Zebravision 7.0

Yashas Ambati, Gabe Hart, Kevin Jaget

During the past few years, we have focused on developing robust sensing and inference systems to accelerate our progress towards a fully autonomous robot. The ability to estimate the robot’s absolute position on the field is critical to achieving this goal. During the 2020 season, we made groundbreaking strides in this vein, developing systems to make our robot more environment-aware. In this paper, we introduce two systems: a neural network-based object detector and a particle filter localization system. We believe that neural networks are the future of computer vision in FRC; to help foster innovation in the FRC community, we release our dataset from this season.