Matthew Lee, dwellSense Lead Researcher, Ph.D. Student in Human-Computer Interaction, Carnegie Mellon University
Long-term deployments of technology give researchers an opportunity to understand how technology can be embedded and incorporated into people's lives, particularly since observations of daily living (ODLs) naturally occur in people's everyday lives. However, long-term deployments are not without their hazards. One of these hazards is that the longer the study, the more likely it is that things (both good and bad) will happen to affect the data collection and even the integrity of the ODL data. I'd like to detail a couple of issues that recently complicated our deployment: Wi-Fi coverage and residents moving apartments.
In our evaluation of the dwellSense system, we have deployed a suite of sensors in 12 apartments in a senior high-rise apartment building. For the last seven months, these sensors have captured information about participants’ daily activities such as medication taking, phone use, and meal preparation. To provide us with the connectivity we needed to collect the data from our sensors, we set up a Wi-Fi network that covered the apartments in our study. Initially, this was a challenge because 1) participants lived in different parts of the buildings and 2) the building’s dated architecture and thick walls reduced Wi-Fi coverage. During the initial deployment, we solved this coverage issue by installing wireless routers that piggy-backed on existing broadband service in two of our participants’ homes. This was enough to cover three floors of the building where many of the study participants lived. All was well for about seven months.
At the beginning of the eighth (and final) month of the evaluation study, we learned that the 90-year-old, historic building was finally going to be renovated. Of our 12 study participants, three had to move apartments. Fortunately for them, they each moved to a beautiful corner apartment with a view of city. Unfortunately for us, the Wi-Fi network and router we had installed in one of their apartments no longer reached any of the other residents.
The first solution we tried was to install a Wi-Fi repeater to increase the range of our network. This attempt didn't work because the repeater would not connect to the router we set up. The repeater connected well with other routers, but not with our router. This was frustrating. After many hours rebooting and configuring the repeater and router, they just would not get along. Even if we had gotten them to connect, the repeated Wi-Fi signal was weak at best when extended two floors down where we would need it.
The second solution we tried was to reinstate a Wi-Fi network in the location where it had been before the move. Fortunately, one of our study residents lived in the center of where we needed Wi-Fi coverage. We asked him if we could pay the additional cost to add broadband Internet service to his already extensive (and expensive) cable television service. He agreed, so we ordered the new service for his account, picked up the equipment from Comcast, and installed the cable modem and a new router in his home. All was well—the dwellSense system in his apartment was once again connected to Wi-Fi and uploading data. However, even though the router was just a few feet away from the original position, the Wi-Fi coverage did not reach all the study participants that it had once covered. After countless visits to apartments to find that the computers having trouble connecting to a Wi-Fi signal that teetered between "weak" and "no signal," we realized that we needed to boost the signal with the Wi-Fi repeater. Finally, after installing two additional Wi-Fi repeaters, we were able to get everyone back on Wi-Fi for the final month of the study. This took nearly a week and a half of trial and error with debugging.
This Wi-Fi-induced nightmare should be lesson that evaluating systems in the field can be difficult, particularly if the evaluation lasts for a long time. The difficulties are twofold: 1) the unpredictability of people's lives (e.g., people had to move apartments) and 2) debugging IT infrastructure that is mostly invisible (e.g., it's hard to know how far Wi-Fi signals will reach except by trial and error). However, when understanding how technology fits into people’s lives, long-term evaluations and field deployment are still the best way to get real, valid data.