ARS: You were the main organiser of Robocup 2004, which was held on the Instituto Superior Tecnico (your home institute) in Lisbon, Portugal in a period 27 June - 5 July 2004. Tell us on the beginning, are you satisfied with whole manifestation?
Lima: First, let me correct you: I and Prof. Luis Custódio, also from the Institute for Systems and Robotics (ISR) at Instituto Superior Tecnico (IST), were the main organizers (General Chairs) of the event. ISR is a research institution, while IST is the higher education faculty where we teach. Answering your question: definitely, RoboCup2004 went pretty well and was a great success, recognized by all the participants and the Trustees of the RoboCup Federation. We had the largest number of participants so far (about 1600, distributed by 330 teams from 37 nations) and our emphasis on the science that RoboCup fosters and on showing it to the public was underlined by the Trustees, corresponding to our main goal.
ARS: How did your institute profit with the Robocup?
Lima: ISR and IST got an even higher visibility at the National level, appearing as the leading Robotics research and Engineering higher education institutions in Portugal, respectively. At the international level, this was yet another step towards showing to the world leading scientists in Artificial Intelligence and Robotics our capacity to organize important international events such as RoboCup, and also to show them the research actually going on at ISR/IST. We have been developing this characteristic of ISR/IST in recent years, by hosting an increasing number of PhD students and several international conferences (IEEE MED2002, IFAC IAV2004, 2003 EURON annual meeting).
ARS: Where there any major team surprises, like Greece in Euro 2004?
Lima: Well, we do not usually think of it that sports-like way. The major "surprises" come typically from the advances in research shown, especially when they lead to better results for a team in the competitions, or provide innovative methodologies described in the Symposium. Some teams bet mainly in technological developments. All these are important in RoboCup. Some points to retain this year: For the first time in RoboCup history, teams played in the Middle-Size League (MSL), Small-Size League (SSL) and Humanoid League (HL) without special spot lights on the top of the field, just the pavilion lights. This did not lead to any major degradation of game quality, which per se is an indicator of the technological and scientific progresses of RoboCup real robot leagues on the perception side. In the SSL, larger fields reduced the robots speed, leading to a less messy game and very interesting game situations where cooperation and robot formations were clearly demonstrated. In the HL, eye-catching advances in the robot capabilities and autonomy clearly point towards the possibility of having one-on-one games very soon (next year?). The Real Rescue League made a significant advance in the information provided to the audience as the robots evolve, showing the images they see, and the map they are building, therefore providing anchors for the understanding of what is actually going on and what are the goals of the "competition".
ARS: From which continent are coming the major players in Robocup?
Lima: Probably from Asia, mainly from Japan. The Japanese always bring a lot of (winning) teams, across all the leagues. However, this year, an European country has shown its brightest face, with the largest number of Symposium papers and several good results in the competitions: Germany. One interesting feature of German teams is the multi-University development of software modules which are then shared by the different teams and across the leagues (e.g., a localization algorithm software module can be used by two different German teams in the MSL, but also by another one in the 4-Legged League). This is definitely the way to go in RoboCup, in order to foster continuous advances. Generally, Europe brings also many teams to RoboCup, from several countries (Portugal amont the largest represented).
ARS: How do you see the role of non-developed countries in robot soccer, or generally in robotics research area?
Lima: One major problem that countries with less funding for research may face is the high cost of Robotics hardware. However, there are a lot of opportunities out there: the simulation leagues allow low-cost participation, with the development of advanced "higher-level" algorithms for planning, decision and coordination. Students and faculty from these countries may well engage in cooperative research programs with groups from other country universities, so that they can be part of joint multi-University/multi-country teams. And entry-level competitive robots are becoming cheaper to build by students at the University level, starting from kits or other groups' experience.
ARS: In which direction in your opinion is going to develop Robot Soccer? Is it realistic that in 2050 humanoid robots could play against the humans?
Lima: Nobody knows if they will, but it definitely seems realistic, given the advances in the humanoids locomotion and in the other leagues sensing, real-time decision-making, teamwork, etc. I believe Robot Soccer will continue displaying games with increasing cooperation among robots, faster decision-making without loss of quality, as well as improved perception and, in general, sensor fusion (e.g., to handle variable light conditions and outdoor environments), across all the leagues. In the legged leagues, particularly the HL, serious advances in biped locomotion are expected. But, above all, methodologies for the integration of all these subsystems in a way re-usable for other cooperative robotic and multi-agent system applications will be (already is!) RoboCup's major breakthrough.
ARS: Why 2 leagues (FIRA and Robocup)? Wouldn't be better two join all Robot soccer researchers in one society?
Lima: I honestly do not know the reasons why these two groups exist separately, but these divisions always happened in research communities, corresponding to different approaches to similar problems, and this is not something we should care too much about, I think.
ARS: Could you tell us something about your institute? What are your major research areas?
Lima: ISR/IST is a non-profit research institution that promotes advanced multidisciplinary R&D in Robotics and associated areas. Quoting our web page (http://www.isr.ist.utl.pt): Ever since its foundation, and in close cooperation with IST, ISR-Lisbon has been committed to meeting the following main strategic objectives: " promote scientific research and technological development in Robotics and associated areas; " promote active intellectual collaboration among researchers with varied backgrounds and perspectives acquired in different kinds of science (e.g., experimental, computational and theoretical), different sectors (university, industry, governmental and regional administration) and different regions; " provide education and research experience for graduate and undergraduate students, post-doctoral researchers, and industrial fellows, providing exposure to leading-edge research and introducing the students to large-scale collaborative research ventures; " promote the diffusion of scientific results through publications and by organizing seminars, conferences, exchange visiting programs and scientific meetings at a national or international setting; " increment the scientific exchange among Portuguese Universities and these and other foreign Universities and similar institutions, through the exchange of researchers and students, and the participation in joint ventures; " promote and support integrated graduate programs leading to Master and Doctor degrees; " promote the participation in research projects with national or international Universities, R&D institutions and industrial companies; The major research areas are Robotic Manipulation and Mobile Autonomous Robotics (Land, Underwater and Space autonomous vehicles), Automation and Control, Dynamic Systems, Signal and Image Processing, Communications, Computer Vision, Biomedical Engineering, Evolutionary Systems, Artificial Intelligence, Manufacturing Systems and Aeronautics. In the particular case of the Intelligent Systems Laboratory that I co-coordinate with Prof. Luis Custódio, our research concentrates on blending Artificial Intelligence and Systems Theory to create innovative methodologies for Multi-Robot Systems.
ARS: Are there any commercial results?
Lima: As a non-profit research institution, commercial results are not among our goals, even though they can appear as side-effects of our work. Some examples are the development of about 4 Small Companies as spin-offs of ISR/IST, in the areas of Educational, Research and Service Robots and Vision Systems, with a moderate success so far.
ARS: What do you believe are future directions of robotics generally, particurarly robot soccer?
Lima: Robotics is turning into the service "arena", after its major focus on Industrial Robots in the past. The latter will still play a major role in upcoming years, of course, but I believe the real research advances will come from the services (vacuum-cleaning, lawn-mowing, inspection, surveillance, active sensor networks, personal robots, etc) applications, as they bring new problems, especially integration, where the whole systems is the important issue, instead of its sub-systems (e.g., very accurate self-localization or complex world modelling may not be as important as having a robot team achieving its task goal(s), even if in a less precise way or with less repeatability). Robotic Soccer is a nice testbed to attract worldwide researchers to the above problems, and will certainly continue its development as such in upcoming years.
Institute for Systems and Robotics, Instituto Superior Tecnico, Lisboa, Portugal
Published in: Volume 1, Number 3, September 2004
For further informations in German please click here!