A 2.5 μW KWS Engine With Pruned LSTM and Embedded MFCC for IoT Applications

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
$\mu \text{W}$
, which is at least
$2\times $
lower than the state-of-the-art.