According to the latest report on the development of world-wide population proposed by the UN, 13 percent of today's population is over 60 years of age. By 2050, this percentage will be almost double and reach 25 percent. Society aging is a reality, and technology is adapting to this demographic change daily.
Researchers at the Universitat Politique de València (UPV), Universidad de Alicante (UA) and Universidade do Minho work on an interactive robot called Pharos that helps the elderly with their daily housework. Their work has recently been published in the journal Sensor .
"Our goal is to turn Pharos into another parent's companion, a virtual assistant, friendly and easy to use, who regularly recommends personal physical activities, promote a healthy life and facilitate the active aging of the population," emphasizes Vicente Julián, a researcher at UPV's computer technology and artificial intelligence group.
Pharos is based on a commercial robot, Pepper, as researchers have added two modules: The first recommends an exercise plan adapted to each user and the other, through advanced AI technicians, evaluates completion of the exercise, verifies if the user has completed it correctly through Deep Learning by comparing it with a library of stored exercises, and registering them on the system.
The robot also contains a visual and physical interface that the user should interact with to identify them through a camera. "Once identified, Pharos determines the most appropriate exercises depending on their ability. In addition, the training program is regularly adapted to the user's development and health. Therefore, it contains a recommendation algorithm that can also help detect health problems, explains Ester Martínez, UA researchers: s Roots and Three-Dimensional Vision Group (RoViT).
The biggest difference in Pharos compared to other similar systems is the detailed monitoring of users, as well as the ability to determine whether they perform the exercises properly or not. It is a very useful tool for caregivers and assistants, as it makes it possible to easily check whether the ability to perform certain exercises decreases, which may reveal progressive physical and / or cognitive problems.
"Low performance in implementing the planned exercises may indicate some type of problem with the user. By registering the exercise history, Pharos can help reveal underlying problems that would otherwise be impossible to see, "explaining the researchers.
The Spanish-Portuguese research group is currently working on the perfection of user robot interactions, optimization of the exercise recognition system, and allowing multiple users to use it without disturbing the others by using planning strategies.
Balance exercises can help people with multiple sclerosis
Angelo Costa et al. PHAROS-Fysical Assistant RObot System, Sensors (2018). DOI: 10.3390 / s18082633