Three UQ students spent four weeks at the University of California, San Diego (UCSD) on sponsored trips to work on robotics projects.
QURPA-funded (Queensland Undergraduate Research Projects Abroad) mechatronic engineering student Joshua Riddell and UCSD Temporal Dynamics of Learning Center-sponsored PhD students Kristyn Hensby (Developmental Psychology) and Nikodem Rybak (Complex & Intelligent Systems, School of ITEE), were in San Diego from early January 2016.
They worked on TDLC research projects, culminating in attendance at TDLC’s All Hands Meeting where they presented posters on their work.
Kristyn worked on the RUBI project at UCSD’s Early Childhood Education Centre. She programmed social robot RUBI to build several toys and then tested how well RUBI could teach one to two-year-old children how to build these toys.
Josh also worked on the RUBI project assisting in movement sequencing for Kristyn's toy building experiment. He programmed an easy to use interface for scheduling and running speech utterances and changing facial expressions in synchronisation with RUBI's physical movements. He also strengthened RUBI's arm joints to suit the new mechanical challenge this experiment presented.
This study will continue at UQ with social robot Opie, who is currently under development at the School of ITEE.
RUBI and Opie belong to a new international network of social robots targeted to work with children of varying ages, with the aim of enhancing learning from technology, by making play with technology a social act.
Nikodem also worked on RUBI, joining UCSD’s Machine Perception Lab to work on the robot’s new type of sensors.
“I wanted to gain some experience in Active Object Recognition (AOR) which refers to problems in which a robot interacts with the world and controls its sensor parameters (e.g., camera orientation, gain, sensitivity) to maximise the speed and accuracy with which it recognises objects,” said Nik.
The three students also worked within UCSD’s Cognitive Development Lab, which has collected a very large data set of video and audio concerning various interactions between mothers and children.
“The aim of our work was to apply machine learning techniques, such as deep neural networks, to this data so we can extract events, such as emotion, from the speech. This would massively decrease the monetary cost and amount of time it takes to retrieve quantifiable data,” said Nik.
The students’ visit continues a long tradition of exchanges between TDLC and UQ. Josh, Nik and Kristyn’s connection with UCSD will continue through participation in regular online meetings.
“I am very happy that our group at UQ, Complex & Intelligent Systems, will be collaborating with UCSD in the field of deep neural networks,” said Nik.
“UCSD has a lot to offer in terms of research in artificial intelligence and robotics. It’s one of the best places in the world to work with experts in these fields, so I would definitely recommend an internship at UCSD to all students who want to pursue a career in AI and robotics.”