In this article, we introduce a visual debugger that helps us develop soccer agents for RoboCup soccer 3D simulation. The visual debugger enables us to graphically monitor the internal state of a soccer agent and the soccer field, such as joint angles, the position of objects, and text messages. We employ a server/client framework where the debugger acts as a server while the agent acts as a client. A soccer agent connects to the debugger using TCP/IP and sends information about the field and the internal status of the agent. The information that is sent from the soccer agent to the visual debugger consists of three parts: visible objects on the soccer field, the joint angles of the soccer agents, and text messages from the agents. These are shown as separate components on the screen of the debugger. The debugger draws the current pose of the soccer agent from the information on joint angles that are sent from the soccer server. Text messages are used as a debugging message. The developer of soccer agents is allowed to check if the developed agent works properly through the screen of the visual debugger. A soccer agent that is manually controlled using a game pad is also included as part of the debugger. Each of the above features is explained in detail.
[1]
Hisao Ishibuchi,et al.
A Fuzzy Reinforcement Learning for a Ball Interception Problem
,
2003,
RoboCup.
[2]
Oliver Obst,et al.
Simulation League: The Next Generation
,
2003,
RoboCup.
[3]
Peter Stone,et al.
A Neural Network-Based Approach to Robot Motion Control
,
2007,
RoboCup.
[4]
Luís Paulo Reis,et al.
Situation Based Strategic Positioning for Coordinating a Team of Homogeneous Agents
,
2000,
Balancing Reactivity and Social Deliberation in Multi-Agent Systems.
[5]
Enrico Pagello,et al.
Balancing Reactivity and Social Deliberation in Multi-Agent Systems
,
2001,
Lecture Notes in Computer Science.
[6]
Peter Stone,et al.
Layered learning in multiagent systems - a winning approach to robotic soccer
,
2000,
Intelligent robotics and autonomous agents.
[7]
James A. Hendler,et al.
Co-evolving Soccer Softbot Team Coordination with Genetic Programming
,
1997,
RoboCup.