Overview
Apex Putter was the final project as a part of Mecheng 495: Embedded Systems in Robotics, a core course of the Master of Science in Robotics program at Northwestern. The project required the use of the Emika Franka Panda robot arm and computer vision while utilizing the capabilities of ROS2.
Project Description
As kids, we thought robots were the coolest things ever. We viewed them as the future that can and would do any task you gave it. We thought robots wouldn’t have limits. Also in our naive ways, many kids believed mini golf was the dominant display of sportsmanship, control, and the overall “cool factor.” A blast from the past, for our final project for ME 495, we’ve decided to integrate these naive views into one project: Robotic Mini Golf.
Mini golf, a flat surface, presents an intriguing challenge for robotic systems. While kids develop mini golf skills through intuition, practice, and a little bit of healthy competition, programming a robot to achieve the same feat requires precision control of position and timing. A young kid needs a putter, a ball, a hole, and probably a bet and competition from their family/member to make a hole-in-one. They can just “eyeball.”
A Franka Robot? A lot more complex. It may not feel competition, but sure does feel the pressure from a badly planned cartesian path. In this setup, a Franka Panda robot arm attempts the fundamental task of mini golf - putting a ball with just enough force to reach the hole.
Our group uses vision integration via April Tag Detection, motion planning arm and joint positions, as well as cartesian paths, and just enough bit physics to prove that robots can play the simplest children’s game (or “perfect” first date), but it’s much more difficult than one would think. A whole lot of math and 3D printed parts later, we give you…Apex Putter
View on GitHub