Embedded Control and Vision Architectures for Humanoid Soccer Robots

Our Robo-Erectus version of humanoid robots has participated in the humanoid league of RoboCup since 2002, collecting several awards since then. Winning a competition not only requires, the vision system to be fast, accurate and reliable under varying lighting conditions but also an  robust and reliable control system in place. In practice, the field conditions for robot soccer can vary enormously from field to field, and even from day to day throughout the competition. This project aims to develop and integrate a robust embedded vision and control architectures for a soccer playing humanoid robots. A framework of hierarchical reactive behaviours is the core of control architecture for Robo-Erectus humanoid robots. This structure restricts interactions between the system variables and thus reduces the complexity. The control of the behaviours happens in three layers: skill, reactive, and planning layer. Object recognition and mapping forms the core of the vision architecture for the Robo-Erectus humanoid robots.

This project also aims to develop a simulator (Virtual-RE) for Robo-Erectus humanoid robots, which provides several possibilities of visualization and interaction with the simulated world. To simulate rigid body dynamics, Virtual-RE uses ODE, since this engine has a wide variety of features. The visualization as well as the computation of imaging sensor data is based on OpenGL, because this industry standard offers the best performance on modern hardware on different software platforms.

Key technologies:

  • Mechanical Design
  • Power Control
  • Embedded System
  • Computational Vision
  • Locomotion Control
  • Distributed Systems
  • Wireless Technologies





Untitled Document © Advanced Robotics & Intelligent Control Centre 2009