Skip to Main content Skip to Navigation

Commande d'un déambulateur robotisé par la caractérisation posturale

Abstract : Mobility is a key factor for maintaining the autonomy of the elderly. This thesis proposes new approaches and solutions for three kinds of assistance in the field of smart walkers: fall prevention assistance; diagnosis and medical follow-up assistance; and mobility assistance. Towards preventing falls, this thesis presents an algorithm for detecting when balance is being lost (available literature mainly focuses on detecting falls, after it occured). Experiments involving healthy subjects show that our algorithm detects losses of balance within 600 ms. Towards assisting diagnosis and medical follow-up, this thesis proposes a portative and affordable device (kinect-like sensor) that offers a relevant tradeoff between portativity, affordability and precision (better precision than the existing literature on markerless diagnosis). Our solution achieves better precision than the existing literature on markerless diagnosis. Our algorithm enables real-time walking analysis. This solution is validated through experiences involving healthy and pathological elderly participants. Towards robotic assistance for mobility, this thesis presents a new prototype, which we called RoAM (Robot for Assisting Mobility), and approaches for controlling it. Experiments on an ecological path included three modes of control: one based on user position (developped in this thesis), another one on interaction forces, and the last one combining the two previous modes. All participants could complete the experiments with all the three modes. We also show that the most promising track seems to consist of a fusion of a force-based control with a position-based control.
Document type :
Complete list of metadata

Cited literature [126 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Tuesday, February 18, 2020 - 3:07:26 PM
Last modification on : Tuesday, December 8, 2020 - 3:36:36 AM


Version validated by the jury (STAR)


  • HAL Id : tel-02483128, version 2


Solenne Page. Commande d'un déambulateur robotisé par la caractérisation posturale. Automatique / Robotique. Sorbonne Université, 2018. Français. ⟨NNT : 2018SORUS221⟩. ⟨tel-02483128v2⟩



Record views


Files downloads