Choosing a Sensor

Once we had the general solution logic sketched out, we started making. The first part we implemented was attaching a sensor to our drone, and the first problem we faced was that the air quality was actually not bad in NYC. Although there were PM2.5 sensors available, we struggled to get a meaningful reading. We had to dial up the particulate diameter that we measured. I looked up online and Haotian met with an expert, Joseph Saavedra, for advices on choosing an effective sensor for our situation. Their meeting helped us clarify a few things about the electronics.

We assessed our options, bought a tiny carbon monoxide sensor and a compact dust sensor, wired them up and tested the results. It appeared that the dust sensor was our best choice, because it gave perceptible readings and had a more windproof enclosure. Even though the carbon monoxide sensor was much lighter and smaller, it was similarly difficult to detect the density of carbon monoxide and PM2.5 in the environments we were able to create.

As for the wireless connection, we took the expert’s advice to settle on bluetooth over heavier and more costly Wi‑Fi enabled boards or drones. We used an add‑on bluetooth module on a bare‑bones Arduino board, because we preferred separable, distributed structure to ready‑made, integrated component.

Weight Lifting Tests

Once we figured out the essential electronic components, we bound them to our drones to see if they could be lifted. Our smallest toy drone could not lift even one regular‑sized Arduino board. The mid‑sized drone would not fly higher than our knees with the load.

We had to buy a more powerful drone within our budget. The shipping took some time, but it turned out great. The final model was powered by a 7.4V 850mAh rechargeable battery, which allowed 10‑minute continuous flying. Including a 9V battery that powered the boards and the sensor, the model perfectly handled the combined weight, saving us time to potentially reframe the circuitry with smaller components.

Making a Cloud

We wanted the Cloudrone to resemble a cloud displaying the shared hope. Soft materials, such as polyester fiber, instantly came to us as a way to make a quick prototype. However, the rotary motion of drone propellers made them prone to tangle with loose fiber. We tried to turn to harder materials, considering paper sculpture, papier mâché, 3D printing with plastic filaments and plastic molding. But for one thing, it would be more time‑consuming to make. For another, it would require extra support to stabilize the electronics at the center of a hollow cloud, which would add more weight.

We focused on building a minimum viable prototype to demonstrate our concept, leaving possible concerns like heat dissipation and discoloration to a future iteration. Therefore what we needed was a lightweight shell to secure the soft materials appropriately. Felt perfectly did the trick. We cut it to shape and sewn the edges together, enveloping all the components inside with a velcroed opening reserved for recharging batteries.

The dust sensor needed to be placed away from drone propellers. We drew inspiration from a precedent project and found it easier to hang the sensor below the drone. The wires between the board and the sensor were lengthened, twisted together and protected with insulating tape. We spray‑painted the body of Cloudrone to be bright blue and white⁠—the colors that were easily recognizable in a smoggy background. To keep the shapes in line with the colors, we used light blue felt to form a droplet silhouette as the enclosure of the dust sensor, with a length of transparent plastic straw clearing the detection hole.

Proof of Concept

Cloudrone is an accessible, drone‑powered, semi‑stationary air quality monitor.

We believe in the value of deploying Cloudrones to make a social and environmental change. On city streets, a Cloudrone would serve as both a static measuring station and a mobile data collector.

Statically, we envisioned to set up Cloudrone stations at existing bus stops, bikeshare stops, subway exits, gas stations, convenient stores, outdoor drinking fountains, news stands, telephone booths or Wi‑Fi stations. These places were economic choices that could provide shelters and supply power for Cloudrone, and had already had their own widespread network in cities.

When people suspect harmful gas release in the city, they can easily control Cloudrone to measure the air quality indices at specific sites, such as a place close to a restaurant’s chimney, near a car’s tailpipe or around a construction site.

This dual‑mode implementation would engage the public with new technology while improving their sense of trust and responsibility on the environmental issues. All the data Cloudrone collects would be open‑source, which could be synchronized to a public database, and reported to monitoring departments.

Diverging and Converging

To name one challenge I faced when implementing wireframes, it was the search for cultural symbols that could resonate with the exhibition and also serve as website navigation. Signs, lintels, plaques, lanterns, firecrackers, and Chinese knots ... came to my mind.

It appeared to be a challenge because
 my client disagreed with all the candidates I explored. They were thought as being too cultural for this one of the 8 neighborhoods and thus lacking generality. This feedback was valuable to me in better understanding my client’s mission, instrument and content.

Accordingly I adjusted my direction, looked beyond Chinatown, and dropped most of the ethnic specifics. But this process of divergent exploration had helped me converge on the generic goal of this website—⁠to inspire diverse audience with different content—therefore, a middle ground was needed.

From the octagonal box of an electric firecracker, I eventually came round to the ubiquitous stop sign. It was created in the States and adopted around the world⁠—it’s common enough to be recognized, and had connotations of intersection, indication and navigation.

Iterating

Similarly, I diverged and converged on many iterations of page layout for mobile and desktop screens. For instance, I tried metaphoric visual element to convey connection, varying sizes to prioritize important item, bare display to list calls to action of equal significance, and reminiscent composition to maintain content continuity.

On a micro level, I tested various designs of modules, components and even map styles generated by JSON files.

Interactive Prototype

The final design was delivered for hand-off using Zeplin with all the parameters and assets in place, and also as an interactive prototype where user flows were intuitively laid out.

Design System

To formulate a design system for this project, I employed most of the preexisting brand guidelines to stay
 in accord with my client’s other visual establishments.