My grandfather grew up a poor farmer in India, with no fancy technology to help him sustain his family - just sweat and muscle. One of the biggest problems he had were pests - crows and rodents flocking for a free meal at his expense. This, coupled with the fact that farms worldwide are blighted by the very same problem (In fact, Africa has a gigantic problem with Red-billed Quelea swarming their crops), inspired us to create Scarecrow and Skywatch - modern solutions for age old problems. We've neatly bundled these two projects into Farm Friend - to be used by any farm, anywhere.
What it does
Scarecrow departs from conventional pest repellents by being eco-friendly, humane, 100% effective, and simple. While other repelling agents are either poisonous, cruel to animals, weak, or just a plain pain to set up, Scarecrow is a piloted drone that plays on the innate reflexes of every wild animal. The premise is simple - we pre-loaded our drone with the call of the natural predator for every major pest, such as the Sparrowhawk call to respond to a threat of Red-billed Queleas.
Skywatch uses open-source fire maps as a reference for current forest fires, ensuring that farmers know where fires are when they need to. We also attempted to use the Planet API & Data, but due to a problem with our authorization key, we were unable to fully implement our vision. A detailed list of what we would have done to interpret Planet data is included in the Skywatch section of our website, given below.
How we built it
We used a Parrot AR Drone to act as the vessel for our MCookie assembly, bound together with a 3D printed basket. We researched specific predator calls to affect the widest array of crop-eating critters, and carefully downloaded the highest quality ones. We repeatedly tested our drone's cargo limit, tried many different MCookie assemblies, and tested out different sounds and speaker configurations. Initially, we were worried about the low volume of our speakers - but research showed us that birds and rodents actually hear higher frequency sounds than we do, and so the high pitched shrieks of a peregrine falcon over the drone of the motors (pun fully intended) stands out significantly more to them than they do to human beings. In addition, our 3D printed carrying basket actually helps amplify the speakers.
Challenges we ran into
Our first challenge was the lack of examples and documentation surrounding the MCookie setup. Fortunately, we were able to finally achieve our intended result through a lot of trial and error, and we overcame this hurdle well as a team. Testing the drone was also quite frustrating, because of the downtime - it was ten minutes of flight time for an hour and a half of charging.
However, our biggest challenge was using the Planet API and Data - We had solid ideas on how to use a lot of their data, but burned too much time trying to figure out why our API key wouldn't work - we asked upperclassmen, they worked their wizardry, but alas; 'twas to no avail.
Accomplishments that we're proud of
However, my happiest moment was when I called my grandfather, now going on 80 years old, and asked if such a product would help him with his gardening work - he goes through great pains to get the squirrels and rabbits away from the fresh tomatoes and Indian veggies that blossom up in our backyard. His response was an enthusiastic yes; a system by which he could monitor his plants and drive off four-legged intruders from the comfort of his favorite armchair was a huge bonus in his book.
What we learned
We learned to maximize uptime when the drones were charged, how to divide up tasks, shower and sleep in rotations, work with a diverse team with people we've never met before (It's our first hackathon, after all), and how to adapt to new systems and mediums of building that we've never used before. A couple of our teammates also learned that we're very much interested in Cybersecurity while talking to sponsors, and we're definitely planning on going deeper into that field.
What's next for Farm Friend
The next step for Farm Friend would be the full implementation of the Skywatch portion, after we figure out how to properly access Planet data. There's so much to be analyzed, so much to learn. We would also like to upgrade our Scarecrow drone to a better model to improve flight time, speaker volume, speed, and auto-recognition of pests. All in all, we've had a great time building a useful project, and look forward to improving it in the days to come.