RoboCup-2000 Simulation League: Team KU-Yam2000

In this paper we described our KU-Yam2000 team that participated in the simulation league of RoboCup-2000 Melbourne. KU-Yam2000 is characterized by soccer agents that are controlled by a hierarchy of actions. Actions in each level of the hierarchy are sep

  • PDF / 101,534 Bytes
  • 4 Pages / 451 x 677.3 pts Page_size
  • 87 Downloads / 216 Views

DOWNLOAD

REPORT


1 Introduction Robot soccer is a candidate for the standard challenging problems in Artificial Intelligence. Our two teams, Team Miya and Team Niken, participated in the simulation league of RoboCup97 Nagoya (Japan, August 1997) [1]. Moreover, we sent Team Miya2 to the simulation leagues of RoboCup98 Paris (France, July 1998) [2] and Team KU-Sakura2 to Robocup99 Stockholm (Sweden, August 1999) [3]. Team Miya was characterized by individual tactical play[4]. Individual tactical play does not require communication between the players, so the speed of passing was rapidly increased in RoboCup 97 games, and the team sometimes behaved as if it had been taught some tactical plays. Team Miya proceeded to the quarterfinal match and was one of the best eight teams in the simulation league. In Team Miya2, a kind of communication between players was realized by using a "say" command so that a passer can make a pass to a receiver without looking around for receivers[5]. Consequently, Team Miya2 was one of the best sixteen teams in RoboCup98. However, more tactical play is required for the following two reasons. First, the top teams of RoboCup98 showed very high-level skill in individual play. For example, we observed a speedy dribble while keeping the ball near the player's body and a safety pass without being intercepted by the opponent players. Second, the offside rule was introduced at RoboCup98. Thus forward players have to check whether they are in an offside position or not at all times. Some tactics are necessary to avoid the opponent's offside trap and to succeed in setting an offside trap against the opponent team. For this purpose, we used communication between players for realizing the tactics in Team KU-Sakura2. We succeeded in moving forward players and defense players to a certain degree, which helped Team KU-Sakura2 to pass the qualifying round and proceed to the final tournament. However, our program became very complicated, and it seemed difficult to further improve individual skills in KU-Sakura2. Thus we decided to make a new team where all procedures are modularized to C-language functions after the RoboCup99 competition. Moreover, we improved the format of a player's world model and used it to estimate positions and velocities of the ball and all of the players. That is a basic concept in developing P. Stone, T. Balch, and G. Kraetzschmar (Eds.): RoboCup 2000, LNAI 2019, pp. 445-448, 2001. c Springer-Verlag Berlin Heidelberg 2001

446

Harukazu Igarashi et al.

Team KU-Yam2000.

2 Team Development We describe the development of our Team KU-Yam2000 in this section. Team Leader: Harukazu Igarashi, Associate Professor of Kinki University, Japan Team Members: Yusuke Yamauchi, Undergraduate Student, and Shuichi Iidoi, Graduate student of Kinki University, Japan. H. Igarashi attended the RoboCup2000 competition. WebPage:http://www.ip.info.hiro.kindai.ac.jp/igarashi/registeration/qualification.htm E-mail address: [email protected]

3 World Model Each soccer agent has a world model in Team KU-Yam2000