Meet Dewey

Shaking its head from left to right, a small robot makes a scratchy noise every few seconds.

“That’s Dewey,” Selma Šabanovic says, laughing and reaching her hand out to pull on one of its elastic spikes.

Created from a plastic bath caddy and rubber ball, both from Target, placed over a mesh of Arduino-brand microcontrollers, sensors, and actuators, Dewey has a light in its belly that turns from cool blue to striking red when it is “hungry” for the set of “fruits” – smart cards slotted into decorated pockets of colored paper – lying beside it.

Šabanovic is cofounder and current director of the R-House Living Lab, an on-campus laboratory on East 13th Street designated for research in human-robot interaction and development of robotic technologies that cater to everyday human life.

Šabanovic gets us as close as we can to actually speaking with the bath caddy-clad robot.

Who is Dewey?

I would call it a socially interactive robot, or maybe an assistive robot. We were interested in helping people who work on computer jobs to take more regular breaks, because what we realized was that when you’re working on a computer job you can kind of forget that you have a body, basically.

And so the question is, how can you remind people that they have bodies, and about their bodily needs? The idea is to have something that’s embodied.

How does Dewey interact with humans?

Dewey feeds on fruit cards, which are radio frequency identification tags, that have unique IDs associated with them.

Generally, the idea would be you’d have a Dewey on your desk, and then these cards would be at the water cooler.

There had been studies about how often you would need to take a break, so we used those studies to make kind of a timer for it.

So it would say you’ve been sitting for 30 minutes, ‘let me remind you to take a little breather,’ and it would move. You would come back with a card, which would reset the timer, back to the work mode.

In some ways, Dewey could be looked at as a little more advanced egg timer.

The first prototype didn’t have any interactive components. For the second iteration, we actually gave it a behavior, so if you want to play with it while it’s sitting on your desk, it responds. So we kind of made it more social. The idea was, maybe if you don’t care enough about yourself to get up and go take a break, maybe you’ll care if you have to go feed your little creature.

Each person had one robot, but the cards were in a common place. One thing was that you had to get up to get a card, so it was forcing you to do that, but also when you were up you might bump into your friend or something, or your colleague, and have a little chat.

How do these robots function if they’ve been made with such simple objects?

Most of them are not really made with “simple objects,” but with specialist prototyping equipment and platforms. The simple objects are sometimes used for the “shells,” partly because they are easily available and also because we want the robots to fit into homes and be somewhat familiar to users, even though they are a novel technology.

Why did you choose to use such hardware?

We use prototyping parts like Arduino because they are widely available and relatively cheap, as one of our goals is to make robotics accessible to the public. We also are interested in designing robots that are socially robust and fit well into their contexts of use.

For this we need to build and test out many design ideas through prototypes that we do not get too attached to. We need them to be cheap, reconfigurable, and not take too long to build, so that we will be fine with changing them when we learn new design requirements through studies.

Using these parts also makes it possible for other researchers to replicate our robots and do studies with them, which is good for science, and also any others who are interested can build them and improve on the designs and applications.

Readers can actually make their own. The parts are all available for purchase online or in stores like Target or Hobby Lobby. We are planning on putting up construction directions for the various robots on our website sometime soon.

How much do they cost?

The first Dewey prototype cost $120, while the second cost $100. The main goal was for the first prototype to be an education and research tool, we intend to put the second iteration in the market. Mugbot, which costs $600, was designed by visiting Japanese scientist Seita Koike, and is available for purchase. Katie and MiRAE, $250 and $300 respectively, were made as research tools.