Abstract
Due to a decline in the mortality of children under the age of one and people over the age of 70, life expectancy is increasing. The number of births is also much lower than during the baby boom years. These reasons have the consequence of making the world's population increasingly aging. The increase in this elderly population brings us new challenges. Indeed, their advanced ages increase the risk of loss of autonomy and tends to make this population more and more dependent. It is therefore essential to think about these issues now and find appropriate solutions. Companies are now offering robot assistants which can provide services and help people at home or in institutions. Health professionals are also interested in the help that these new tools can offer in order to lighten some of their repetitive tasks, or directly to provide services to their patients and thus supplement their existing care.The objective of our research is to set up a robotic assistance service capable of bringing an everyday object to an elderly person in a nursing home or in even from his home. The aim of this service is to reduce the movement of the person in order to limit their risk of falling, to optimise their living environment, to improve their comfort, and finally to preserve their autonomy. We propose a framework based on the bio-mechanical model of the human upper limbs and the effect of manipulability and gravity on the transfer position. The contribution of our research focuses on the generation of a transfer space adapted to the morphology and preferences of the user. We introduce a new relevant parameter for the comfort of the individual, which is the gravity index. It minimises the forces exerted on the joints of the human arm during the transfer. We adapt to the physical characteristics of the individual, i.e., we are able to define a personal transfer zone that takes into account their weight, height, joint amplitudes and their preferences. Finally, in order to bring objects as close as possible to dependent people, we also integrate the geolocation and navigation of the robot within a nursing home.All of our developments have been tested and validated during experiments involving elderly people with the Pepper robot. These experiments aimed to analyse the relevance of the generated transfer zone. The results show that the robot’s movements and the area reached for the transfer are satisfactory for retrieving the object in good conditions. The answers to the questionnaires proposed for these studies tend to show a general preference for the areas closest to them. That is to say with a high availability of the object and therefore few movements required to retrieve it. People with joint limitations are also satisfied with the proposed transfer position which seems well suited to their conditions.