploaded: