Embedded machine learning on accelerometer data for exercise classification

Rufyid U. Nissa, Nemai Chandra Karmakar, Maryam Shojaei Baghini

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

1 Citation (Scopus)

Abstract

This paper describes the use of embedded machine learning (ML) on raw accelerometer data to classify three lower-limb exercises. The developed model, which uses supervised ML can accurately classify the exercises. The model performed with an accuracy of 99.41 % on the test data and F1 score of 0.99. The trained model is deployed on the microcontroller board, Arduino Nano 33 BLE Sense (ARM Cortex M4F processer). The system has been developed using open-source platforms. The developed wearable device, PhysioSens has the potential to enhance limb exercise monitoring in home-based physical therapy.

Original languageEnglish
Title of host publicationAPSCON 2023 Symposium Proceedings
EditorsS. Gopalakrishnan, Hamida Hallil
Place of PublicationPiscataway NJ USA
PublisherIEEE, Institute of Electrical and Electronics Engineers
Number of pages3
ISBN (Electronic)9781665461634
ISBN (Print)9781665461641
DOIs
Publication statusPublished - 2023
EventIEEE Applied Sensing Conference 2023 - Bengaluru, India
Duration: 23 Jan 202325 Jan 2023

Conference

ConferenceIEEE Applied Sensing Conference 2023
Abbreviated titleAPSCON 2023
Country/TerritoryIndia
CityBengaluru
Period23/01/2325/01/23

Keywords

  • embedded machine learning
  • inertial sensor
  • mems accelerometer
  • physical therapy

Cite this