ZJUNlict
Team Description Paper for RoboCup 2014
Yue Zhao, Rong Xiong, Hangjun Tong, Chuan Li and Li Fang
National Laboratory of Industrial Control Technology
Zhejiang University, P.R.China
rxiong@iipc.zju.edu.cn
http://www.nlict.zju.edu.cn/ssl/WelcomePage.html
Abstract.
The Small Size League is one of the most important events
in RoboCup. The league is devoted to the advancing in mechanical de-
sign, artificial intelligence and multi-agent cooperation of mobile robots.
ZJUNlict from Zhejiang University has participated in this league for ten
years since 2004 and got the first place in RoboCup 2013. In this paper,
we introduce the main ideas of the robots’ hardware design of ZJUNlict,
and emphasize the intelligent control system including the hierarchical
architecture of strategy, the Lua script architecture, learning and select-
ing trajectories methods based on Dynamic Movement Primitives.
1
Introduction
Small Size League (SSL) is an important part of the RoboCup event. It is the
fastest and most intense game in RoboCup’s soccer competitions. The basic
rules of SSL are based on the rules of a FIFA’s soccer game, but each team
consists of only six robots playing on field that is 6.05m long by 4.05m wide.
There are two cameras mounted over the field to capture images, which are
processed in real time by a shared vision system, SSL-Vision [1]. This vision
system recognizes and locates the position and orientation of the robots and
the position of the ball, then broadcasts the information package to each team
via network. Of course, the objective of the game is to score more goals than
the opponent. SSL emphasizes on the fast movement of robots in a complex,
dynamic and competitive environment.
ZJUNlict from Zhejiang University has participated in this League for ten
years since 2004. We won the regional competitions in RoboCup ChinaOpen
2006, 2007, 2008, 2011. In RoboCup world-wide competition, we run into the
quarter finals in 2006, 2009, 2011, and got the semi finals in 2007 and 2008.
Since 2012, we have made a great progress and received the second place in
RoboCup2012, and won the championship in RoboCup2013.
The remainder of this paper is organized as follows. Section 2 briefly intro-
duces the hardware design of ZJUNlict’s robot. Section 3 introduces the intelli-
gent control system including strategy selection and trajectory generation based
on learning method. We present a play script writing in Lua and the results of
the DMPs’ execution in section 4, and section 5 concludes the paper.