Chance to teach a robot new skills

Professor Kerstin Dautenhahn and her team at the university’s School of Computer Science, who are developing robots as personal companions, will present KASPAR II, a child robot, which is learning autonomous social skills. The robot originated from KASPAR I, which was developed to work with children with autism.

At the STRI Showcase, which will take place at the university from 5-7 March, participants will be invited to drum with KASPAR II and encourage it to decide of its own accord what actions it should take next. This is part of the 4-year European Project Robotcub which investigates how a robot can “grow up” and develop behaviours and interact with people.

“We are using this robot to answer the basic research question of how can robots develop social skills autonomously,” said Professor Dautenhahn.

The researchers use various probabilistic models to enable KASPAR II to decide when it is appropriate to start drumming and when it should stop. The study also investigates which of these computational models people prefer in order to play an enjoyable drumming game with the robot.

“This is all part of developing robots as social companions,” added Professor Dautenhahn. “Once the robot develops further, it will be able to gauge the human’s mood and intention and act accordingly.”

The STRI Showcase, which will be held at the Weston Auditorium, de Havilland Campus from 5-7 March 2008, will demonstrate how technology enhances our world. As well as providing a chance to meet robots, it provides an opportunity to debate the big questions facing astrophysics, witness research on climate change, understand how biosensors could improve security and learn about the wonders of 21st century materials. There will also be an Open Evening at Bayfordbury Observatory on Friday 7 March. For further information about the Showcase, please visit: go.herts.ac.uk/strishowcase

A press reception will be held from 10.30-12pm on Thursday 6 March during which journalists can visit the exhibition stands and interview the researchers.