DOI QR코드

DOI QR Code

Recognition of Indoor and Outdoor Exercising Activities using Smartphone Sensors and Machine Learning

스마트폰 센서와 기계학습을 이용한 실내외 운동 활동의 인식

  • Received : 2021.11.02
  • Accepted : 2021.11.11
  • Published : 2021.11.30

Abstract

Recently, many human activity recognition(HAR) researches using smartphone sensor data have been studied. HAR can be utilized in various fields, such as life pattern analysis, exercise measurement, and dangerous situation detection. However researches have been focused on recognition of basic human behaviors or efficient battery use. In this paper, exercising activities performed indoors and outdoors were defined and recognized. Data collection and pre-processing is performed to recognize the defined activities by SVM, random forest and gradient boosting model. In addition, the recognition result is determined based on voting class approach for accuracy and stable performance. As a result, the proposed activities were recognized with high accuracy and in particular, similar types of indoor and outdoor exercising activities were correctly classified.

스마트폰은 다양한 고성능의 센서가 포함되어 있으며 센서에서 발생하는 데이터를 이용하여 인간의 활동을 분석하는 연구가 진행되어왔다. 이러한 인간 활동 인식은 생활 패턴 분석, 운동량 측정, 위험 상황 감지 등 다양한 분야에서 활용될 수 있다. 그러나 기존 연구의 경우 인간의 기본 행동의 인식에 초점을 두거나 효율적인 배터리 사용을 위해 최적의 인식 결과를 내는 방법을 연구하는 경우가 많았다. 본 논문에서는 기본 행동에 건강 관리 목적으로 실내 및 실외에서 행해지는 운동 동작을 총 10가지로 정의하여 인식하도록 하였다. 이를 위해 가속도, 자이로 및 위치 센서의 값을 수집하고 데이터 전처리 과정을 거치고, 활동을 인식하기 위해서 SVM 모델 외에 안정적인 성능을 가진 앙상블 기반의 랜덤 포레스트, 그라디언트 부스팅 모델을 결합하여 투표 기반으로 인식 결과를 결정하였다. 그 결과 높은 정확도로 정의된 활동의 인식이 가능하였으며 특히 유사한 종류의 실내 및 실외 운동 활동의 분류가 가능하였다.

Keywords

References

  1. M. Meng., "AlphGo Shock and Changes of A.I. Technology Society", in Proceeding of 2017 Korean Association Of Science And Technology Studies, Busan, Korea, pp.237-237, 2017. (in Korean)
  2. Y. Lee and J. Suh, "Big Data Platform for Utilizing and Analyzing Real-Time Sensing Information in Industrial Sites", Journal of Creative Information Culture, vol.6, no.1, pp.15-21, 2020. (in Korean) https://doi.org/10.32823/JCIC.6.1.202004.15
  3. D. Anguita, A. Ghio, L. Oneto, X. Parra and J. Reyes-Ortiz, "Human Activity Recognition on Smartphones Using a Multiclass Hardware-Friendly Support Vector Machine", in Proceeding of 2012 International Workshop of Ambient Assisted Living, Vitoria-Gasteiz, Spain, pp.216-223, 2012.
  4. D. Anguita, A. Ghio and L. Oneto, "Energy Efficient Smartphone-Based Activity Recognition using Fixed-Point Arithmetic", Journal of Universal Computer Science, vol.19, no.9, pp.1295-1314, 2013.
  5. E. Kim, H. Jun, J. Gwak and J. Fenyi, "Development of physical activity classification table for Koreans: using the Compendium of physical activities in the United States", Journal of Nutrition and Health, vol.54, no.2, pp.129-138, 2021. https://doi.org/10.4163/jnh.2021.54.2.129
  6. S. Dernbach, B. Das, N. C. Krishnan, B. L. Thomas and D. J. Cook, "Simple and Complex Activity Recognition through Smart Phones," in Proceeding of 2012 Eighth International Conference on Intelligent Environments, Guanajuato, Mexico, pp.214-221, 2012
  7. E. Ha, J. Kim and K. Ryu, "Ensemble of Nested Dichotomies for Activity Recognition Using Accelerometer Data on Smartphone", Journal of Korea Intelligent Information System, vol.19, no.4, pp.123-132, 2013. (in Korean) https://doi.org/10.13088/jiis.2013.19.4.123
  8. J. Cho, "Human Activity Recognition Using Sensor Fusion and Kernel Discriminant Analysis on Smartphones", Journal of the Korea Convergence Societ, vol.11, no.5, pp.9-17, 2020. (in Korean) https://doi.org/10.15207/JKCS.2020.11.5.009
  9. METs for Activities with Fitness , https://community.plu.edu/~chasega/met.html, (visited Oct. 2021).
  10. D. Anguita, A. Ghio, L. Oneto, X. Parra and J. Reyes-Ortiz, "A Public Domain Dataset for Human Activity Recognition Using Smartphones", in Proceeding of 21th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, Bruges, Belgium, pp.24-26, 2013.