RADIUS: Looking for Robots' Help in Computer Science Research and Education

by Jozef Kelemen and Ales Kubík

RADIUS is a project at the Institute of Computer Science at the Silesian University in Opava, Czech Republic, which reflects the shift from viewing computation as an execution of a sequence of basic computational steps, towards viewing it as the activity of a community of interacting basic computational entities. The acronym RADIUS is inspired by the robot of the same name from the play 'R.U.R' (Rossum's Universal Robots) by the Czech writer Karel Capek.

RADIUS, based on the experience of the project participants in such varying scientific areas as formal language theory, artificial intelligence, artificial life, natural computing, multi-agent systems and agent-based economies, reflects this paradigmatic change in two ways; research in robotics and decentralised intelligence underlying systems, and the incorporation of ideas from studies in this field into the computer science and applications curricula.

Our results and experiences in the theory of grammar systems (GSs), a recent area of formal language theory, have proven that formal languages can also be described and studied in terms of (finite) collections of grammars rather than individual grammars. This theory provides us with an effective framework for research in fields such as multi-processing, decentralisation, cooperation and emergence of collective phenomena, and can thus be viewed through the lens of the above new computational paradigm.

In addition to research, grammar systems as a part of Masters and doctoral curricula have been presented by Alica Kelemenová and Jozef Kelemen at Czech and other European universities (Silesian University, Opava, University at Hradec Králové, both Czech Republic, University of Economics at Bratislava, Slovakia, Vienna University of Technology, Austria, and Rovira i Virgili University, Tarragona, Spain). A tutorial was delivered by Jozef Kelemen on the connection of a new, non-functional modularisation of systems and grammar systems at the 9th ECCAI Advanced Course on AI in Prague, 2-13 July 2001.

We have supervised Masters and PhD theses in GSs; for example, a formal framework and computer simulation as a tool for experimentation have been used in order to prove emergence of some phenomena in agent-based computational economies in the PhD dissertation of the second author of this article.

Besides investigating multi-agent systems as grammar systems, we have experience with teaching agent-based software engineering and the theory and practice of multi-agent systems as a regular Masters-level university course. A collection of lecture notes prepared by Ales Kubík and published by the Silesian University for this purpose was the first comprehensive introductory textbook on the topic published in the Czech language.

In spring 2002, we set up a laboratory for experimentation in information-processing aspects of collective robotics. It is equipped with K-team Khepera robots. We study societies of robots as open information systems of physically embodied agents executing tasks in a shared dynamic environment. For programming the robots we use C and Java, with a Lisp interface for controlling them remotely.

RADIUS is focused on studying the emergence of complex behaviours from the interactions of simple agents, in order to make emergence more tractable through the design and implementation of multi-agent architectures on the theoretical ground provided by grammar systems. Proven capabilities of formal tools can thus be experimentally tested in real environments. In conjunction, we study how the behaviour of a group of robots can evolve from an initial specification, and how to design robot controllers using neural nets, interactive Turing machines, biologically inspired computing etc.

Their modularity and easy-to-use interface make Kheperas suitable for teaching distributed programming and the information-processing base of robotics, and for teaching basic computer programming in a non-traditional manner, stressing the ideas of Stein's new modularity rather than the traditional functional modular decomposition. By programming robots, students must interact with autonomous units that are driven by the controller, with several processes with 'endless' loops that compete for the robot control. Students face the problem of programming real devices equipped with sensors and effectors which interact with the dynamic environment. This helps them to realise not only the particular characteristics of programming open information systems, but the limitations - which from our perspective is the most important point - of the object-oriented analysis and design of information systems in comparison with agent-based analysis and design of distributed applications or even robotic societies.

Figure 1: The Khepera robots.	Figure 2: The authors playing with the robots.

It is intended that RADIUS will continue to be used in the formulation and solution of further problems of new modularity, collective robotics, artificial life, computational economies and other fields within the area of grammar systems. It will also find employment in the experimental verification of emerging hypotheses (we plan to build our own peripherals in order to enhance the sensory/effectory capabilities of robots), and in incorporating new ideas into our programming courses, making them more effective, attractive and interesting for our students.

Looking forward to further common work in order to achieve these goals, we acknowledge the long-term, intensive and fruitful cooperation with our partners, especially with the Research Group on Modelling Multi-Agent Systems, SZTAKI, and the above-mentioned universities.

Link:
http://ui.fpf.slu.cz

Please contact:
Jozef Kelemen
Silesian University, Opava, Czech Republic
Tel: +420 653 684 361
E-mail: kelemen@fpf.slu.cz

Ales Kubik
Silesian University, Opava, Czech Republic
Tel: +420 653 684 367
E-mail: kubik@fpf.slu.cz