27 April 2011
Urbana is an annual fundraiser hosted by Dharma Developments where proceeds raised support community programs with the goal of preventing homelessness in Ottawa. This year, URBANA is helping Action Housing / Action-Logement, an Ottawa-based charitable organization that assists low-income individuals and families to find and maintain safe, affordable and adequate rental housing.
Each annual Urbana hosts an artistic feature from local artists. This year, Dharma contacted Deb and I for an interactive lighting installation by The Latest Artists. The theme of “an on-going light for housing rights” was developed by integrating our medium of light with the need for the awareness of housing (and other social/charitable) issues to remain front and center: these issues don’t simply get solved with single instances of involvement or donation, but require constant attention to address.
This theme lead to the proposed installation: a large lighting structure, architecturally-inspired, requiring on-going crowd interaction to propagate its behavior, or,” keep the lights on.”
Some Google Sketchups were created:
This first sketch shows the six foot structure with an assortment of illuminated windows, akin to an apartment tower at night, w/ residents inside watching tv. The planned interaction for this mode was to have the crowd turn on the windows by touching them, after which the window remains lit for a period of time before turning off again; constant interaction is needed to keep the piece illuminated.
The sketch below shows off the full color (RGB) capabilities of the piece. Animations like this could be used in a game play mode to signify winners:
Read on for technical details of the project and to see the finished piece . . .
Design + Fabrication
Fabrication was key to this project, for providing a robust, polished product. With the help of Flux Lighting, we finalized the structure design and they provided some exciting renderings:
The bare structure
Structure with acrylic windows installed
The illuminated piece
In parallel to the design and fabrication of the structure, the electronics and lighting were designed and assembled.
The greatest technical challenge was making each window individually touch responsive. Our first impulse was to have an IR transmitter/emitter pair behind each window to sense a hand blocking the window; this was quickly put to rest by trying to fire IR through the acrylic only to have it bounce directly back and flood the sensor, making it only useful for detecting the acrylic windows themselves. Perhaps a combination of IR and IR-response cameras would do the trick, but we avoided this approach due to complexity and space issues inside the structure. The thought of wiring up individual piezo elements to each window, or some sort of physical switch, was too labour-intensive seeing as there are a total of 120 windows.
The most promising and scalable solution seemed to be capacitive touch sensing. We originally thought of using a dedicated chip like this, but after learning about the CapSense library — which works with minimal external parts on an Arduino — we decided to test a mock-up touch surface.
The advantage of the capacitive sensing was that we could use a row/column matrix arrangement, thus only needing 11 sense electrodes for the 30 windows of each fixture side (five for the columns, and six for the rows). This conveniently used 11 of the Arduino digital pins, while leaving the two serial pins free for communication. Four dedicated Prototinos were used, one for sensing the windows on each side of the fixture.
In consideration of budget vs. labor, Shiftbrites were used for the lighting. Each window has a single Shiftbrite mounted behind it, allowing for full color (RGB), individual control of each window’s illumination.
With lighting and touch sensing in place, we focused on a control scheme, deciding on an external control scheme, allowing the fixture’s behavior to remain easily reprogrammable. Our specific arrangement made use of a custom computer-based control panel written in Processing.
Here’s a schematic the control scheme:
The top row shows the four Shiftbrite panels that make up the four walls of the light cage; the 120 Shiftbrites are controlled by a single Prototino, connected to an XBee (and breakout adapter). The bottom row shows the four acrylic panels, each having a capacitive sensing matrix connected to a corresponding Prototino and XBee/breakout combo (I attempted to utilize serial messaging between the touch sense controllers with a single XBee to take care of all the messaging for the four sides, but I obtained much better results using individual XBees than ironing out my own protocol). On the right is a laptop, running the control software, and communicating with the touch panels and the light cage: the touch panels only send data to the control software when a touch is registered, and the control software then processes that information and sends appropriate commands to the light cage; calibration and debugging messages can also be sent between the control software and the touch panels and light cage.
Here’s an early pic of the control software and corresponding lights:
When the galvanized steel structure was available from the fabricators, it was time to assemble everything. This stage took a considerable amount of time since all the elements had exact mounting points that had to be accurately drilled to match the precise machining of the structure.
The below pic shows one of the window touch panels mounted to the inside of its corresponding structure wall:
You can clearly see the row/column touch electrodes; each intersection corresponds with the middle of a window, so when someone touches a window, which particular one can be determined from the corresponding row/columns registering a touch. This arrangement worked surprisingly well (after debugging some grounding and interference issues caused by the metal structure itself).
Here’s a pic showing three assembled sides of the structure while the power to the light cage is checked and double-checked:
The light cage is attached to a steel plate that mates with the top of the structure, so it literally hangs in the middle of the inside of the structure, each light aligned with a specific window.
The below pic better illustrates how I actually felt getting everything together under the tight deadline for the debut of the piece at Urbana:
After getting everything assembled and tested, it then all had to be torn down again for transportation to the event. I think once re-assembled and powered up at the venue, the final result was well worth the effort:
The crowd also seemed to enjoy it and it certainly made for an interesting conversation piece:
We were pretty proud as well:
The Latest Artists (Deborah + Andrew O’Malley)
. . . and Aksash Sinha of Dharma Developments was very happy w/ the literal and illuminating tie-in to the evening’s cause:
Andrew + Deborah O’Malley and Akash Sinha
So what did we learn?
Timeline management of supply chain and fabrication is crucial. Thank goodness for overnight shipping and a highly helpful and skillful fabrication team.
Capacitive sensing is touchy! A major problem we encountered was tuning the capacitive touch system. The circuit exhibited varying behavior between all environments from proof-of-concept to final integration with the structure. We experienced noise problems when using the touch panels near the lights, and further interference when the arrangement was placed into the metal structure (common grounding solved most of these issues — subject to a blog post all their own). The biggest downfall to this sensitivity was experienced during the move from studio testing to the actual venue: in the studio we had to use sensitive settings to obtain usable touch data, while once in the venue, the control software was receiving so much touch data that it became overwhelmed and couldn’t control the lighting exactly has planned.
Shiftbrites are simple in theory — the control data gets passed along nicely from the controller to each respective light along the chain — but have many practical caveats: power needs to be injected along the chain to keep things stable; the power supply used has implications; a long chain of Shiftbrites seems very susceptible to interference/glitches on the data lines, resulting in unwanted behavior (solved largely in software); and you must be very careful interconnecting Shiftbrites as they are easily damaged irreparably (make sure to order a contingent stash!). Diffusion needs to be well thought-out as well.
We received a lot encouraging feedback on the fixture from the Urbana crowd and several offers of interest to host it at other events/venues, so we’re currently considering these options along w/ plans to tour the piece around other events/festivals where it can be programmed to compliment their themes.
In the meantime, we’ll be refining the control software and programming/capabilities of the fixture. We also hope to improve the internal light cage, as it was quite fragile for transportation/installation; along with adding a grid of blinds inside the structure to better isolate each Shiftbrite’s light output to it’s specific window as we experienced a bit of light overlap between windows.
Stay tuned for updates . . .
Writing this summary, I’d like to thank Deb for taking this on with me as one of our more complex projects as The Latest Artists to date.
Together, we’d like to thank the team at Dharma Developments for inviting us to participate in Urbana 2011; we’ve been keen to work on something inspired by the evolving, ambient light patterns of apartment towers at night, and Urbana 2011 provided the perfect impetus to get a project like this started.
Big thanks to Dan and Caleb at Flux Lighting for all their interest in, enthusiasm for, and enabling of this project on the fabrication side.
Garrette at Macetech was extremely helpful with his knowledge of Shiftbrites and seems very comitted to helping customers get their ideas up and running.
Thanks also to the City of Ottawa, their funding and support has also been instrumental in realizing this project.