Our wonderful community at Petoi consistently leads the way in innovation and technological advancement. We are excited to feature Reid Graves, a winner of Petoi Spring 2024 Robotics Contest, in this article as he pushes the boundaries, innovates with Petoi Bittle smart coding robot dog, and discovers new applications for these open source AI robot pets.
When we launched the Spring Robotics Challenge and Contest in Q2, our goal was to encourage robotics education and innovation!
One of our contests is called Bittle/Nybble's Got Talent. The scope of this contest was to showcase Petoi open source Nybble robot cat or Bittle robot dog to perform new tricks or doing something that one could see in a talent competition, like singing and dancing. While all the participants had fantastic ideas, a submission by Carnegie Mellon University graduate student Reid Graves stood out! He sent us 2 videos of Petoi Bittle educational robot kit as part of his robot project entries that explore the AI robotics projects in action!
Reid's Bittle/Nybble's Got Talent contest winning video - Bittle four-legged robot searches for an acorn and picks it up
As you can see in the video, Reid's ingenious idea showcased Petoi quadruped robotics talent: not only the hardware and software, but also the AI automation as it was autonomously collecting acorns from the floor. What made this entry so special was that it outlined the extendable capabilities of the smart IOT quadruped robot as well!
In both videos, Reid attached a Raspberry Pi to Arduino Uno-compatible NyBoard microcontroller powering Bittle robotics kits. The Raspberry Pi extension gives Reid more CPU power and memory and enables him to prorgram AI algorithms & do more processing. Next, Reid attached a small, inexpensive smart camera that he got from Amazon to Bittle robot dog, which served as the guide for the autonomous movement. He designed a light robot arm to pick up acorns from the floor:
The arm is driven by a Petoi servo. I hooked it up to one of the unused ports on the Nyboard and defined a custom action to move it to the right angle. The white thing on the back is to deposit physical pheromones. This is so I can run experiments with multiple robots and they can communicate with the pheromones.
For now, Reid used a custom You only look once (YOLO) model to program the robot to move, detect and pickup acorns autonomously in real time.
Reid's second video: robotics pheromone following
This video dives deeper into the project and showcases the AI Robot dog from the POV. In this video, Reid sets a bunch of pheromones down on the ground and has the robot follow the path. As you can see in the video, the camera window shows the surroundings and marks the yellow pills. The other three windows show the recognition's inference time, command issued, and operational time per frame.
Reid's Bittle integration with RaspBerry Pi and ROS2
Reid shared his autonomous swarm robotics Python projects from the CMU-MAIL(Mechanical and AI LAB) lab with Bittle robot dogs:
Reid's robotics learning journey with Bittle Open Source Robot Dogs
When we spoke with Reid, we were delighted to learn of his story and how he got his idea for these projects. Reid told us that he started working with Bittle open source robotics kits last September for a research project. His ultimate goal for the project is autonomous swarm behavior with a collection of Bittle robot dogs. (For instance: having 3 robots explore their environment looking for food, such as acorns, and working together to collect all the acorns.)
Starting off, Reid had minimal experience using Arduino, Raspberry Pi, or ROS. We were extremely happy to know that, like many other students and enthusiasts who started working with Petoi Robots, Reid mentioned that he learned many new practical skills, such as Python socket communication, Linux command lines, GIT, OpenCV, YOLO, Raspberry Pi, and ROS2, just to name a few!
He also enlightened us about his ideas for the future. He plans to have a search function where his Petoi AI Robot pets automatically finds the acorns or any objects themselves. Further building on this, he wants to extend another servo so that he can unlock a new camera angle. He intends to develop adaptive gate control so Bittle can hopefully walk on rough and uneven surfaces without falling.
About the Petoi Robot Challenge and Contest
The Petoi Robot Challenge and Contest is a fantastic way to learn and showcase robotics! For the 1st edition, we were amazed by the fantastic level of participation: the simple programming by young kids to show engaging robotics movement, the use of artificial intelligence, and the excellent robot project entries we received! The challenge was simple: based on the given contest themes (like Bittle/Nybble plays in the park and Young Roboticist, etc.), participants had to record a video and showed the video to the world. The contest participants would get a chance to win exciting prizes like Amazon gift cards, coupons, or even a new Petoi Bittle or Petoi Nybble Robot!
To participate in the next installment of this edition, you can start learning robotics and coding, and combine your school project or college project with the robotics challenge and enter the competition. Here's a quick tip: Make sure to start your robot project a few months prior, so you, just like Reid, can explore more and dive deeper into your findings as well!
So which other robotics project won? Looking to know more about all the winners of the Robot Challenge? Read about all the Spring 2024 Robot Challenge winners: