The UNSW RoboCup 2000 Sony Legged League Team
We describe our technical approach in competing at the RoboCup 2000 Sony legged robot league. The UNSW team won both the challenge competition and all their soccer matches, emerging the outright winners for this league against eleven other international t
- PDF / 232,995 Bytes
- 12 Pages / 451 x 677.3 pts Page_size
- 36 Downloads / 164 Views
Abstract. We describe our technical approach in competing at the RoboCup 2000 Sony legged robot league. The UNSW team won both the challenge competition and all their soccer matches, emerging the outright winners for this league against eleven other international teams. The main advantage that the UNSW team had was speed. The robots not only moved quickly, due to a novel locomotion method, but they also were able to localise and decide on an appropriate action quickly and reliably. This report describes the individual software sub-systems and software architecture employed by the team.
1
Introduction
Each team in the Sony legged robot league consists of three robots that play on a pitch about the size of a ping pong table. All teams use the same Sony quadruped robots. The 2000 competition included entries from twelve international laboratories. Since all teams use the same hardware, the difference between them lies in the methods they devise to program the robots. The UNSW team won the championship as a result of their innovative methods for vision, localisation and locomotion. A particular feature of these methods is that they are fast, allowing the robots to react quickly in an environment that is adversarial and very dynamic. The architecture of the UNSW United software system consists of three modules that provide vision, localisation and action routines. A strategy module coordinates these capabilities. Currently two strategy modules implement the roles of forward and goalkeeper. Each role can invoke a set of behaviours to achieve its goal. In the following sections, we describe the infrastructure modules that perform the vision processing, object recognition, localisation, and actions. We then describe the basic behaviours and strategies.
2
Vision
Since all the objects on the field are colour coded, the aim of the first stage of the vision system is to classify each pixel into the eight colours on the field. The colour classes of interests are orange for the ball, blue and yellow for the goals and beacons, P. Stone, T. Balch, and G. Kraetzschmar (Eds.): RoboCup 2000, LNAI 2019, pp. 64-75, 2001. c Springer-Verlag Berlin Heidelberg 2001
The UNSW RoboCup 2000 Sony Legged League Team
65
Fig. 1 A polygon growing program finds regions of pixels with the same colour.
pink and green for the beacons, light green for the field carpet, dark red and blue for the robot uniforms. Currently, we only use the medium resolution images (88 x 60 pixels) available from the camera. The information in each pixel is in YUV format, where each of Y, U and V is in the range 0 to 255. The U and V components determine the colour, while the Y component represents the brightness. The Sony robots have an onboard hardware colour look up table. However, for reasons that will be explained later, we have chosen to perform the colour detection entirely in software. Our vision system consists of two modules: an offline training module and onboard colour look up module. The offline software generates the colour tables and stores them in
Data Loading...
