DOI QR코드

DOI QR Code

Accuracy Improvement Methode of Step Count Detection Using Variable Amplitude Threshold

가변 진폭 임계값을 이용한 걸음수 검출 정확도 향상 기법

  • 류욱재 (대진대학교 컴퓨터공학과) ;
  • 김은태 (대진대학교 컴퓨터공학과) ;
  • 안경호 (대진대학교 컴퓨터공학과) ;
  • 장윤석 (대진대학교 컴퓨터공학과)
  • Received : 2013.02.07
  • Accepted : 2013.05.07
  • Published : 2013.06.30

Abstract

In this study, we have designed the variable amplitude threshold algorithm that can enhance the accuracy of step count using variable amplitude. This algorithm converts the x, y, z sensor values into a single energy value($E_t$) by using SVM(Signal Vector Magnitude) algorithm and can pick step count out over 99% of accuracy through the peak data detection algorithm and fixed peak threshold. To prove the results, We made the noise filtering with the fixed amplitude threshold from the amplitude of energy value that found out the detection error was increasing, and it's the key idea of the variable amplitude threshold that can be adapted on the continuous data evaluation. The experiment results shows that the variable amplitude threshold algorithm can improve the average step count accuracy up to 98.9% at 10 Hz sampling rate and 99.6% at 20Hz sampling rate.

본 연구에서는 3축 가속도 측정을 위한 LSM을 개발하고 가변 진폭을 이용하여 걸음수 검출 정확도를 향상시킨 가변 진폭 임계값 알고리즘을 설계하였다. 테스트 프로토콜에 따라 실험하여 수집한 x, y, z 값을 SVM(Signal Vector Magnitude) 알고리즘을 사용하여 하나의 에너지값($E_t$)으로 변환하고 Peak 데이터 검출 알고리즘과 고정 Peak 임계값을 사용하여 평균 99%이상의 정확도로 걸음수를 검출하였다. 그러나 검출한 걸음이 정확한 걸음임을 증명하기 위해 에너지값($E_t$)의 진폭 크기로부터 고정 진폭 임계값을 구하고 노이즈를 필터링 한 결과 걸음수 검출 오차율이 증가하였다. 따라서 본 연구에서는 오차율을 줄이기 위하여 고정 진폭 임계값이 아닌 데이터를 관찰하여 적응적으로 변화하는 가변 진폭 임계값 알고리즘을 설계하였다. 가변 진폭 임계값 알고리즘을 적용한 결과, 걸음수 검출의 평균 정확도는 샘플링 주기 10Hz에서 평균 98.9%, 20Hz에서는 99.6%로 높아졌다.

Keywords

References

  1. Dong-Soo Ha, Sung-June Park, "Smart-Phone based User Movement State Identification Algorithm", Journal of the Korea Society of Computer and Information, Vol.16, No.3, pp.167-174, 2011 https://doi.org/10.9708/jksci.2011.16.3.167
  2. Tan Joshep, "E-Health Information System," San Francisco : Jossey-Bass, 2005.
  3. Adnan Saeed, Miad Faezipour, Mehrdad Nourani, Subhash Banerjee, Gil Lee, Gopal Gupta, and Lakshman Tamil, "A Scalable Wireless Body Area Network for Bio-Telemetry," Journal of Information Processing Systems. Vol.5 No.2, pp.77-86, June, 2009. https://doi.org/10.3745/JIPS.2009.5.2.077
  4. R.boulic, N,M. Thalmann, "A Global Human Waling Model With Realtime Kinematic Rersonification", The Visual Computer, Vol.6, pp.344-358, 1991.
  5. I. H Lee, J. C. Kim, S. M. Jung, Sun K. Yoo, "The Detection of Gait Cycle and Realtime Monitoring System Using the Accelerometer", The Korean Institute of Electrical Engineers, CICS 2008, pp.476-477, 2008.
  6. Han-Jin Jang, Jeong Won Kim, Dong-Hwan Hwang, "Design of a Robust Pedometer for Personal Navigation System against Ground Variation and Walking Behavior", The transactions of the Korean Institute of Electrical Engineers. D, Vol.55, No.9, pp.420-422, 2006.
  7. Kim Nam-Jin, Hong Joo-Hyun, Lee Tae-Soo, "Motion Sensor Data Normalization Algorithm for Pedestrian Pattern Detection", The Journal of the Korea Contents Association, Vol.5, No.4, pp.94-102, 2005.
  8. Yoo Hyang-Mi, Suh Jae-Won, Cha Eun-Jong, Bae Hyeon-Deok, "Walking Number Detection Algorithm using a 3-Axial Accelerometer Sensor and Activity Monitoring", The Journal of the Korea Contents Association, Vol.8, No.8, pp.253-260, 2008. https://doi.org/10.5392/JKCA.2008.8.8.253
  9. R.W DeVaul, S.DUNN, "Real-time motion classfication for wearable computing applications", Technical report, MIT media LAB, 2001.
  10. S.H. Shin, C,G. Park, "Adaptive Step Length Estimation Algorithm Using Low-Cost MEMS Inertial Sensors", IEEE Sensors Applications Symposium, San Diego, California USA, pp.1-5, Feb., 2007.
  11. "BMA250 Datasheet", BOSCH.