Online Learning and Classification of EMG-Based Gestures on a Parallel Ultra-Low Power Platform Using Hyperdimensional Computing

• Published in: IEEE transactions on biomedical circuits and systems Vol. 13; no. 3; pp. 516 - 528
• Main Authors: Benatti, Simone, Montagna, Fabio, Kartsch, Victor, Rahimi, Abbas, Rossi, Davide, Benini, Luca
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.06.2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Autonomy; Batteries; Classification; Computation; Distance learning; Electromyography; Embedded systems; Energy budget; Energy efficiency; Energy management; Gesture recognition; Gestures; Hidden Markov models; HMI; Humans; hyperdimensional computing; Internet; Learning; Machine learning; Machine learning algorithms; Pattern Recognition, Automated; Personal computers; Power consumption; Power management; Prototyping; Pulp; PULP platform; Signal Processing, Computer-Assisted; Support vector machines; System on chip; Training; Wearable Electronic Devices; Wearable technology
• ISSN: 1932-4545; 1940-9990; 1940-9990
• DOI: 10.1109/TBCAS.2019.2914476

This paper presents a wearable electromyographic gesture recognition system based on the hyperdimensional computing paradigm, running on a programmable parallel ultra-low-power (PULP) platform. The processing chain includes efficient on-chip training, which leads to a fully embedded implementation with no need to perform any offline training on a personal computer. The proposed solution has been tested on 10 subjects in a typical gesture recognition scenario achieving 85% average accuracy on 11 gestures recognition, which is aligned with the state-of-the-art, with the unique capability of performing online learning. Furthermore, by virtue of the hardware friendly algorithm and of the efficient PULP system-on-chip (Mr. Wolf) used for prototyping and evaluation, the energy budget required to run the learning part with 11 gestures is 10.04 mJ, and 83.2 <inline-formula><tex-math notation="LaTeX">\mu</tex-math></inline-formula>J per classification. The system works with a average power consumption of 10.4 mW in classification, ensuring around 29 h of autonomy with a 100 mAh battery. Finally, the scalability of the system is explored by increasing the number of channels (up to 256 electrodes), demonstrating the suitability of our approach as universal, energy-efficient biopotential wearable recognition framework.