We made our air quality mapping robot navigate around the house like a iRobot Roomba using the same technology Tesla uses to train its computer vision-based navigation system!!

Me and my teammates have recently been working on an air quality mapping project for Alaska Airlines Environmental Innovation challenge organized by Buerk Center for Entrepreneurship at the University of Washington.

The idea to combine a Roomba with an air purifier such that it eliminates the need for multiple air purifiers in the house dawned upon me when two months ago I saw my Roomba get tangled in the wires of my air purifier. We have built a robot that goes around the house and maps the air quality at different parts of the house. Based on the historical data, the robot should be able to predict where the air quality is potentially going to be bad and takes a position such that it can improve the air quality at source of generation (e.g., Kitchen).

Thanks to the generous $2500 funding that we received after clearing the initial checkpoint, we were able to buy fancy electronics for mapping the air quality and for the robot.

Here is a video of our robot going around my house, navigating the complex environment while creating the map of my house. Here I am using SLAM (Simultaneous Localization and Mapping using a Lidar for preserving privacy) for getting the coordinates of the robot in the map. Notice how the path planning algorithm changes the planned path of the robot when obstacles are introduced.

I am excited to work on this section of the project since it has applications in self-driving cars, consumer electronics, and game AI development. This project is fairly complex and I will share my progress on different sections of this project in the upcoming posts.

A big thanks to my teammates Sreejith Sivan Mayank Kumar Sanskar Naik Rishabh Gupta for all the bug squashing. It’s always a pleasure to work with you all.