Développement d'un robot d'analyse de la locomotion et d'entrainement
|Abstract:||Whether for performance enhancement or physical rehabilitation, the analysis of fitness and locomotion occupies an important place in today's healthcare. The recent democratization of microcontrollers and their suite of sensors has allowed the emergence of new methods for evaluating locomotion (e.g., kinematic analysis with inertial systems). This dissertation presents the development of an autonomous vehicle of small size allowing the video capture of a participant according to an adaptive tracking that could be used for movement analysis and physical condition evaluations performed outside of laboratory environments and clinics (e.g., Instrumented room, treadmills). The robot, manufactured at low cost, is based on the Raspberry Pi platform. In order to conduct a video data acquisition in an ecological environment, it is placed on an indoor running track on which it can move independently in the corridors by following the lane lines according to a visual recognition and automation algorithm. The onboard instrumentation of the vehicle allows the participant to be evaluated in "follow-up" mode, that is to say by following the pace while maintaining a constant and safe distance with the assessed participant. The ‘pacesetter” mode, for its part, imposes a pace on the participant. This mode can also be used for performance analysis or training purposes. Throughout this document, the design and manufacturing methods will be presented. Computer vision methods for autonomous driving developed for a low cost, low-power computer will be detailed. The results of tests carried out on an indoor running track to demonstrate the performances and limitations of the vehicle are presented. Finally, new approaches for fitness assessment will be proposed.|
|Document Type:||Mémoire de maîtrise|
|Open Access Date:||13 September 2021|
|Collection:||Thèses et mémoires|
All documents in CorpusUL are protected by Copyright Act of Canada.