How to feed training examples for a class called "competing robot" if I don't know how the robots of other competitors will look like?


I’m planning to go on a robot football competition where two robots are placed in an arena and have to collect and score the most ping pong balls into the opponent’s goal.

I’ve managed to build a custom classifier based on the Keras ver. of You Only Look Once that can run in reasonable time on a small computer like the Raspberry Pi. I can train for classes such as “ball”, “goal”, “corner” etc. My problem is that I don’t know how the enemy robots will look like. All I know is the maximum allowed dimensions. They might have visible wheels, or not. They might show that they contain balls, or not.

How should I go about training for such a class if the real-world situation might have a robot very different from what has been seen before?

1 Like

You could show it pictures of past competitions’ robots. If your other classes are “ball” ,“goal” etc, then its very likely that the classifier will still predict the unknown robot as a robot (its unlikely that it will look more like a ball than a robot).

Hi…there are some organizations for operating and supporting competitions for some of the world’s leading robotics platforms and organizations, including VEX, TSA, and BEST Robotics, the foundation also provides program support and workshops focused on technology and professional development for educators.

lead free assembly