Robot Soccer
Build two robots from parts and let them play on their own. Each one only knows what its camera and its wheel sensors tell it, and both are imprecise. From that it works out where it is, where the ball and the goals are, and what to do next.
Blue0:0Orange0:00
What you are looking at
- Solid robot: where it really is. Coloured outline on the floor: where it believes it is. The soft disc around the outline is how unsure it is.
- Thin lines run to the marker poles the camera recognises at this moment. The small rings at the poles show where the robot has mapped them; it starts without a map and builds one while playing (SLAM: simultaneous localisation and mapping).
- Ring near the ball: where the robot believes the ball is. When the ball leaves the camera's view, the robot keeps guessing from its last speed, then searches.
- Try Wheels only: the outline slowly drifts away from the robot, and it starts shooting past the goal. Or turn the noise sliders up.
- A collision or a shove moves the robot in a way its wheels do not measure. A bump switch tells it to trust its position less until the camera has fixed it again.
- Building a robot: every part costs something. Weight makes a robot slower to get going, a tall body must accelerate gently and falls over easily, a wide camera measures less precisely. Use Change robots to rebuild them.
- Gadgets are aimed at the other robot: a pusher shoves it away, a flipper tips it over, a firecracker shoves everything nearby and dazzles cameras, an oil patch makes any robot on it slide while its wheels spin. A robot on its side sees nothing and gets up after a few seconds, unsure of where it is.
A simplified simulation: flat physics seen from above (only the ball can leave the ground), cameras that report distance and direction to things they recognise, and a simple rule-based player. Drag to turn the view, scroll or pinch to zoom.