IRSST - Institut de recherche Robert-Sauvé en santé et en sécurité du travail

Validation of a New Inertial Measurement System to Estimate Human Body Kinematics: The Case of Manual Material Handlers

Summary

It is now possible to carry out field studies with wearable motion capture systems capable of measuring the kinematics of the human body. However, these systems are often very expensive, which may prevent companies or professionals from purchasing them. The objective of this research was to validate a new and affordable wearable motion capture system (Noitom Ltd.’s Perception Neuron, which costs less than CAN$2000) to measure the complete kinematics of the human body and to provide accurate feedback of the task being performed, using the example of the work of manual material handlers.

For this purpose, five subjects were recruited and laboratory measurements were taken. The system consists of 17 inertial sensors (inertial measurement units (IMU)) placed on the various segments of the human body: feet, calves, thighs, pelvis, sternum, head, shoulder blades, upper arms, forearms and hands. These IMUs are made up of three 3-axis sensor units: accelerometers, gyroscopes and magnetometers. The measurements of these three sensors were combined to estimate the orientation of the IMUs in 3D space. The system was compared in the laboratory to a reference system recognized as valid, NDI’s optoelectronic Optotrak system.

The laboratory results revealed that the Neuron system was able to measure the kinematics of the majority of body segments with root mean square errors (RMSE) that fluctuated around an acceptable 5° error margin in the sagittal plane.

The prospects for using this type of system are promising. In addition to providing feedback to workers in training on how to perform their tasks, the system could also be used for partial measurements of workers’ physical exposure, as long as the workplace does not present strong magnetic disturbances.

Additional Information

Category: Research Report
Author(s):
  • Hakim Mecheri
  • Xavier Robert-Lachaîne
  • Antoine Muller
  • Christian Larue
  • André Plamondon
Research Project: 2018-0036
Online since: June 23, 2020
Format: Text