Published October 20, 2023 | Version v1

A hardware-aware neural architecture search algorithm for wearable robotics

Description

Hardware-aware neural architecture search (HW NAS), the process of automating the design of neural architectures taking into consideration hardware constraints, has already outperformed the best human designs on many tasks. However, it is known to be highly demanding in terms of hardware, thus limiting access to non-habitual neural network users. Fostering its adoption for the next-generation wearable robotic devices design, we propose an HW NAS that can be run on laptops, even if not mounting a GPU. The proposed technique, designed to have both a low search cost and resource usage, produces tiny convolutional neural networks (CNNs) targeting low-end microcontrollers, typically applied in developing wearable robotic devices. Such CNNs can be used to analyse multiple sEMG or force signals, like in the force myography use case, to control wearable robotic devices without the need for costly and powerful hardware specifically designed to run CNNs on the edge. It achieves state-of-the-art results in the human-recognition tasks, on the Visual Wake Word dataset a standard TinyML benchmark, in just 3:37:0 hours on a laptop mounting an 11th Gen Intel(R) Core(TM) i7-11370H CPU @ 3.30GHz equipped with 16 GB of RAM and 512 GB of SSD, without using a GPU.

Files

file.pdf

Files (97.4 MB)

Name Size
md5:8e4528f6018e6492c76a3887da377612
547.6 kB Preview Download
md5:273edb16882da521cf75511914f6b676
96.8 MB Download