A few weeks back, we held our second formal hackathon, where DoorDash engineers tackled projects like a tool to improve customer service communication, a lightweight recommendation engine, and an employee name game.
Oh, and we also crashed some drones.
As a technology company focused on delivery, we’ve always been fascinated with different ways to get goods from point A to point B, whether it be by foot, bike, car… or quadcopter. So we used the hackathon as a chance to roll up our sleeves and figure out whether drone delivery was actually feasible.
https://youtu.be/RbzBa5Dv_zoWe tested out two different drones for the hackathon: an off the shelf X8+from 3D Robotics and a prototype drone from a drone delivery startup. We had to modify the body of the X8+ by adding a delivery payload to the bottom of the copter while attaching a GoPro to capture the flight. Meanwhile, since the prototype drone was built with deliveries in mind, it already had a pretty sweet dispatch controlled auto-grasping mechanism built in that allowed for super simple payload pickup and dropoff capabilities.
On the day of the hackathon, a team of 8 or so software engineers started by assembling the quadcopter — where we quickly learned that the orientation of the blades has a material impact on whether the drone flies or not (tl;dr: don’t install them upside down). Once we finally got the drones assembled, we traveled to the baylands in Mountain View to conduct a few test flights. We used several open source projects to get these test flights working. We used the Arducopter Mission Planner to upload a few simple waypoint flight plans and tracked the progress of our drone in real time.The Arducopter software allows you to download mission log files and analyze them, which gave us a good sense of the large number of external variables that can influence the simplest of flight plans. We also performed a test takeoff and landing in our office parking lot to simulate the real restaurant pickup/delivery experience.
After a day of flights — and a few crash landings — we came away with a much better understanding of the opportunities for drone deliveries. While it’s clear that we’re still a long way off from actually delivering food with drones, we were impressed with how much we could learn in such a short time span.
One of the most exciting parts about working at DoorDash is the chance to explore, experiment, test, and play with ideas that don’t necessarily relate to our day to day responsibilities. While DoorDash won’t become DroneDash any time soon, it’s clear that drones could eventually help us and other delivery companies provide a faster and more reliable customer experience, and simplify a lot of the very tricky aspects of our logistics platform, like estimating when a Dasher will be available or how long it will take to perform a delivery.
For now though, the DoorDash Engineering team just enjoyed the opportunity to get out of the office and kick the tires — err, blades — of a new technology. If you love experimenting with new toys, and want to join us as we build the future of delivery, drop us a note here.