Exploration of the potential of computer vision for object recognition towards the design of an intelligent device in an industrial context.

Résumé

The development of intelligent wheelchairs is a challenge that attracted several research laboratories working on the technological aspect. The use of these wheelchairs is hampered by the identification of real needs which pose a problem of acceptance and feasibility of the proposed solutions. The problem is all the more important when it comes to populations living in unfavorable regions of the world such as developing countries. People with reduced mobility in the Algerian context encounter several problems related to the adaptability of the wheelchairs, proposed in the market, to their real needs. On the other hand, the technological evolution has allowed the development of intelligent and automatically adaptable products at lower cost. This thesis aims to develop an intelligent wheelchair that meets people with reduced mobility requirements in the Algerian context. A first study allowed to develop a behavior model of people with reduced mobility based on a requirements collection and analysis. Then, the thesis addresses the design of the intelligent wheelchair based on system engineering (SE) and proposes a hybrid approach. This approach offers a design process based on the hybridization of the APTE method and the SysML annotation to allow a dynamic value analysis. Dynamic value analysis involves modeling and optimizing the impact of requirements change on the product, the process and the supply chain. The obtained results allowed us to propose an approach based on the evolution of a manual wheelchair towards an intelligent wheelchair at lower cost. This evolution is done through the re-design with the integration of a computer vision module in a manual wheelchair. Finally, the obtained results were implemented and modeled in CAD. Thus, a first assembly version is proposed.

Citer ce document

Accès au document

Ce document est hébergé sur son dépôt institutionnel d'origine.

Auteur(s)

Statistiques

Consultations : 3

Téléchargements : 0