KIT | KIT-Bibliothek | Impressum | Datenschutz

How to use human pose estimation to measure the hand-arm motion in craft application with no influence on the natural user behavior

Helmstetter, Sebastian; Sänger, Johannes ORCID iD icon; Germann, René; Matthiesen, Sven


The interaction between human and machine plays an important role in the design and optimization of human-machine systems. This interaction is characterized by human motion using the technical system. Especially in the field of hand and power tool applications, the motion capture should be performed under the real working condition and without influencing the user. There are already motion tracking systems that allow capturing the motion during the interaction, but there is no mobile motion capture system that allows an individual analysis of the user for biomechanical analysis in the normal work process without influencing him. Therefore, requirements for a motion capture system are derived and a system is presented that meets these requirements. This system consists of two cameras and is based on the pose estimation algorithm OpenPose. The comparison of the presented system and the state-of-the-art system Xsens is performed and based on the measured elbow angle and the wrist position. The results show a very good correspondence between the curve characteristic of the elbow angle and the wrist position of both systems. However, inexplicable values shifting at two different levels still occur, which need to be investigated further. ... mehr

Verlagsausgabe §
DOI: 10.5445/IR/1000133742
Veröffentlicht am 10.06.2021
Cover der Publikation
Zugehörige Institution(en) am KIT Institut für Produktentwicklung (IPEK)
Publikationstyp Zeitschriftenaufsatz
Publikationsjahr 2021
Sprache Englisch
Identifikator ISSN: 2212-8271
KITopen-ID: 1000133742
Erschienen in Procedia CIRP
Verlag Elsevier
Band 100
Seiten 631–636
Nachgewiesen in Scopus
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page