Project

RinoBot System

First C++/Qt control system for UFJF's robot soccer team — computer vision, strategy, and robot control.

Archived Repository →

Overview

Between 2016 and 2018 I was Diretor de Visão Computacional on UFJF’s Rinobot team and wrote the first version of the C++/Qt system that controlled the robots’ AI in the IEEE Latin American Robotics Competition’s soccer category — small autonomous robots playing matches under an overhead camera.

What’s in it

  • Computer vision: real-time detection and tracking of robots and the ball from the overhead camera feed.
  • Strategy: game-state representation and per-role behavior selection (attacker, defender, goalkeeper).
  • Robot control: the low-level loop that turned strategic decisions into wheel commands sent to the robots over the radio link.
  • Tooling: a Qt-based control panel for tuning, calibration, and match-time overrides.

Why it still matters to me

It’s the earliest version of the pattern that shows up in everything I work on now: a perception layer, a decision layer, and a control loop, with the failure modes of one layer hidden behind interfaces the next layer trusts. The agent framework I work on today is the same shape with different lenses.

Status

Archived. Linked here as academic / early-career credit, not active work.