Ballerina Cappucina, a graceful omnidirectional robot with a critical mission. This project brings her to life by designing and building a robot that can autonomously glide across a “dance floor,” identify scattered colorful “bombs” (balls) and scoop them up with precision.
A custom omnidirectional mobile base is created that allows for fluid movement, and a vision system using OpenCV for color detection. Integrating these systems we enable Ballerina Cappucina to autonomously seek, collect, and sort the objects, clearing the stage.
working_model.mp4
The robot begins by rotating in place, scanning the surroundings until it detects a ball of a specific color, which is chosen beforehand using OpenCV for color detection. Once it identifies the target ball, it moves toward it, aligning itself over the ball.
Underneath the robot, there’s a unique cavity mechanism, resembling a revolving door. As the robot moves over the ball, this mechanism traps the ball inside by closing a flapping door-like structure. After the ball is secured inside, the robot resumes its search, rotating again to find another ball, and the process repeats itself.
This folder contains the source files and configuration necessary to run the robot simulation in ROS. It includes the core scripts for controlling the robot’s movement, as well as pre-configured launch files. Additionally, you’ll find detailed instructions on how to set up and run the simulation in a ROS environment. ROS files of Jetson are in jetson branch.
In this folder, you will find the source code for the ESP32 as well as detailed instructions on how to flash the code onto the device.
To implement this project in real life, you’ll need the following hardware components:
-
Jetson Nano (Main computing unit)
-
4 Omni-Wheels (For omnidirectional movement)
-
4 Rhino GB37 Servo Motors (For movement control)
-
4 Cytron DC Motor Drivers (For motor control)
-
Webcam (For visual perception)
-
ESP32 (For communication and control)
-
MG995 Servo Motor (For ball trapping mechanism)
This project was completed under the Society of Robotics and Automation (SRA), VJTI. Special thanks to the SRA for their continuous support, guidance, and encouragement throughout the development process.
