Introduction

The ultimate goal in AI, and probably in robotics, is to build intelligent systems capable of displaying complex behaviors to accomplish the given tasks through interactions with a dynamically changing physical world. Traditional AI research has been mainly pursuing the methodology of symbol manipulations to be used in knowledge acquisition and representation and reasoning about it with little attention to intelligent behavior in dynamic real worlds [Brooks1991]. On the other hand, in robotics much more emphasis has been put on the issues of designing and building hardware systems and their controls. However, recent topics spread over the two areas include design principles of autonomous agents, multi-agent collaboration, strategy acquisition, real-time reasoning and planning, intelligent robotics, sensor-fusion, and behavior learning. These topics expose new aspects with which traditional approaches seem unable to cope.

In coping with these issues and finally achieving the ultimate goal, physical bodies play the important role of bringing the system into meaningful interaction with the physical environment - complex and uncertain, but with an automatically consistent set of natural constraints. This facilitates the correct agent design, learning from the environment, and rich meaningful agent interaction. The meanings of ``having a physical body'' can be summarized as follows:

1. Sensing and acting capabilities are not separable, but tightly coupled.
2. In order to accomplish the given tasks, the sensor and actuator spaces should be abstracted under resource-bounded conditions (memory, processing power, controller etc.).
3. The abstraction of the sensor and actuator spaces depends on both the fundamental embodiments inside the agents and the experiences (interactions with their environments).
4. The consequence of the abstraction is the agent-based subjective representation of the environment, and it can be evaluated by the consequences of behaviors.
5. In the real world, both inter-agent and agent-environment interactions are asynchronous, parallel and arbitrarily complex. There is no justification for adopting a particular top-down abstraction level for simulation such as a global clock 'tick', observable information about other agents, modes of interaction among agents, or even physical phenomena like slippage, as any seemingly insignificant parameter can sometimes take over and affect the global multi-agent behavior.
6. Natural complexity of physical interaction automatically generates reliable sample distributions of input data for learning, rather than from an a priori Gaussian distribution in simulations which does not always correctly capture the characteristics of the system.

Even though we should advocate the importance of ``having a physical body,'' it seems required to show that the system performs well coping with new issues in a concrete task domain. In other words, we need a standard problem which people regard as a new one that expose various various aspects of intelligent behaviors in real worlds.

RoboCup (The World Cup Robot Soccer Games:[Kitano, et al., 95, Kitano, et al., 97]) is an attempt to promote AI and robotics research by providing a common task for evaluation of various theories, algorithms, and agent architectures, and was proposed as a new standard problem. Not only the integration of a wide range of technologies but also accomplishment of technical breakthroughs are required for the robots to play a soccer game reasonably well. RoboCup consists of three competition tracks: (1) Real Robot League: using physical robots to play soccer games, (2) Software Robot League: using software agents to play soccer games on an official soccer server over the network, and (3) Expert Robot Competition: competition of robots which have special skills, but are not able to play a game.

In this paper, we propose the RoboCup Physical Agent Challenges as new research issues of physical agents. First, we show how the challenge is significant for physical agent research with long range issues. Then, we show mid- and short-term issues which spans from simple skill acquisition to a simple teamwork behavior. In particular, we pick two single agent skills (ball moving and ball catching) and one cooperative skill (passing the ball between two players) with different situations (no obstacles, stationary or moving obstacles) as the RoboCup physical agent challenge Phase I. We describe how these techniques can be evaluated in terms of various kinds of design issues.