My experience working on the grant project has been productive but stressful—though not necessarily because of the project itself. At the beginning of the month, I had a family emergency that required me to travel to India. While I was there, the city I was staying in became a target during a skirmish between India and Pakistan. Although I managed to get a significant amount of work done, it’s difficult to evaluate my experience based solely on the application itself.
Between refreshing CNN Live news and attending family events, I was working with Android Studio and Kotlin. We chose Android because it offers the most streamlined access to the DJI API, and because I have used similar technologies in the past. This also makes sense given that specialized versions of Android are commonly used in these types of systems and can provide portable control for drones. Currently, we plan for the app to have two essential features. The first is a location labeler: the drone operator will provide IP coordinates for buildings of interest. At present, the operator manually adds each point of interest, though this implementation may change.
One of our biggest challenges has been acquiring data to test our models. As I mentioned in my last blog post, our side-view model processes four images—one of each side of a building—to estimate a potential cost range for repairs. This model requires inputs in a very specific format, which is difficult to find online. Due to drone flight range limitations in many cities, collecting the data ourselves has also proven difficult.
Another issue we encountered was a lack of communication on my part. I felt that what I was working on was poorly understood. A large part of this has to do with how I would try to explain my vision for large sums of the project in what was essentially an email message.
This method of communication was sufficient for explaining the “what” but not so much the “why.” A lack of initial documentation and face-face team meetings also contributed to the confusion; which is why we decided to commit to both
As for data collection, we found a database containing satellite imagery of buildings damaged during Hurricane Helen. We believe that many of these buildings may still be in varying states of disrepair. We had begun making travel plans to collect data from the region, but unfortunately, a second family emergency occurred. As I write this, I’m currently on a plane to Amritsar. The IDEA grant project has been both fun and rewarding, but this month has been especially stressful due to unexpected personal events.
Toward the end of the month, we decided to pivot and pursue both a research project and a potential startup launch. This shift means we’ll need to produce significantly more documentation than before as well as need to network with insurance companies as well as state agencies- something I haven’t had to tackle before. It’s been a hectic month for our team—hopefully, the next one will be more straightforward.
HSF House Blogs 