Inspiration

Pepper and other communication robots become popular recently, and they are often treated as not "machines" but "partners" of human. So what will happen when they become more natural to be together with human beings? When you go lunch with your friends, your robot and your friend’s robot may accompany with you. The goal of our application is to achieve communication between multiple users and multiple robots.

What it does

When each user is in his/her house, the robot stores the communication logs, what user did in each application, and retrieves data from user’s wearable device. Based on that information, the robot talk with user’s friend and the friend’s robot when they are together. (ex. Robot says like “Tomo-chan ate pancakes in the morning. (Communication log)” “Yuka doesn’t do excersise recently (retrieve from activity log from fitbit)”) Also robots are sometimes doing secret talk when they meet each other, and sharing each user’s information. The robot may not be talking about the information in front of their users, but telling the user after they go home. For example, “Your friend birthday will come soon!” “Her robot told me that she looks tired recently, how about sending text to her?”... such kind of information will be a kind reminder for user or helpful information for communication between friends.

How we built it

Built with Choreographe(Pepper SDK), Python programming, and fitbit API.

Challenges we ran into

Connection between multiple Peppers. Basiclly, Pepper application is not designed to control multiple robots, but we achieved it using master and slave connection by python programming.

Accomplishments that we're proud of

There are less application which achieves communication between multiple robots with users. We are proud of our application which will be pioneer of future communication between human and robots.

What we learned

Potential and difficulties of running application on multiple robots

What's next for Pepper Partners

  • Implementation for storing communication logs and training process for more intelligent communication
  • Use Watson API to train from the communication log.

Built With

  • fitbit-api
  • pepper
  • python
  • qiscript
Share this project:

Updates