Towards Deduction in RoboCup

We are currently working on designing clients for Robocup-Simulation-League in Prolog and C++. We feel that logic and deduction is an appropriate approach for this task. A single player is part of a team and he has to deduce information on his situation. For example, players have to recognize when passing the ball is possible or a player is ooside. But, almost naturally, tasks to be solved by a team of autonomous agents are many-sided and complex. To achieve a goal a single agent has to use a set of complementary subtasks. Some of these subtasks consist of solving numerical equations to enable a player to handle tasks like dribbling or actually passing the ball. On the other hand, we have to derive new information from a given set of facts. So we were lead to the idea to combine both the advantages of procedural and logic programming and decided on a hybrid system. As a result, we implemented the Robolog Prolog extension. This extension is an enhanced RoboCup soccerserver interface for ECLiPSe-Prolog 3]. Time critical and computational expensive tasks are handled within the Robolog module, as well as the exchange of data. The developer of a soccer client accesses the calculated data via Prolog predicates. The module provides the atomic soccerserver commands and some more complex actions. Our C++-Module simulates the player's nerves in so far as it is the reasoners connection to the soccer-server. It supplies the reasoner with sensory input and initiates actions. The Prolog engine handles the player's reasoning and planing, i.e. to a certain degree it models a real soccer player's reasoning. However, it is diicult to say, in how far cognitive actions that are done on a subconscious level by humans, e.g. calculating the amount of kickpower needed to pass (or stop) the ball, should or could be added to the "nerves-layer" of our clients.