Physical Data Collection and Transpondence System for Beetleweight Combat Robotics Arena Using Computer Vision

Abstract

Beetleweight combat robotics tournaments are spectator events in which two robots, weighing less than three pounds, are remotely controlled to mechanically damage and disable one another. The controlled environment of the arena, a square plexiglass box with a wooden floor, enables the monitoring of the fight from all angles using cameras. Using an axonometric viewpoint of the arena from above the fight, image detection software can track the position of each robot throughout the match. Both robots are understood to be roughly three pounds. This means that fundamental physical properties, such as velocity, acceleration, force, and energy in impacts are all approximately measurable. While these are all interesting figures to display on- screen when televising tournaments, this system can also be used to quickly measure the effectiveness of drivetrain and weapon systems in a practical environment. Any metric of a combat robot’s performance can be calculated empirically using simple physics, but there are far more factors to consider in an actual fight. This system would enable the approximate measurement of several key metrics of a robot in practice. Moreover, this project provides an opportunity for Wildcat Robotics members at the University of Arizona to gain experience working with a tangible application of computer vision.