This week, the team focused on refactoring the project to meet the project goals with the coronovirus quarantines in place, and each member working remotely. Aditi is the only once in Pittsburgh. We had a few zoom meetings to discuss the future trajectory of the project.

We came up with the following concerns over out original project –

– All parts located in Pittsburgh while most of the team is not.
– Not able to test on original test cases.
– No access to TechSpark for the mechanical assembly of our robot
– Missing equipment such as monitors and peripherals to interface with the PI at home
– Long integration and communication cycles.

As a result, we decided to remove the computer vision component completely. This means we will no longer use a camera for door detection and text labeling. Removing this part of the project will allow for simpler mechanical assembly of the robot; we won’t have to create a stand and fixtures to hold the camera and sensors. This will save us time with assembly, integration and testing.

We also want to add to the user-facing web application. Originally, our user interface would have been a basic image display. Some possible enhancements to the web application include: stitching multiple floors together, user correction or control, Google Maps API integration, and a visualization of the mapping progress in real time.These software additions are much easier for us to test and contribute to remotely and during asynchronous work periods. They can be tested instantaneously from our own machines and do not have many dependencies, and still add to the same user experience.

Because Aditi is in Pittsburgh, she will have access to all the collective parts. We will be pair programming a lot of the navigation algorithm and SLAM together, and Aditi will test extensively on her own. The web application component will mostly be done by Shanel and Alex. Our expected output of the project will still be a robot that generates a map of an indoor space, using SLAM and a custom navigation algorithm. During the mapping process, the user will be able to see progress of the generated maps on a web application. The quantitative requirements of power, speed, and accuracy will remain the same.