Meet the real-life robots of UC San Diego, leading the way for a world of robot helpers, teachers, maybe even friends.
|Illustrations by Manny Pantoja '12|
Creators: Javier Movellan, cognitive scientist | Machine Perception Lab | Hanson Robotics | Kokoro Co.
Origins: No, researchers did not set out to create the world’s most adorable baby-headed robot. That was just a bonus. Diego-San is rather a multidisciplinary inquiry into robotics, computer science, developmental psychology and machine learning. This boy-bot was created to mimic the actions of a one-year-old child in order to study how young children learn to control their bodies and interact with people. On the technical side, the robot lends a greater computational understanding of how humans develop sensory motor intelligence.
Superpower: Smiling! Diego-San’s most popular finding has been revelations into the nature of babies’ smiles—what seems random and innocent actually serves a clear purpose. In fact, babies smile to make their mothers smile in return, yet they do this while smiling as little as possible. “If you’ve ever interacted with babies, you suspect that they’re up to something when they’re smiling. They’re not just smiling randomly,” says Movellan.
Cute or Creepy? Diego-San’s anamorphic noggin places him on the edge of the “uncanny valley,” the territory assigned to objects so lifelike they actually become unsettling. You can judge for yourself in the video below, but the bot is nevertheless the next step in the development of “emotionally relevant” robotics.
What you see here is RUBI-5. She’s gone through several designs, with early versions even using parts bought from IKEA.
|What you see here is RUBI-5. She’s gone through several designs, with early versions even using parts bought from IKEA.|
Creators: Deborah Forster, ’91, M.S. ’95, Ph.D. ’12, project scientist at the Qualcomm Institute | Javier Movellan | Machine Perception lab
Origins: RUBI was designed to study the development of young children, whose ability to speak is limited. “We’re working on understanding what it takes to have a natural interaction between robots and humans,” says Movellan. “We believe that to be useful to people, in educational settings or elsewhere, robots will have to get better at what humans do brilliantly without thinking—recognizing a voice, for example, or smiling back at just the right time.”
Superpowers: Equipped with a Computer Expression Recognition Toolbox (CERT), RUBI teaches preschoolers colors and shapes as well as songs. When she sees a child smiling, she giggles and encourages them in the lesson. This smile detection technology led to an algorithm derived from a dataset of more than 60,000 individuals; the algorithm was later developed into Shutter Smile Technology in Sony digital cameras.
Appearances: RUBI has been a hit with children at UC San Diego’s Early Childhood Education Center (ECEC), who love how RUBI giggles when tickled and flinches when bumped. “It’s an enriching experience for both the children and the teachers,” says Lydia Morrison, a retired ECEC teacher. “They help us imagine new ways of teaching.”
Shape-shifter: RUBI has undergone several redesigns—from being soft and plump with a computer monitor display to a boxy shape with flexible arms and a tablet
Einstein may be “head and shoulders” above other bots, but occasionally he’s given a body too. Check out his dance moves in the video below.
|Einstein may be “head and shoulders” above other bots, but occasionally he’s given a body too. Check out his dance moves in the video below.|
Creators: Javier Movellan | Machine Perception Lab | Hanson Robotics
Origins: The Einstein robot was created in order to better understand how humans and robots can perceive emotions. “When a robot interacts in a way we feel is human, we can’t help but react,” says Movellan. “Developing a robot like this one teaches us how sensitive we are to biological movement and facial expressions, and when we get it right, it’s really astonishing.”
Superpowers: Einstein’s facial recognition software, developed by Movellan and a team of graduate students at the Qualcomm Institute, is based on a series of computational algorithms derived from an analysis of more than 1 million facial images. It allows Einstein to understand and respond to a number of “perceptual primitives,” such as expressions of sadness, anger, fear, happiness and confusion. This kind of “empathy” allows it to interact with humans in a relatively natural, conversational way.
Composition: The robot is equipped with 31 motors—17 of which spring into action when it needs to smile. That smile is convincing thanks to a patented, flesh-like material called, Frubber™, or “flesh rubber,” complete with realistic pores that measure from 4 to 40 nanometers.
|Got a better name for this guy here? Leave a comment or let us know at #namethatbot|
The (Amazing) 3D-Printed Jumping Robot
Creators: Michael Tolley, assistant professor of mechanical engineering | Nicholas Bartlett, Ph.D. student in Harvard’s Microrobotics Laboratory
Origins: Researchers were inspired by the way some organisms’ bodies blend soft and hard materials. Mussels, for example, have a foot that starts out soft yet becomes rigid at the point of contact with rocks. Engineers brought this concept to the first-ever 3D-printed robot that transitions from a rigid core to a soft exterior. “Bringing together soft and rigid materials will help create a new generation of fast, agile robots that are more robust and adaptable than their predecessors, and can safely work side by side with humans,” says Tolley.
Composition: The robot is made of two nestled hemispheres, one of which has nine layers of stiffness that go from a flexible, rubber-like exterior to full rigidity near the core. Researchers struck a fine balance between hard and soft—where a fully rigid top would make for higher jumps, a more flexible top would survive impacts on landing, allowing the robot to be reused.
Superpowers: Powered by a mix of butane and oxygen, this bot is capable of more than 30 untethered jumps. In tests, the robot jumped six times its body height (or 2 ½ feet) and survived multiple falls from four feet—10 times its body height!
Creators: Jeffrey Friesen, ’13, Ph.D. student, Jacobs School of Engineering | Coordinated Robotics Lab
Origins: With Americans spending an estimated 90 percent of their time indoors, air quality is a major concern for public health. Researchers at UC San Diego’s Coordinated Robotics Lab developed the DucTT robot to navigate HVAC (heating, ventilation and air conditioning) systems better than current devices, which are similar to miniature street sweepers. DucTT was designed using principles of tensegrity, which uses components under pure tension and pure compression. “Tensegrity robots have the advantage of being light and flexible,” says Friesen. “DucTT is built from lightweight tubes and cables arranged to provide an extensive range of motion with a small number of motors.”
Superpowers: Where other devices are unable to access more remote parts of air systems, DucTT uses an inchworm-like movement to explore ducts more efficiently and achieve a range of motion that can negotiate the intersection of two or more ducts.
DucTT is funded by a grant from NASA, and is developed in collaboration with the tensegrity group at NASA/Ames Research Center.
The Many Modes of MiP
The latest products developed by the Coordinated Robotics Lab and its director, Thomas Bewley, aren’t just fun and games, but could revolutionize robotics education.
MiP is the original poster-robot proving that complex algorithms and innovative engineering can make for some serious fun. Short for Mobile Inverted Pendulum, MiP can balance itself and navigate on two wheels. “MiP’s organic nature is particularly engaging: when it stands, it gently sways back and forth; when pushed, it takes a step back to regain its balance. In a very real way, the dynamics of MiP mimic life,” says Bewley.
Superpowers: MiP is not only a master of balance, but uses this talent to be interactive and fun for all ages. Users can install a serving tray and test their skill at stacking games, or control the bot via a smartphone app to make special deliveries. Always the life of the party, MiP also responds to simple hand gestures and dances to the beat of built-in tunes or any music on the user’s mobile device.
Celebrity Status: MiP has hung out with LL Cool J, Lorde, Stan Lee and other celebrities at various Hollywood award ceremonies. The bot makes frequent cameos on television’s The Big Bang Theory, and is a fun fixture in millions of households worldwide.
Built to Build Upon: Robotics enthusiasts and coders rejoiced when MiP co-creator and manufacturer WowWee released a software development kit for the bot, allowing code-savvy users to program many other actions into MiP’s bag of tricks.
EduMiP & EduRover
The MiP-makers at the Coordinated Robotics Lab have not only mastered the intersection of technology and toymaking, but are on a mission to democratize robotics education. Within UC San Diego’s Mechanical and Aerospace Engineering department, they have developed an extremely popular Embedded Control & Robotics capstone course, during which each student builds, models, programs and stabilizes their own educational Mobile Inverted Pendulum (EduMIP). Students start from a bag of parts and an inexpensive credit card-sized Linux computer, while each step along the way drives toward the review and synthesis of key elements of modern engineering disciplines. This educational curriculum is currently being extended to incorporate a small four-wheel-drive car(EduRover) with extreme four-wheel steering.
Superpowers: Both EduMIP and EduRover can be modified in an infinite number of ways via 3-D printing and additional sensors, actuators and subsystems of the students’ choosing. As these systems are added on, EduMIP and EduRover become simply the chassis upon which budding roboticists realize their own visions for what small mobile robots can accomplish by leveraging situational awareness and feedback control.
Broad Impact: In partnership with WowWee, hardware kits for this course are in the process of being mass produced at very low cost. Once available on a large scale, these materials will become the core of a two-part course for aspiring roboticists in high school, college and beyond. Using a Massively Open Online Course (MOOC) format, the course and materials will teach the salient features of robotics systems to a broad audience worldwide. Says Bewley, “We look forward to empowering end users to achieve their own unique visions, and helping to inspire them to do something new—to use emerging technologies to make the world a better place. At its core, that is what this effort is all about.”