Creating a robot that elicits love. How do you make people protect a robot? By eliciting love through designed human-robot interaction. Our robot's form had to be perceived as friendly, small, and approachable, all within strict engineering contraints. We transformed a looming 6-foot-tall robot into a short and friendly robot, without making it appear interactive.

I was the designer for this project done at Bossa Nova Robotics in 2015.

Challenge

Bossa Nova uses a robot to collect data in big box retail stores. Through HRI research, we knew Bossa Nova’s current robot looked techy, fragile, and expensive. We needed to design a new form that presents a persona that is accepted and loved, so that shoppers allow the robot to do its job.

Small wins

Through HRI research we knew there were aspects of the form of Bossa Nova’s robot that needed to change to gain acceptance from shoppers. In collaboration with industrial design and mechanical engineering, we designed a robot that shoppers will love.

The back of the old generation of the Bossa Nova robot.

The previous generation of Bossa Nova’s robot balanced on two wheels, much like a Segway. While this is a cool engineering feat, the sight makes shoppers uncomfortable—especially when the robot transitions from a balancing to non-balancing state by falling back onto casters. The robot appeared fragile, and its movements were jerky. Shoppers were uncertain about how to move around the robot because they were afraid they would knock it over. We don’t want our robot to negatively affect the shopping experience, and making customers uneasy was a negative. Fortunately, it was to the benefit of both engineering and design goals to have a statically stable robot, so we could make the next-generation robot non-balancing.

Considerations for industrial design

Our robot needed to appear robust and friendly. This meant moving beyond the dark colors and hard lines of the masculine stereotype seen so frequently in today’s technology. This was an opportunity to make robotics more relatable. If this robot becomes part of everyday life, the way it looks could go a long way to changing stereotypes. At the same time, the robot couldn’t look too friendly, or it would invite interaction from shoppers. We don’t want people touching it. The robot can’t have traditional humanoid features, otherwise it would appear too interactive. This all made designing communication patterns much more interesting.

Another consideration was that the robot is going to be on its own without a handler, and people would be more likely deter the robot from its goal if it elicited feelings of disgust and rage. Instead, we had to find a form that would elicit love and adoration with the right balance of cuteness; too far on the cute side of the spectrum and it could invite more interaction than we’d like.

It’s really hard to make a 6-foot-tall robot cute. To make it even more difficult, required the next version of the robot to be even taller. Somehow, we needed to shrink the robot. We had a hunch that we could do this by visually separating the parts of the robot that were required to be tall from the parts of the robot that drove it around. This way, what shoppers would describe as the “robot” appeared to be only 3 feet tall, dragging along the part that was 6 feet tall.

Validating the form through testing

We worked with an industrial design firm that created a new shape to meet all of these goals. We then had a foam model of that shape made, and we attached it to a drive train. We took this mobile foam model into retail stores, drove it around, and asked shoppers what they thought. To our delight, the visual separation of the shapes worked. When shoppers talked about the robot, they looked down at the small 3-foot form instead of up at the top. Shoppers thought of the robot as small, despite the fact that it was over 6 feet tall.

Pieces of a foam model before they were glued together and painted.

People also had behaviors towards the robot that suggested they found it cute. One store employee even patted the top of the small form while saying, “Hey, buddy,” with an endearing tone. Kids also loved it. They ran up and gave it bear hugs, and they stood in front of it trying to play.

A kid playing with the robot model.

Through these interactions we realized that we may want to increase the height of the form perceived by shoppers as the actual robot. Despite how cute the interactions with the kids are, we want to avoid having the robot sit at at exactly kid height because it could interfere with the robot’s data collection. Testing also revealed some other negative connotations, so we are still working with another industrial design firm to iterate on the shape. This is an ongoing project, and the new form will be released mid-2017. Once we have a shape, we will be integrating communication patterns and new behaviors.