RoboCup Soccer

Introduction:

RoboCup is an international research and education initiative. It is an attempt to foster AI and intelligent robotics research by providing a standard problem where a wide range of technologies can be integrated and examined, as well as being used for integrated project-oriented education.

For this purpose, RoboCup chose to use soccer game as a primary domain, and organizes RoboCup: (The Robot World Cup Soccer Games and Conferences). In order for a robot team to actually perform a soccer game, various technologies must be incorporated including: design principles of autonomous agents, multi-agent collaboration, strategy acquisition, real-time reasoning, robotics, and sensor-fusion. RoboCup is a task for a team of multiple fast-moving robots under a dynamic environment. RoboCup also offers a software platform for research on the software aspects of RoboCup.

While the soccer game is used as a standard problem where a broad-range of efforts will be concentrated and integrated, competition is only a part of RoboCup activity.

Activities of the RoboCup consist of:

 

Objective:

RoboCup is used as a vehicle to promote robotics and AI research, by offering publicly appealing, but formidable challenges. One of the effective ways to promote engineering research, apart from specific application developments, is to set a significant long term goal. When the accomplishment of such a goal has significant social impact, it is called the grand challenge project. Building a robot to play soccer game itself, do not generate significant social and economic impact, but the accomplishment will certainly be considered as a major achievement of the field. This kind of project is called as a landmark project. RoboCup is a landmark project as well as a standard problem.

The ultimate goal of the RoboCup Initiative is: by mid-21st century, a team of fully autonomous humanoid robot soccer players shall win the soccer game, complying with the official rule of the FIFA, against the winner of the most recent World Cup.

This goal is proposed to be one of the grand challenges shared by robotics and AI community for the next 50 years. This goal may sound overly ambitious given the state of the art technology today. Nevertheless, it is important that such a long range goal is claimed and pursued. It took only 50 years from the Wright Brother's first aircraft to Apollo mission to send man to the moon and safely return them to the earth. Also, it took only 50 years, from the invention of digital computer to Deep Blue, which beat the human world champion in chess. Building humanoid soccer player requires equally a long period and extensive efforts of broad range of researchers, and the goal will not be met in any near term.

The RoboCup is designed to meet the need of handling real world complexities, though in a limited world, while maintaining an affordable problem size and research cost. RoboCup offers an integrated research task covering the broad areas of AI and robotics. Such areas include: real-time sensor fusion, reactive behaviour, strategy acquisition, learning, real-time planning, multi-agent systems, context recognition, vision, strategic decision-making, motor control, intelligent robot control, and many more.

 

Rules:

The regulations provided constitute official regulations of RoboCup games, administered by The RoboCup Federation. The initial regulations were drafted in 1994. Several revisions were made from technical and logistical point of view. The basis of the current regulation was drafted in the 1996 version of the RoboCup regulations, and modified each year for RoboCup World Championship games. Currently rule revision for RoboCup-2005 is under discussion.

 


Competitions:

RoboCup has different kinds of tournaments and each one has different categories and rules in it, namely:

RoboCup Junior is a project-oriented educational initiative that sponsors local, regional and international robotic events for young students. It is designed to introduce RoboCup to primary and secondary school children, as well as undergraduates who do not have the resources to get involved in the senior leagues. The focus in the junior league is on education, and we will not describe it further in this report.

 

History:

The idea of robots playing soccer was first mentioned by Professor Alan Mackworth (University of British Columbia, Canada) in a paper entitled "On Seeing Robots"  presented at VI-92, 1992, and later published in a book Computer Vision: System, Theory, and Applications, in 1993. A series of papers on the Dynamo robot soccer project was published by his group.

Independently, a group of Japanese researchers organized a Workshop on Grand Challenges in Artificial Intelligence in October, 1992 in Tokyo, discussing possible grand challenge problems. This workshop led to serious discussions of using the game of soccer for promoting science and technology. A series of investigations were carried out, including a technology feasibility study, a social impact assessment, and a financial feasibility study. In addition, rules were drafted, as well as prototype development of soccer robots and simulator systems. As a result of these studies, it was concluded that the project was feasible and desirable. In June 1993, a group of researchers, including Minoru Asada, Yasuo Kuniyoshi, and Hiroaki Kitano, decided to launch a robotic competition, tentatively named the Robot J-League (J-League is the name of the newly established Japanese Professional soccer league). Within a month, however, they received overwhelming reactions from researchers outside of Japan, requesting that the initiative be extended as an international joint project. Accordingly, they renamed the project as the Robot World Cup Initiative, "RoboCup" for short.

Concurrent to this discussion, several researchers had already been using the game of soccer as a domain for their research. For example, Itsuki Noda, at ElectroTechnical Laboratory (ETL), a government research center in Japan, was conducting multi-agent research using soccer, and started the development of a dedicated simulator for soccer games. This simulator later became the official soccer server of RoboCup. Independently, Professor Minoru Asada's Lab at Osaka University, and Professor Manuela Veloso and her student Peter Stone at Carnegie Mellon University had been working on soccer playing robots. Without the participation of these early pioneers of the field, RoboCup could not have taken off.

In September 1993, the first public announcement of the initiative was made, and specific regulations were drafted. Accordingly, discussions on organizations and technical issues were held at numerous conferences and workshops, including AAAI-94, JSAI Symposium, and at various robotics society meetings.

Meanwhile, Noda's team at ETL announced the Soccer Server version 0 (LISP version), the first open system simulator for the soccer domain enabling multi-agent systems research, followed by version 1.0 of Soccer Server (C++ Version) which was distributed via the web. The first public demonstration of this simulator was made at IJCAI-95.

During the International Joint Conference on Artificial Intelligence (IJCAI-95) held at Montreal, Canada, August, 1995, the announcement was made to organize the First Robot World Cup Soccer Games and Conferences in conjunction with IJCAI-97 Nagoya. At the same time, the decision was made to organize Pre-RoboCup-96, in order to identify potential problems associated with organizing RoboCup on a large scale. The decision was made to provide two years of preparation and development time, so that initial group of researchers could start robot and simulation team development, as well as giving lead time for their funding schedules.

Pre-RoboCup-96 was held during the International Conference on Intelligence Robotics and Systems (IROS'96), Osaka, from November 4 - 8, 1996, with eight teams competing in a simulation league and demonstration of real robot for middle size league. While limited in scale, this competition was the first competition using soccer games for promotion of research and education.

The first official RoboCup games and conference was held in 1997 with great success. Over 40 teams participated (real and simulation combined), and over 5,000 spectators attended. The number of teams and spectators has been growing successfully. In Bremen (2006, Germany) 15,270 people were present in the exhibition halls from 14 to 18 June and 2,613 of them (440 participating teams from 35 countries) took an active part in the contests.

The 12th RoboCup International Competitions and Conferences, RoboCup-2008 Suzhou is to be held on July 14th - July 20th, 2008, Suzhou, China. The RoboCup-2009 will be held in Graz, Austria.

 

Participation:

Place
Year
Participating teams
Countries/regions
Atlanta
2007
300
37
Bremen
2006
440
35
Osaka
2005
330
31
Lisbon
2004
346
37
Padova
2003
277
34
Fukuoka/Busan
2002
188
29
Seattle
2001
119
22
Melbourne
2000
110
19
Stockholm
1999
120
35
Paris
1998
63
20
Nagoya
1997
40
10

 

Further information about the competitions:

 

RoboCup Simulation League

RoboCup Simulation League consists of a number of competitions with simulated soccer matches as the main event.
Categories:

If the proposed ideas get accepted by the technical committee, they will be incorporated into the official simulator.
One of the goals for future 3D soccer competitions is to have simulated robots with articulated bodies, for example like humanoid robots.

 

Small Size Robot League

Small size robot soccer is one of the RoboCup league divisions. Small size robot soccer, or F180 as it is otherwise known, focuses on the problem of intelligent multi-agent cooperation and control in a highly dynamic environment with a hybrid centralized/distributed system

A small size robot soccer game takes place between two teams of five robots each. Each robot must conform to the dimensions as specified in the F180 rules, the robot must fit within a 180mm diameter cylinder. If a team is using the global vision system, each robot on that team must have height of 150 mm or less. In all cases, a robot must have height less than 225 mm. The robots play soccer on a green carpeted field that is 5m long by 3.5m wide with an orange golf ball. Robots come in two flavours, those with local on-board vision sensors and those with global vision. Global vision robots, by far the most common variety, use an overhead camera and off-field PC to identify and track the robots as they move around the field. The overhead camera is attached to a camera bar located 3m above the playing surface. Local vision robots have their sensing on the robot itself. The vision information is either processed on-board the robot or is transmitted back to the off-field PC for processing. An off-field PC is used to communication referee commands and, in the case of overhead vision, position information to the robots. Typically the off-field PC also performs most, if not all, of the processing required for coordination and control of the robots. Communications is wireless and typically uses dedicated commercial FM transmitter/receiver units although at least one team has used IRDA successfully.

Building a successful team requires clever design, implementation and integration of many hardware and software sub-components into a robustly functioning whole making small size robot soccer a very interesting and challenging domain for research and education.

 

Middle Size Robot League

Two teams of mid-sized robots with all sensors on-board play soccer on a field. Relevant objects are distinguished by colors. Communication among robots (if any) is supported on wireless communications. No external intervention by humans is allowed, except to insert or remove robots in/from the field.

 

 

 

 

 

Other challenges

Ball Control and Planning
Six to eight black obstacles (length/width 40 cm, height 60 cm) are put at arbitrary positions on the field. The ball is put on the middle of the penalty area line, and a robot inside the same goal. The robot should dribble the ball into the opposite goal within 90 seconds, while it avoids all obstacles. One point is awarded to the robot if the ball has passed the centre line, another point when a goal is scored.
Penalty points are given each time the robot or the ball touches an obstacle. The challenge is repeated three times with various setups. An extra point is awarded to the team with the fastest robot. In order to be eligible for this extra point the robot may not have touched any of the obstacles. In total a team can be awarded up to seven points for this challenge.

Cooperative behaviour
Teams should demonstrate cooperative behaviour between at least two robots. The selection of the activity to be performed is free, but it should last at most 90 seconds. A jury will evaluate the quality of cooperation and cooperative behaviour and will assign up to six points to each team.

Cooperative Mixed-Team Play
Teams should demonstrate cooperative mixed-team play between at least two robots from different teams. The selection of the activity to be performed is free, but it should last at most 90 seconds. A jury will evaluate the quality of cooperation and cooperative behaviour and will assign up to six points to each team.

Play with an arbitrary FIFA ball
The aim of this challenge is to encourage teams to improve their vision routines. This challenge is carried out with three different standard FIFA balls. A robot is placed on the field and the ball is placed in front of the robot for 5 seconds. Afterwards the ball is placed at an arbitrary position on the field. The robot has now 60 seconds to find the ball and to dribble it into a predefined goal. One point is awarded to the robot for correctly identifying the ball, i.e. the robot has found and touched the ball for the first time. A second point is awarded if the robot has scored a goal. In total this challenge is repeated three times with varying balls but always with the same robot. In total a team can be awarded up to six points for this challenge.

Play with arbitrary goals
A robot is placed in a random position within its own half of the field and a regular RoboCup ball is placed in a random position in the other half. The robot should kick the ball in the opposite goal (the one in the other half of the field with respect to its initial position). No assumptions should be made on the colors of the two goals. The organizer will cover the yellow and blue goals with a different color pattern (except for the red). The challenge is repeated three times with different initial positions of the robot and of the ball. Each attempt has a time limit of 60 seconds. One point is awarded to each score and an additional point is given to robots that score three times in less than 30 seconds each. In total a team can be awarded up to four points for this challenge.

Show scientific or engineering achievements
Teams are free to show one significant achievement each, and all the other team leaders, together with the TC members will judge them. Achievements in the present list are encouraged.

Passing
At the beginning, the ball is on the own penalty kick point. A player (#1) is on the center point. Another player (#2) is in the opposite half field. A still goalie (a black box, 50 cm wide) is in the middle of the opponent goal. Player #1 has to get the ball, pass it to player #2 which should score the goal. 1 point if the ball passes midline, 1 point if it is dribbled by player #1 over midline, 3 points if it hits player #2, 3 points if player #2 tries to score in the opponent goal (either by kicking or bringing the ball there, always following the rules), 2 points if the goal is scored. Time is also taken and used to rate the teams in case of same number of points.

Four-Legged Robot League

One of the ultimate dreams in robotics is to create life-like robotics systems, such as humanoid robots and animal-like legged robots. Quadruped robots have major potential for future robotics, including entertainment applications for personal robots. However, a number of challenges exist before any such robot can be fielded in the real world. Robots have to be reasonably intelligent, maintain a certain level of agility, and be able to engage in some collaborative behaviours.RoboCup is an ideal challenge to foster robotics technologies for small personal and mobile robotics systems. Through the research and competition in this league, technology will be developed and will improve robot performance not only in entertainment but also in non-entertainment applications such as rescue robots and other dangerous jobs currently performed by humans, and that robots will perform these tasks in the future by working together. In the league teams consisting of four Sony Aibo robots each play on a field of 5.4 m x 3.6 m. The robots operate fully autonomously, i.e. there is no external control, neither by humans nor by computers.

The Open Challenge
This challenge is designed to encourage creativity within the Legged League, allowing teams to demonstrate interesting research in the field of autonomous systems. Each team will be given three minutes of time on the RoboCup field to demonstrate their research. Each team must distribute a technical description of their research (1-2 pages) before the round-robin starts. Teams who do not submit a description or propose a non-technical challenge will be ineligible to compete. The winner will be decided by a vote among the entrants. In particular:

The winner will be decided by a vote among the entrants using a Borda count (http://en.wikipedia.org/wiki/Borda_count). Each entering team will list their top 10 teams in order (excluding themselves). The teams are encouraged to evaluate the performance based on the following criteria: Technical strength, novelty, expected impact and relevance to RoboCup. At a time decided by the designated referee, within 30 minutes of the last demonstration if not otherwise specified, the captain of each team will provide the designated referee with their rankings. Each ranking is converted to points: ten points for the top ranked team, nine for the team ranked second and so on down to one point for the team ranked tenth. Any points awarded by a team to itself will be disregarded. The points awarded by the teams are summed and the team with the highest total score shall be the winner.

The Passing Challenge
This second challenge is intended to encourage teams to develop passing and catching skills. In this challenge each team will be required to provide three robots; all robots must be in the same coloured uniform (the decision on red or blue uniforms can be made by each team).
Each robot will be placed on the field inside a circle of radius 35cm. The centre of the circles will be no closer then 80cm and no further then 200cm apart. The triangle formed by the circles will not be equilateral, i.e. the distances between robots will be different.
Initially the robots will be in the ‘set’ state for 15 seconds, this will enable them to localise. The robots will then be placed into ‘playing’ and given two minutes to pass the orange ball around. A pass will be regarded as successful when:

Robots may pass between each other in any order, but will be rewarded for passing to a different robot then that which passed to it.
All normal game rules apply in the challenge, except when a ball leaves the field it will be replaced back in the closest circle. If a rule is violated then any pass resulting from this violation will receive no points.

The New Obstacle Avoidance Challenge
The purpose of this challenge is to successfully solve a series of problems. In summary robots must navigate from one goal to the other goal while avoiding obstacles and performing tasks.
This challenge will require the team to use two robots, one red and one blue. The blue robot will start in the blue goal while the red robot will start in the yellow goal. The task of each robot is to reach the opposing goal. "Reaching" means to have at least two feet in the opposite goal (that is two legs behind the goal line). Along the way the two robots will face a series of obstacles. The obstacles may include:

At the middle of the field the two robots will be asked to perform a simple cooperative task. In this case both robots will have to 'cross' the halfway line at the same time. The rule will be defined as 'the two robots can not be in the same half of the field for more than 3 seconds'.

 

Humanoid League

The Humanoid Leagues are special within RoboCup. The human-like embodiment of the robots leads to extraordinary tasks and challenges that are different from the other RoboCup leagues. In addition, humanoid robots are essential to fulfil the aim of beating the human world champion in soccer by the best humanoid team in the year 2050. Humanoid robots show basic skills of soccer players, such as shooting a ball, or defending a goal. Relevant objects are distinguished by colors. External intervention by humans is allowed, as some of the humanoid robots are tele-operated.

Strategy for the next years:

The robots are diverse: The size of humanoid robots that participated since the start of the league varies from 10cm to 2m. Robots of so different sizes can hardly play soccer together. For the next years one important aim is to bring the robots together in one league. The number of leagues is 2: Kidsize and Teensize.
Robots participating in the Humanoid League competitions must have a human-like body plan, as shown in the figure. They must consist of two legs, two arms, and one head, which are attached to a trunk. The robots must be able to stand upright on their feet and to walk on their legs. The only allowed modes of locomotion are bipedal walking and running.

The Footrace

n the TeenSize class, the footrace replaces soccer games until enough TeenSize teams are able to play soccer games. The footrace is done as a 1 vs. 1 competition between two teams. It takes place on a TeenSize field. Access to the field is given to both teams at least 10 minutes prior to the scheduled starting time.
A footrace match consists of five runs. A run goes from one touch line to the other touch line. Both teams place their robot in front on the border strip, outside the field, in front of the touch line. The robots must be in an upright standing posture. One team uses the yellow side of the field and the other team uses the blue side. Different robots might be used for multiple runs.
After the referee gives the start signal, both robots walk as fast as possible across their half of the field towards the opposite touch line. Robot handlers are not allowed to enter the field, unless the referee asks them to remove a robot.
If a robot touches the start line before the referee gave the start signal, the start is invalid. This robot receives a warning and the start is retaken. The warnings of a team accumulate within a match. Every third warning results in the technical winning of a run for the other team.
The robot which first crosses the goal line wins the run. Both feet must be outside the field on the border strip again. If a robot falls, it must get up by itself to continue the run.
If a robot leaves its half of the field, its run terminates at the position where it left the field. If 60s after the start signal no robot crossed the goal line, the referee decides which robot advanced more towards the goal line. This robot wins the run. If the referee cannot determine which of the robots advanced more the run ends in a draw. Both teams must place their robot in front of the start line within 60s after the end of a run.
The team which wins most of the runs wins the footrace match. If both teams win an equal number f runs, the match ends in a draw. To decide knock-out matches, the runs are continued until one teams wins a run.


Challenges