Nowadays, many elderly people live in loneliness. They might feel reluctant to call for help from family members living or working far away, fearing that they might become a burden. The neighbourhood provide chances of social life and interaction for them.

Besides, many have difficulty with daily routines in their own house. Alzheimer patients often forget concepts and past activities they take during the day. The problem is not only a nuisance but also hampers their confidence, making them dependent and therefore even more lonely.

What it does

Our solution help an elderly person with both aforementioned problem. In one hand, Pepper works as a friend and personal assistant that is always present in the house. It can help alleviate forgetfulness. For example, when the user forgets a word, he or she can describe the object they are trying to recall and Pepper will guess the word for it. Another use case is when the user forget whether he or she have taken medicine/breakfast/exercise, Pepper can work as a talking diary.

On the other hand, Pepper is connected to sensors in the house and wearable watches to detect the emotion of user. It can then react to sadness, boredom or other negative feelings of the user by playing with the person, suggesting activities or connecting he or she with relatives or neighbours.

How we built it

We make use of FollowFlow infrastructure and Pepper robot. The intelligence of the robot is program with Choregraphe and Python scripts.

Challenges we ran into

Complexity is a big problem. The rules that a Pepper robot can handle is limited and we need to push to its limit to handle as much use cases as possible.

Accomplishments that we're proud of

We manage to instruct Pepper to have a meaningful and, more importantly, helpful conversation with users.

What we learned

We learn Choregrahpe for the first time here. It was also our first encounter with Pepper.

What's next for Pepper as a connecting portal

There are many opportunities such as trying machine learning, animate the robot more.

Built With

Share this project: