Mind the gap - development of conversion models between accelerometer- and IMU-based measurements of arm and trunk postures and movements in warehouse work

Mikael Forsman; Xuelong Fan; Ida-Märta Rhen; Carl Mikael Lind

doi:10.1016/j.apergo.2022.103841

Mind the gap - development of conversion models between accelerometer- and IMU-based measurements of arm and trunk postures and movements in warehouse work

Appl Ergon. 2022 Nov:105:103841. doi: 10.1016/j.apergo.2022.103841. Epub 2022 Jul 30.

Authors

Mikael Forsman¹, Xuelong Fan², Ida-Märta Rhen³, Carl Mikael Lind⁴

Affiliations

¹ School of Engineering Sciences in Chemistry, Biotechnology and Health, KTH Royal Institute of Technology, SE-141 57, Huddinge, Sweden; IMM Institute of Environmental Medicine, Karolinska Institutet, SE-171 77, Stockholm, Sweden; Centre for Occupational and Environmental Medicine, Stockholm County Council, SE-113 65, Stockholm, Sweden.
² IMM Institute of Environmental Medicine, Karolinska Institutet, SE-171 77, Stockholm, Sweden. Electronic address: xuelong.fan@ki.se.
³ IMM Institute of Environmental Medicine, Karolinska Institutet, SE-171 77, Stockholm, Sweden; Centre for Occupational and Environmental Medicine, Stockholm County Council, SE-113 65, Stockholm, Sweden; Department of Industrial and Materials Science, Chalmers University of Technology, SE-412 96, Gothenburg, Sweden.
⁴ IMM Institute of Environmental Medicine, Karolinska Institutet, SE-171 77, Stockholm, Sweden.

PMID: 35917697
DOI: 10.1016/j.apergo.2022.103841

Abstract

Sensor type (accelerometers only versus inertial measurement units, IMUs) and angular velocity computational method (inclination versus generalized velocity) have been shown to affect the measurements of arm and trunk movements. This study developed models for conversions between accelerometer and IMU measurements of arm and trunk inclination and between accelerometer and IMU measurements of inclination and generalized (arm) velocities. Full-workday recordings from accelerometers and IMUs of arm and trunk postures and movements from 38 warehouse workers were used to develop 4 angular (posture) and 24 angular velocity (movement) conversion models for the distributions of the data. A power function with one coefficient and one exponent was used, and it correlated well (r² > 0.999) in all cases to the average curves comparing one measurement with another. These conversion models facilitate the comparison and merging of measurements of arm and trunk movements collected using the two sensor types and the two computational methods.

Keywords: Field measurement; Inertial measurement unit; Sensor conversion.