Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

Soccer player behavior development for the RoboCup Standard Platform League

Tsagarogianni Eleni

Full record

---

Summary

Humans are used to operate in dynamic environments, which nevertheless can be quite unpredictable and complicated for a robot. Therefore, for an autonomous robot to be able to adapt and act in such an environment, it must tackle a large variety of the problems, using a top-down approach that breaks them down into simpler ones. The annual robot soccer (RoboCup) competition offers such a dynamic environment and one of the main components of any RoboCup player is the algorithm responsible for decision making, known as robot behavior control. In this thesis, we aim to develop from scratch soccer player behaviors for the Standard Platform League (SPL), where all teams share the same robot platform, but develop different software, in an attempt to better understand the process of behavior development. Behavior development can be very tricky, since it combines information from different modules (vision, motion, communication, estimation, etc.) and decides the actions that need to be taken for the agent to achieve its goal. We worked on a recent code base framework developed by SPL Team B-Human that uses the C-based Agent Behavior Specification Language (CABSL) for the behavior development, which is freely available, however with all behavior code turned to void on purpose. We implemented three soccer behaviors for three basic soccer roles: a striker, a defender and a goalkeeper. Furthermore, we implemented an additional behavior, called ball control, in order to assess the utilization of the framework beyond typical robotic soccer behaviors. We aim for this work to serve as the initial steps for new code base of our SPL team Kouretes and as a guide for assisting the understanding and the use of the chosen framework and its capabilities.

Available Files

Services