Team, Visitors, External Collaborators
Overall Objectives
Research Program
Application Domains
Highlights of the Year
New Software and Platforms
New Results
Bilateral Contracts and Grants with Industry
Partnerships and Cooperations
XML PDF e-pub
PDF e-Pub

Section: New Results

Motion Sensing of Human Activity

MimeTIC has a long experience in motion analysis in laboratory condition. In the MimeTIC project, we proposed to explore how these approaches could be transferred to ecological situations, with a lack of control on the experimental conditions. In the continuation of 2018, we have proposed to explore the use of cheap depth cameras solution for on-site motion analysis in ergonomics.

Motion Analysis of Work Conditions Using Commercial Depth Cameras in Real Industrial Conditions

Participant : Franck Multon [contact] .

Based on a former PhD thesis (of Pierre Plantard) we have demonstrated the use of depth sensors in industry to assess risks of musculoskeletal disorders ar work. It has leaded to the creation of the the KIMEA software and of the Moovency start-up company in November 2018. In 2019 we published a synthesis work with new results [48] to demonstrate that such an approach can actually support the work of ergonomists in their goal to enhance the quality of life of workers in industry.

Hence, measuring human motion activity in real work condition is challenging as the environment is not controlled, while the worker should perform his/her task without perturbation. Since the early 2010s, affordable and easy-to-use depth cameras, such as the Microsoft Kinect system, have been applied for in-home entertainment for the general public. In this work, we evaluated such a system for the use in motion analysis in work conditions and propose software algorithms to enhance the tracking accuracy. Firstly, we highlighted the high performance of the system when used under the recommended setup without occlusions. However, when the position/orientation of the sensor changes, occlusions may occur and the performance of the system may decrease, making it difficult to be used in real work conditions. Secondly, we propose a software algorithm to adapt the system to challenging conditions with occlusions to enhance the robustness and accuracy. Thirdly, we show that real work condition assessment using such an adapted system leads to similar results comparing with those performed manually by ergonomists. These results show that such adapted systems could be used to support the ergonomists work by providing them with reproducible and objective information about the human movement. It consequently saves ergonomists time and effort and allows them to focus on high-level analysis and actions.