LuckyStar II - Team Description Paper
Our first robotic soccer team, LuckyStar, competed in Stockholm 1999 and was lucky enough to win third place. From that experience, we found that our team weaknesses were due to lack of vision reliability, smooth robot movement control, shooting capabilit
- PDF / 148,196 Bytes
- 4 Pages / 451 x 677.3 pts Page_size
- 30 Downloads / 204 Views
1
Introduction
Our first robotic soccer team, LuckyStar, competed in Stockholm 1999 and was lucky enough to win third place. From that experience, we found that our team weaknesses were due to lack of vision reliability, smooth robot movement control, shooting capability and team cooperation. We decided to concentrate on areas with the most immediate impact on the game, i.e. vision, robot movement control and shooting mechanism.
2
System Description
We continue to use a global-vision-based system except that the vision system now resides in the host computer. The host computer computes the next move for each robot. The outputs, which are the robot’s translational and rotational speeds, are sent to the robots via a Radiometrix RF transceiver.
3
Robot
Figure 1. Picture of robot
Figure 2. Kicking mechanism exposed
The new robots are similar to the old robot except that they have kicking mechanism. The kicking mechanism is based on a rack and pinion system driven by a DC motor. This allows the robot to kick both way and with varying amount of force. The robot is P. Stone, T. Balch, and G. Kraetzschmar (Eds.): RoboCup 2000, LNAI 2019, pp. 543-546, 2001. c Springer-Verlag Berlin Heidelberg 2001
544
Ng Beng Kiat et. al
able to shoot with full force at opponent goal and with just enough force when clearing the ball towards the sloped wall so as to avoid putting the ball off field. The host system decides completely when to kick the ball. In general, once the ball is within kicking distance, angular error between the robot orientation and the balltarget-to-robot angle is used to determine shooting criterion. The further the robot is from the ball-target, the smaller the angle error allowance.
4
Vision system
One of the problems faced last year is the detection of opponent color. The ping pong balls are small and not evenly illuminated as they are not flat. In order to be able to use only simple vision algorithm, the camera and the frame grabber must be carefully chosen so as not to lose any ping pong ball color pixel unnecessarily. 4.1
Hardware
We tested some cameras and concluded that 3-ccd cameras are the best in term of resolution but they are extremely expensive. Not being able to afford one, we settled for the second best, a single-ccd camera with RGB output. With composite video, the color information are encoded at the camera end and decoded by the frame grabber board. The result is the degradation of signal-noise ratio, hence poorer color resolution. It is also important that camera has electronic shutter capability in order to be able to capture high-speed object. The shorter the exposure time, the sharper the high-speed object. Of course, we must make sure the lighting requirement for short exposure can be met. For the frame grabber, we chose the FlashBus MV Pro from Integral Technologies as it provides RGB input and field capture interrupt capability for fast vision processing response.
4.2
Software
We rely on color patches on top the robot to determine robot orientation. In order to play a
Data Loading...
