A Localization Method for a Soccer Robot Using a Vision-Based Omni-Directional Sensor
In this paper, a method for robot self-localization based on a catadioptric omni-directional sensor is introduced. The method was designed to be applied to fully autonomous soccer robots participating in the middle-size league of RoboCup competitions. It
- PDF / 441,985 Bytes
- 12 Pages / 451 x 677.3 pts Page_size
- 64 Downloads / 244 Views
Carlos F. Marques? , Pedro U. Lima Instituto de Sistemas e Robotica, Instituto Superior Tecnico Av. Rovisco Pais, 1 | 1049-001 Lisboa, PORTUGAL
fcmarques,[email protected] http://socrob.isr.ist.utl.pt/
In this paper, a method for robot self-localization based on a catadioptric omni-directional sensor is introduced. The method was designed to be applied to fully autonomous soccer robots participating in the middle-size league of RoboCup competitions. It uses natural landmarks of the soccer eld, such as eld lines and goals, as well as a priori knowledge of the eld geometry, to determine the robot position and orientation with respect to a coordinate system whose location is known. The landmarks are processed from an image taken by an omni-directional vision system, based on a camera plus a convex mirror designed to obtain (by hardware) the ground plane bird's eye view, thus preserving eld geometry in the image. Results concerning the method's accuracy are presented. Abstract.
1
Introduction and Motivation
The navigation system is perhaps the most important sub-system of a mobile robot. In many applications, especially those concerning indoors well-structured environments, one important feature of the navigation system concerns the ability of the robot to self-localize, i.e., to autonomously determine its position and orientation (posture). Once a robot knows its posture, it is capable of following a pre-planned virtual path or to stabilize its posture smoothly[1]. If the robot is part of a cooperative multi-robot team, it can also exchange the posture information with its teammates so that appropriate relational and organizational behaviors are established[9]. In robotic soccer, these are crucial issues. If a robot knows its posture, it can move towards a desired posture (e.g., facing the goal with the ball in between). It can also know its teammate postures and prepare a pass, or evaluate the game state from the team locations. An increasing number of teams participating in RoboCup's middle-size league is approaching the self-localization problem [2]. The proposed solutions are mainly distinguished by the type of sensors used: Laser Range Finders (LRFs), visionbased omni-directional sensors and single frontal camera. The CS-Freiburg and Stuttgart-Cops teams can determine their position with an accuracy of 1 and ?
This work was supported by grant PRAXIS XXI /BM /21091 /99 of the Portuguese Foundation for Science and Technology
P. Stone, T. Balch, and G. Kraetzschmar (Eds.): RoboCup 2000, LNAI 2019, pp. 96-107, 2001. c Springer-Verlag Berlin Heidelberg 2001
A Localization Method for a Soccer Robot
97
5 cm, respectively, using LRFs. However, LRFs require walls surrounding the soccer eld to acquire the eld border lines and, in a sense, correlate them with the eld rectangular shape to determine the team postures. Should the walls be removed, the method becomes not applicable. RoboCup's Agilo team proposes a vision-based approach to the self-localization problem too. A single frontal camera is used to match
Data Loading...