A New NN-Based Approach to In-Sensor PDM-to-PCM Conversion for Ultra TinyML KWS

• Published in: IEEE transactions on circuits and systems. II, Express briefs Express briefs, 2023-04, Vol.70 (4), p.1595-1599
• Main Authors: Vitolo, Paola, Liguori, Rosalba, Di Benedetto, Luigi, Rubino, Alfredo, Pau, Danilo, Licciardo, Gian Domenico
• Format: Article
• Language: English
• Subjects: Artificial neural networks; audio processing; Circuits; CMOS; Conversion; custom digital design; edge computing; Finite impulse response filters; in-sensor computing; Microelectromechanical systems; Micromechanical devices; Microphones; Neural network; Neural networks; Power consumption; Pulse code modulation; Pulse duration modulation; Pulse modulation; Signal to noise ratio; Training; ultra-low-power
• ISSN: 1549-7747; 1558-3791
• DOI: 10.1109/TCSII.2022.3224022

This brief proposes a new approach based on a tiny neural network to convert Pulse Density Modulation (PDM) signals acquired from digital Micro-Electro-Mechanical System (MEMS) microphones into the standard Pulse Code Modulation (PCM) format for any further digital audio processing. The proposed approach allows for a compact and ultra-low-power hardware implementation of the conversion, suitable for ultra tinyML Key Word Spotting (KWS) applications, closely coupled with the sensor itself and tightly coupled with a neural network classifier. The converter achieves a signal-to-noise ratio value of 48 dB, which enables KWS accuracy of 89% over 12 classes. Implementation on a Xilinx Artix-7 FPGA results in 917 LUTs, 361 FFs, and 182 \mu \text{W} Dynamic Power (DynP) consumption. By targeting the TSMC 0.13 \mu \text{m} CMOS technology, the synthesis reports an area occupation of 0.086 mm2 and a DynP of 128.7 \mu \text{W} /MHz. These results enable the integration of the proposed design into the CMOS circuitry closely coupled with the MEMS microphone.