CMPUT503 (Robotics)

This Repository is for sharing codes and exercises of CMPUT 503 course at the UofA.
More information can be found on this personal website: https://golnazmes.github.io/Robotics.html

Running the Codes

Each exercise has its own directory.
All the exercises follow the format of this repository.: https://github.com/duckietown/template-ros
This template provides a boilerplate repository for developing ROS-based software in Duckietown.

To run the code, you need to cd to the directory of that exercise, and build and run the code on duckiebot with the following commands:

cd ExerciseX
dts devel build -f -H DUCKIEBOT_NAME
dts devel run -H DUCKIEBOT_NAME

Description of Exercises

Exercise 1

This exercise was about building the duckiebot and running some simple code on it.(make the duckiebot say hello!)

Exercise 2

This exercise was about the introduction to ROS and kinematics. We built the following tasks for this lab:
1- Make the duckiebot move in a multi-state task
2- Make the duckiebot change its LED colors
3- Estimate the robot’s pose in the world frame, and an introduction to odometry We worked as teammates for this exercise.

Exercise 3

This exercise was about Computer Vision and Sensor Fusion.
We implemented apriltag detection, a PID controller for lane following, performed localization, and worked with transformations.

Exercise 4

We implemented a tailing behavior for the duckiebot in this exersice.

Exercise 5

We implemented a machine learning model for digit classification in the duckietown!

Final Project

The final project was a demonstration of using all the concepts we learned during the semester. The duckiebot should perform a set of tasks in the duckietown including apriltag detection, lane following, avoiding other robots, and parking.

###Final Grad Project Code is adapted from https://github.com/jihoonog/CMPUT-503-Exercise-5 for expercise 5 and modified for using human feedback to improve classification.