Obstacle Detection, Identification and Sharing on a Robotic Soccer Team

When building a representation of the environment for a robot in a multi-agent application, as is the case of robotic soccer, sensor and information fusion of several elements of the environment are an important task. To build an increasingly better world model, one of the aspects that one should consider is the treatment of obstacles. This paper gives an insight of the general steps necessary for a good obstacle representation in the robot world model. A first step is the visual detection of the obstacles in the image acquired by the robot. This is done using an algorithm based on radial search lines and colour-based blobs detection, where each obstacle is identified and delimited. After having the visually detected obstacles, a fusion with a-priori known information about the obstacles characteristics allows the obstacle separation and filtering, so that obstacles that don't fill the criteria are discarded. With the position information shared by team mates, the matching of the ...

• Type: Conference
• Year: 2009
• Where: EPIA
• Authors: João Silva, Nuno Lau, António J. R. Neves, João Rodrigues, José Luís Azevedo
• Keywords: Artificial Intelligence, EPIA 2009, Obstacles Characteristics, Robotic Soccer, World Model