Inspiration
Inspired on projects in where a drawing is projected onto a surface and based on the techniques acquired in the image processing and robotics field during our Master's degree.
What it does
The user acquires a picture and send it to the processor, which gets the contour of this image and send commands to the robot to autonomously begin the drawing of these contours.
How we built it
The redBot Inventor's kit made by sparkfun is used as the autonomous vehicle; along with the dragonboard 410c a ROS has been integrated and the processing of the images and self navigation can be remotely controlled from a computer over the network.
The path is computed by processing an image, extracting its contours and a final algorithm finds an acceptable path to follow.
Challenges we ran into
Getting used to the hardware was one of the first problems as connection problems were faced; and some not supported hardware was acquired and not able to be used.
Some other issues implies the control of the navigation using just the encoders of the wheels which makes the navigation a little uncertain, which is reasonable acocording to the kit used for the development.
Accomplishments that we are proud of
Achieving robot navigation with some hardware limitations and the integration of simple hardware along with ROS.
Setting up a generic controller based on ROS, easily configurable for any other differential vehicle.
What we learned
Used of different hardware and sensors, along with the programming of the arduino that we have barely used through our master.
What's next for autonomous drawing robot powered by ROS
Development of a better navigation and position estimation by adding more robust sensors; and a camera to get a better exploration of the environment.
Adding a pen and controling it (i.e servo motor) in order to draw the desired sketch. Currently its path is published to RVIZ using markers.
Log in or sign up for Devpost to join the conversation.