What robotics learned from Pixar

Each year brings us closer to the day when robotic companions will become an integral part of our homes, schools, hospitals and offices. However, for robots to be truly accepted in our personal space, their social interactions with us must acquire the kind of fluency and coordination that humans expect from each other. This is one of the challenges addressed by Guy Hoffman, the co-director of the Media Innovation Lab at IDC Herzilya in Israel and possibly one of the most original thinkers in robotics today.

Collaborative fluency implies a coordinated and synchronised meshing of joint activities between several participants. Among the most significant parameters that affect the level of fluency and coordination are the anticipation and timing of robotic movements. The problem with this is that the majority of existing robots are designed and programmed in such a way that requires them to first analyse human movements, calculate the appropriate response and only then act accordingly, all of which delay the robot’s movements and contribute to their jerkiness and unnaturalness.

Guy Hoffman was one of those researchers who realised that eliciting emotional response has more to do with how a robot moves than how it looks. Hoffman was initially inspired by Pixar’s animated short film that featured a pair of desk lamps, who, despite their non-anthropomorphic appearance managed to provoke a strong emotional response exclusively by means of right timing and sound effects.

His subsequent experiences with computer animation in combination with his enthusiasm in robotics led him to MIT where he created AUR, a real-world robotic counterpart of Pixar’s lamp, capable of quietly assisting a human based on anticipating his movements rather than providing a straightforward calculated response. Thanks to AUR’s smooth and obedient behaviour, people who interacted with the lamp had a more positive and fulfilling emotional experience.

Shimon robot

Shimon can improvise music together with human muscicians

According to Hoffman, robotic intelligence can be essentially classified either as a traditional “calculated” intelligence that works in a chess-like manner or a more intuitive “adventurous” intelligence that tries to anticipate its partner’s movements. Anticipating the full range of movement, however, is tricky and Hoffman’s robots still tend to commit more mistakes along the way. Even so, studies demonstrate that people prefer such less perfect robots to their more accurate, but less understanding twins.

Travis, a robotic speaker dock released in 2012

Travis, a robotic speaker dock released in 2012

With one of his latest robotic creations Shimon, Hoffman ventured into the world of music improvisation, where he tried to apply the same principles of fluent collaboration. Why music improvisation? Because it is a time-critical interaction that Hoffman saw as an ideal ground to test his ideas. Shimon is basically a robotic marimba virtuoso that can jam with human musicians in real time.


You can also check out Travis (aslo Shimi), a cute speaker dock released by Hoffman in 2012, which not only plays music, but also enjoys it himself.

Share this article:

    One thought on “What robotics learned from Pixar

    1. Pingback: These robots have some serious dance moves | Convergent Science Network

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>