Human activity recognition using inertial measurement units and smart shoes

Prudhvi Tej Chinimilli, Sangram Redkar, Wenlong Zhang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

8 Citations (Scopus)

Abstract

This paper presents an intelligent fuzzy inference (IFI) algorithm using inertial measurement units (IMUs) and smart shoes to recognize human activities. IFI algorithm recognizes the activities based on ground contact forces (GCFs) and the knee joint angles. The smart shoes are designed to measure GCFs exerted by the wearer. A total of four IMUs are mounted on bilateral thighs and shanks to provide acceleration and angular rate data. To calculate knee flexion extension, a calibration procedure is adopted which eliminates the need for an external camera system. Then, an extended Kalman filter (EKF) is used to estimate the relative orientations of thigh and shank segments, from which knee angle is calculated. Random forest search (RFS) technique is used as a baseline to compare with the performance of the IFI algorithm. To evaluate the performance of this algorithm, several outdoor experiments are conducted on two healthy subjects for six activities including sitting, standing, walking, going upstairs, going downstairs and jogging. The results show that the algorithm is capable of classifying six activities with higher precision and less update time compared to the baseline approach for both subject dependent and independent tests. Also, the algorithm detects transitions between all the activities smoothly such as sit-to-stand or stand-to-walk with higher precision.

Original languageEnglish (US)
Title of host publication2017 American Control Conference, ACC 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1462-1467
Number of pages6
ISBN (Electronic)9781509059928
DOIs
StatePublished - Jun 29 2017
Event2017 American Control Conference, ACC 2017 - Seattle, United States
Duration: May 24 2017May 26 2017

Other

Other2017 American Control Conference, ACC 2017
CountryUnited States
CitySeattle
Period5/24/175/26/17

Fingerprint

Units of measurement
Fuzzy inference
Extended Kalman filters
Cameras
Calibration
Experiments

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Chinimilli, P. T., Redkar, S., & Zhang, W. (2017). Human activity recognition using inertial measurement units and smart shoes. In 2017 American Control Conference, ACC 2017 (pp. 1462-1467). [7963159] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.23919/ACC.2017.7963159

Human activity recognition using inertial measurement units and smart shoes. / Chinimilli, Prudhvi Tej; Redkar, Sangram; Zhang, Wenlong.

2017 American Control Conference, ACC 2017. Institute of Electrical and Electronics Engineers Inc., 2017. p. 1462-1467 7963159.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Chinimilli, PT, Redkar, S & Zhang, W 2017, Human activity recognition using inertial measurement units and smart shoes. in 2017 American Control Conference, ACC 2017., 7963159, Institute of Electrical and Electronics Engineers Inc., pp. 1462-1467, 2017 American Control Conference, ACC 2017, Seattle, United States, 5/24/17. https://doi.org/10.23919/ACC.2017.7963159
Chinimilli PT, Redkar S, Zhang W. Human activity recognition using inertial measurement units and smart shoes. In 2017 American Control Conference, ACC 2017. Institute of Electrical and Electronics Engineers Inc. 2017. p. 1462-1467. 7963159 https://doi.org/10.23919/ACC.2017.7963159
Chinimilli, Prudhvi Tej ; Redkar, Sangram ; Zhang, Wenlong. / Human activity recognition using inertial measurement units and smart shoes. 2017 American Control Conference, ACC 2017. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 1462-1467
@inproceedings{d4e1dd22f98d4f01a25c488fb2d38a16,
title = "Human activity recognition using inertial measurement units and smart shoes",
abstract = "This paper presents an intelligent fuzzy inference (IFI) algorithm using inertial measurement units (IMUs) and smart shoes to recognize human activities. IFI algorithm recognizes the activities based on ground contact forces (GCFs) and the knee joint angles. The smart shoes are designed to measure GCFs exerted by the wearer. A total of four IMUs are mounted on bilateral thighs and shanks to provide acceleration and angular rate data. To calculate knee flexion extension, a calibration procedure is adopted which eliminates the need for an external camera system. Then, an extended Kalman filter (EKF) is used to estimate the relative orientations of thigh and shank segments, from which knee angle is calculated. Random forest search (RFS) technique is used as a baseline to compare with the performance of the IFI algorithm. To evaluate the performance of this algorithm, several outdoor experiments are conducted on two healthy subjects for six activities including sitting, standing, walking, going upstairs, going downstairs and jogging. The results show that the algorithm is capable of classifying six activities with higher precision and less update time compared to the baseline approach for both subject dependent and independent tests. Also, the algorithm detects transitions between all the activities smoothly such as sit-to-stand or stand-to-walk with higher precision.",
author = "Chinimilli, {Prudhvi Tej} and Sangram Redkar and Wenlong Zhang",
year = "2017",
month = "6",
day = "29",
doi = "10.23919/ACC.2017.7963159",
language = "English (US)",
pages = "1462--1467",
booktitle = "2017 American Control Conference, ACC 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - Human activity recognition using inertial measurement units and smart shoes

AU - Chinimilli, Prudhvi Tej

AU - Redkar, Sangram

AU - Zhang, Wenlong

PY - 2017/6/29

Y1 - 2017/6/29

N2 - This paper presents an intelligent fuzzy inference (IFI) algorithm using inertial measurement units (IMUs) and smart shoes to recognize human activities. IFI algorithm recognizes the activities based on ground contact forces (GCFs) and the knee joint angles. The smart shoes are designed to measure GCFs exerted by the wearer. A total of four IMUs are mounted on bilateral thighs and shanks to provide acceleration and angular rate data. To calculate knee flexion extension, a calibration procedure is adopted which eliminates the need for an external camera system. Then, an extended Kalman filter (EKF) is used to estimate the relative orientations of thigh and shank segments, from which knee angle is calculated. Random forest search (RFS) technique is used as a baseline to compare with the performance of the IFI algorithm. To evaluate the performance of this algorithm, several outdoor experiments are conducted on two healthy subjects for six activities including sitting, standing, walking, going upstairs, going downstairs and jogging. The results show that the algorithm is capable of classifying six activities with higher precision and less update time compared to the baseline approach for both subject dependent and independent tests. Also, the algorithm detects transitions between all the activities smoothly such as sit-to-stand or stand-to-walk with higher precision.

AB - This paper presents an intelligent fuzzy inference (IFI) algorithm using inertial measurement units (IMUs) and smart shoes to recognize human activities. IFI algorithm recognizes the activities based on ground contact forces (GCFs) and the knee joint angles. The smart shoes are designed to measure GCFs exerted by the wearer. A total of four IMUs are mounted on bilateral thighs and shanks to provide acceleration and angular rate data. To calculate knee flexion extension, a calibration procedure is adopted which eliminates the need for an external camera system. Then, an extended Kalman filter (EKF) is used to estimate the relative orientations of thigh and shank segments, from which knee angle is calculated. Random forest search (RFS) technique is used as a baseline to compare with the performance of the IFI algorithm. To evaluate the performance of this algorithm, several outdoor experiments are conducted on two healthy subjects for six activities including sitting, standing, walking, going upstairs, going downstairs and jogging. The results show that the algorithm is capable of classifying six activities with higher precision and less update time compared to the baseline approach for both subject dependent and independent tests. Also, the algorithm detects transitions between all the activities smoothly such as sit-to-stand or stand-to-walk with higher precision.

UR - http://www.scopus.com/inward/record.url?scp=85027034764&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85027034764&partnerID=8YFLogxK

U2 - 10.23919/ACC.2017.7963159

DO - 10.23919/ACC.2017.7963159

M3 - Conference contribution

SP - 1462

EP - 1467

BT - 2017 American Control Conference, ACC 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -