GALEN project team is part of an international consortium: the I-Support project
Launched as a consortium in March 2015, I-Support aims to design a robotics system to assist elderly people bathe independently. Nine research laboratories and companies are involved in the development of I-Support. The GALEN team at Inria Saclay - Île-de-France Research centre is working on the computer vision for the system.
A home support tool
In light of Europe's rapidly ageing population, the issue of support for the elderly so that they can remain in their own homes is an increasingly crucial one; according to forecasts by INSEE (France's National Institute for Statistics and Economic Studies), the number of dependent people in France is set to double by 2060. In addition, paying for a place in a care home or for home support is extremely onerous for the family. The starting point for the teams working on I-Support was the idea that bathing is the first Activity of Daily Living (ADL) that elderly people become unable to perform on their own. Together with dressing, getting around, feeding oneself and going to the toilet, bathing is key to retaining one's independence. When an elderly person loses any one of these faculties, they immediately become dependent on their family or require home care or care in a specialised residence. A robotics system designed to help a person bathe will therefore enable an elderly person to remain at home for longer, respecting their privacy, and, thereby, their feelings of dignity and self-esteem.
The I-Support consortium takes these issues very seriously. It has been discussing ethical, sociological and gender concerns with geriatricians specialising in such matters. Their discussions focus on what the robot should take care of: should it only assist the person in washing the most difficult-to-reach parts of the body, such as the back and feet? Or should it be able to help wash every part of the body? The answers to such questions will help make the machine more acceptable to the people who will use it. Other questions are also raised within the teams, for instance, regarding the control interface: the elderly are not necessarily used to using digital tools, so the control systems need to be simple and intuitive. The teams are primarily looking into a hybrid system integrating a basic remote control unit and voice command.
An international consortium
In March 2015, an international, multidisciplinary consortium was set up to develop the project. It involves research centres (the Karlsruhe Institute of Technology, the Institute of Communication and Computer Systems, the Scuola Superiore Sant’Anna BioRobotics Institute, Inria, the Fondazione Santa Lucia, Frankfurt University of Applied Sciences and the Bethanien Hospital/Geriatric Centre at the University of Heidelberg, together with the companies Robotnik Automation and Omega. It is funded by the European Commission under the Horizon 2020 research and innovation programme. The consortium aims to develop, within three years of research, a product that will be more or less ready for commercial development. With this in mind, the researchers hope to keep costs down by trying to adapt existing robotics solutions for this specific purpose.
If it wasn't multidisciplinary, the project would never have got off the ground.
Within the consortium, the GALEN team from the Inria Saclay - Île-de-France Research Centre and the Ecole Centrale de Paris is working on the computer vision design for I-Support. This is one of the robot's key features, since it must be capable of seeing all the different parts of the user's body. Iasonas Kokkinos is lead researcher for this project.
Why was the GALEN team chosen to be part of the consortium?
Our research mainly focuses on computer vision. Before joining I-Support, my research was already focused on healthcare, and in particular on designing computer vision systems to improve medical devices. For example, before joining I-Support, I was working on a tool to help elderly people walk. The main problem is the same: you need a computer vision system that can recognise the position of the person. For example, I-Support must be able to understand whether the person that needs assistance is sitting down, if their arm is raised, etc. When asked to scrub the lower back, it must be able to find it.
What technologies are you using to resolve this problem?
First, we are trying to adapt depth-sensing technology, as used for the Kinect sensor in the Xbox console, for example. This technology can recognise the player's movements, using the body like a joystick. However, it can only be used under certain conditions: the user must be at least 1 metre away and facing the system. For I-Support, neither of these conditions apply, so we need to adapt it. For even greater precision, we are also using machine learning methods to determine the user's position and gestures based on accumulated collected data. Last, with Galen, we are developing systems that combine 3D vision and machine learning to assess the position of the articulated arms in space.
What are the advantages of a consortium like this?
We are working with multidisciplinary teams, some of which specialise in robotics, while others focus on the language aspect, because I-Support will be voice-controlled. The doctors and geriatric centres are helping us to design a robot that will be the most suitable for and the most acceptable to elderly people. The project would never have got off the ground if it didn't combine all these specialisms.
© i-support project
I-support looks like a shower unit without sides. The user sits on a motorised chair and, for example, can turn the chair to wash their back. The unit is equipped with two articulated arms, one of which is a shower head and soap dispenser, while the other can hold a sponge for washing or a towel for drying.