Lessons from the SPL for Humanoid leagues

Abstract

Standard platforms in robotics competitions have several advantages. As well as easing the cost of entry, having similar hardware makes the sharing of software very easy. The RoboCup Soccer Standard Platform league has progressed rapidly because teams can take the best ideas from across the league to build high performing robots. The increasing popularity of ROS now makes sharing of software somewhat easier even with different hardware, so some of the advantages of the SPL can be carried across to other leagues. The rUNSWift team has been porting its software base to ROS2 so that we can take advantage of the many libraries available and to ease the transition to new humanoid robots, not only for soccer, but also RoboCup@Home and potentially other leagues, as new and powerful hardware platforms are becoming available. This talk will give an overview of our experiences wth software sharing in the SPL and the transition to ROS2.

Biography

Claude Sammut is a Professor of Computer Science and Engineering at the University of New South Wales, Sydney, Australia. He leads the Robotics Research Group, which focusses on methods that combine classical AI with subsymbolic AI, including machine learning, knowledge representation and reasoning and cognitive robotics. His early work on relational learning helped to the lay the foundations for the field of Inductive Logic Programming (ILP). With Donald Michie, he also did pioneering work in Behavioural Cloning. His current research is developing machine learning techniques for robotics. He was leader of the UNSW teams that won the RoboCup Standard Platform League competitions in 2000, 2001, 2003, 2014 and 2015 and in the RoboCup Rescue Robot competition, his team won the award for best autonomous robot in 2009, 2010 and 2011.  He is a past president of the RoboCup Federation and was the chair of RoboCup 2019 in Sydney. He is co-editor-in-chief of Springer’s Encyclopedia of Machine Learning and Data Mining.