Desktop Robot Soccer

Abstract

Robot soccer pits teams of fast-moving robots in a dynamic environment (Sng et al., 2002).
Robot soccer fosters AI and intelligent robotics research by providing a standard problem
where a wide range of technologies can be integrated and examined (Asada & Kitano, 1999).
Today two international robot soccer federations, RoboCup (RoboCup, 2007) and FIRA
(FIRA, 2007), organize competitions in an eclectic range of categories. Those competitions are accompanied with technical conferences. The first international robot soccer tournament MiroSot'96 was held at Korea Advanced Institute of Science and Technology (KAIST), in November, 1996. At the time of writing, we can count more than ten different robot soccer leagues from RoboCup and FIRA.
Taxonomy of the robot soccer leagues could start with the vision system used. The global
vision group contains all the leagues that allow a global vision system (camera that gives an eye-bird view of the playing field). The image processing is done on a PC that controls the robots via a radio link. Whereas the local vision group contains all the leagues that require the vision processing to be done on the robots themselves. In this second group, the robots achieve a higher level of autonomy. Only wheeled robots are used in the global vision group. Whereas, the local vision group can be subdivided into wheeled robots and legged robots. Finally there are simulation leagues that provide a test bed for multi-agent research for those who do not have access to real robots.
Robot soccer not only stimulates robotic research, but also provides a platform for
computational intelligence education that allows the development of engaging
undergraduate level assignments. However, there are several limiting factors for the
widespread use of robot soccer as a research platform or a teaching tool. Most robot soccer leagues like the popular RoboCup Small Size and FIRA Mirosot leagues require a large playing field and a team of several postgraduate students to build the hardware and
develop the complex software. The least resource-demanding robot soccer league is the
simulation league. Unfortunately, by its very nature, this league does not provide the
invaluable experience of real robots. With the constraint of using real robots, the least
resource-demanding robot soccer league is arguably the FIRA KheperaSot league. This
league represents Desktop Robot Soccer, in the sense that the playing field fits on a desktop or a computer laboratory bench.
The regulations of the KheperaSot league impose size restrictions on the robots. The size
limitation lowers the entry barrier for participants in relation to other robot soccer
tournaments, making it more accessible to individuals and small teams with modest
funding and infrastructure support. The size limitation also poses challenge for hardware
technology. It pushes the limits of how much processing and sensing can be put into the
small package at a reasonable cost. However the size is not as small as to requiring
miniaturisation technology beyond the reach of standard electronics and construction
techniques. The KheperaSot league was the first fully autonomous robot soccer league of
FIRA.