This project is all about building an Open Source robotic arm whose main purpose is to sort objects using a Pi-Cam based on criteria like color and shape based on the robot operator’s program.

It is based on a combination of programming capabilities such as Python because of the use of a Raspberry PI MCU, as well as electronics and building skills. Some more competencies that could be developed through this project are Linux knowledge, MATLAB, GPIO interfacing and 3D printing among many more.

This project is ongoing during Fall 2017 to Winter 2018. Interested in joining the team? Contact us or drop by the lab!

Progress:

CVRA Flow Chart

Technical Component:
At the moment, the arm is run by a simple Arduino code, it has been taught some basic movements and key locations in order place small objects into sorting bins on its left and right. A color detection script using OpenCV is run from a raspberry pi to detect whether an object is red or green and the relevant information is transmitted to the Arduino over its GPIOs. (see attached diagram).

The CVRA project began in the beginning of the fall 2017. The goal was to create a learning opportunity for students on matters of robotics, computer vision, microcontrollers and networking early on in their undergrad by taking on the challenge of building a robotic arm with computer vision functionalities. Since then, the team grew to a team of six student from unique backgrounds owning a broad palette of skills. The robot arm along with a camera stand has been designed and 3D printed assembled interfaced to a microcontroller and an onboard computer and machine vision python scripts have been successfully ran and tested.

Although 3D printing the arm provided us with great learning opportunities, it also taught us more about the technologies limitations. Gears are mechanical components that require a high degree of precision which increases as the complexity of the gear system itself increase. Unfortunately, warping and imperfections from the printing process eventually translated into “dead-zones” under which movement was significantly constrained. To overcome these limitations, a pre-fab arm (lynxmotion AL5B) was purchased and is now the main arm used within the project.

  • Timeline: Ongoing (contact us for details)
  • Project e-mail: cvraproject@ieee.concordia.ca