Picture is from 24 Spring TAMU Robomaster Roster.
We won against UT.
I joined the prestigious TAMU Robomaster team since 2024 Fall. Our team had a decorated history, usually regarded as one of the strongest teams in the international division. I was a part of the Computer Vision Team, for developing algorithms for our robots to aim and shoot at the enemy robots before they shot us too much.
This is a wonderful journey as it's my first grand design on realtime system on a embedded system, with latency, CPU utilization, and all those metrics of image processing and recognization at bench competitively. When working with consumer products or company servers, even though they say performance is good, in the end there are a lot of problems we ignore simply because computing power is too cheap and Moore's law caught up too fast. However, when everyone uses similar enough hardware and there could only be one winner, we have to go optimizing everything, rewriting stuff in C, compiling with PyPy, developed our own custom model, so much tediousness and toughness I won't be able to experience if I stay afloat in the software development world, but I'm glad I touched grass.
We currently on working on Fast Fourier Transformation in Autoaim process, used primarily in Sentry category and secondarily in Hero and Standard robots. We're also trying to pretrain model to accelerate our model against the peers.
