🤖 UGOT Robot Auto Demo

Introduction

Welcome to the UGOT Robot Auto Demo! This project showcases the autonomous control of the UBTech UGOT Robot in a soccer game scenario. The robot can independently lock onto and kick a ball, demonstrating a small but significant part of the UGOT soccer game.

Control Methods

• Manual Command Selection Control
• Full Autonomous Mode

Features

Command Set Control

Easy to expand!

• Commands implement methods declared under the Command class.
• Inspired by the WPILib command set control logic from the FIRST Robotics Competition.

Webpage Control Panel

Integrated with a local server, communicates via WebSocket

• Manual Command Selection Buttons
• Video Stream Preview
• PID Tuning with Reaction Curve
  • Maintains data from the past 50 ticks of the program run.
• PS: this front end webpage might look terrible, as I am not a frontend person, it is just usable, lol

Computer Vision

• YOLOv8 Model
  • Trained with a customized dataset of ping-pong balls and goals used in the UGOT soccer game.
• Monocam Distance and Angle Measure Algorithm
  • No depth camera or depth prediction model is involved.
  • Algorithm located in utils.get_relative_position.py.

🎥 Video Demo on YouTube

🤩 Get Started

1. Installed all the required packages (docker support already added to repo)
2. Make sure that the UGOT robot is connected to the same WiFi as your computer
3. Config the IP of the robot in the config.py file
4. Open a terminal and go to the ~/static directory to run python -m http.server to start the webpage service
5. Open http://localhost:8000/index.html in the browser to view the control panel (including the PID Tuning panel)
6. Open another terminal and go back to the root directory then run python main.py to start the program
   • the default mode is the autonomous mode, so the robot will start locating the ball as soon as your computer is connected to the robot
   • an Opencv frame will be pulled up on your computer, this frame should have better video quality than the frame shown on the webpage
   • to switch to manual command selection mode, go to ~/threads/control_thread.py then comment out the line command_planner.update(update_signal) at line 46, then you will be able to control the robot fully through the buttons on the webpage