Testing in the Australian skies
Project Wing is collaborating with communities and partners in southeastern Australia to trial our delivery drones
Over the past few years, Project Wing has conducted thousands of flights to get our drone delivery technology ready for everyday use. In early 2016, we successfully completed our first drone deliveries to members of the public in an open field at Virginia Tech University. This fall we’ve been testing in a rural community on the border of the ACT and NSW and tackling an entirely different level of operational complexity: making deliveries directly to people’s yards.
Our testers — alpaca farmers, math professors, equestrians, and artists (not to mention a few curious kangaroos) — have been helping us fine-tune how our drones move goods from where they’re located to where they’re needed. And today we’re announcing that two Australian merchants are joining our tests, as they’re eager to understand how drone delivery could help them serve their customers better. Guzman y Gomez, a Mexican food chain, and Chemist Warehouse, a chain of pharmacies, will receive orders from our testers who’ve purchased items using the Project Wing app on their smartphones. We’ll dispatch our drones to pick up the order from our partners’ loading sites and then transport and deliver the goods to testers at their residences.
Here’s a bit of an update on what we’ve been learning over the last several weeks and what we’re going to be focusing on in the weeks ahead.
Residents near our testing area on the outskirts of the ACT live an idyllic country lifestyle on 10-acre blocks of rolling land spotted with gum trees and horses. But they face a 40-minute round trip in the car for almost anything, whether it’s a carton of milk, veggies for dinner, or a cup of coffee. Our testers, including young families, busy professionals and retirees, had many suggestions for how our technology could address this fundamental inconvenience. They wanted fresh meals delivered at dinner time. Some who run small businesses at home wanted to be able to send customer orders from their doorstep. A few with farms wanted supplies to arrive at their paddocks, or spare parts delivered to the ailing vehicle on their property. Almost all said that they’d value having medicine delivered to their door, especially when they’re unwell.
They also had ideas about delivery drones being used to transport drinking water, food, medical supplies, and mechanical parts to emergency service workers operating in rural areas or places cut off due to floods and fires. As part of our upcoming tests, we’ll help the Australian Capital Territory Rural Fire Service assess how our technology could aid their efforts.
Last year at Virginia Tech, our first deliveries with members of the public were in an open field, not to a specific address or location. Now, with each delivery, we encounter a new yard space with its own layout of trees, sheds, fences, and power lines. That means that in addition to learning what people want delivered, we also have to learn how to best deliver items to people.
Our drones are able to deliver items almost anywhere — backyards, public parks, farmlands or even fire-breaks. But we need to train our systems to reliably identify safe and convenient delivery locations. This is more complicated than it looks. We have to incorporate customer preferences — e.g. many of our testers would like packages delivered to backyards so they’re not visible from the road, or near kitchens so food items can be unpacked quickly. And we have to be ready to accommodate changing conditions at the delivery location. While our unmanned traffic management (UTM) platform lets us pre-plan a flight route, the sensors on our aircraft are responsible for identifying obstacles that might appear during a flight or delivery, like a car parked in an unexpected spot, or outdoor furniture that’s been moved. The more test deliveries we do, exposing the sensors on our aircraft to new delivery locations, the smarter our aircraft’s algorithms will one day become at picking a safe spot for deliveries.
To operate an effective drone delivery system, Project Wing must be able to pick up packages from anyone, in almost any location. This presents an interesting design challenge: our technology must be intuitive and easy to use, so packages can be loaded and received without any specialized infrastructure and by people without specialized experience.
Our partners Guzman y Gomez and Chemist Warehouse will teach us what we need to do to ensure that orders are channeled to their staff smoothly and that they can easily load goods onto our delivery drones. In the case of Guzman y Gomez, who is our first delivery partner for this trial, we’ll need to make sure our technology fits in smoothly into their kitchen operations, as their staff have to juggle many orders at once to ensure that every customer is served fresh, hot food in a timely fashion. We want to learn how much notice to give them for a drone’s arrival so that they can cook, pack, and load it in one well-timed workflow.
Through our partnership with Chemist Warehouse, we want to ensure our system is able to support merchants with a wide variety of products. As part of this test, they’re offering nearly 100 products across categories like vitamins, dental care, sun care, and over-the-counter medicines. By practicing how we pack items of very different shapes and sizes into our fixed-sized package, we’ll learn how to optimize how many items we’re able to deliver per flight.
The information we gather from both of these test partners will help us build a system so that merchants of all kinds can focus on what they’re good at — like making food or helping people feel healthier — rather than being distracted by complex delivery logistics.
We know the weeks and months ahead will be filled with unexpected challenges as we undertake these new tests. We’re grateful to the communities in the ACT and Queanbeyan regions who’ve let us into their yards, so we can learn even more about building a delivery network ready to fly in the open skies.
Originally published at blog.x.company on October 16, 2017.