Part of the challenge for a project like Step Up! is engaging the user beyond the short-term. Curiosity is enough to get most people to try similar systems, like the Piano Stairs, and their desire to explore will keep them engaged for the short term. But how to get users to come back again and again?
For Step Up!, we wanted to give users a record of the exercise they performed on the stairs, data they could track over time to see the progress they have made in terms of steps climbed and calories burned. For this reason, we envisioned coupling Step Up! with Salud!, a database designed for health promoting systems. But how to get that data? We needed sensors that could track user movement, with granularity down to the individual stair.
To this end, we explored a number of options. Vibration and force sensors were initially considered, but we had concerns about noise from steps on other stairs. We were not sure if the force of the user’s foot landing on one stair would trigger more than one sensor or not. Next we considered infrared technologies. We looked into PIR sensors, which are designed to detect motion in an open space. This brought with it the issue of triggering the sensor by the user simply moving within its field of view, but we quickly devised a solution where we could create a hood for the sensor, narrowing the area for detecting motion.
However, we found that when detecting, the PIR brought with it noise of its own. Triggering it would make it unstable, so that it would shift between the triggered and untriggered states for a short period of time. Because of this, we decided to try a solution more closely designed for our needs: infrared emitter/detector pairs. This solution consists of an emitter, that generates IR light, coupled with a detector of IR light. When an object is placed between the two, the detector is tripped. In theory, this seemed like an ideal solution because the sensor was focused on a small area, unlike the PIR sensors, minimizing the amount of false detections. In reality, it did not provide the range we required. We were only able to detect IR across a span of about a foot, which was much shorter than the width of a stair.
In our final prototype, we went back with the PIR sensors. They were not perfect, but we devised a software work around for them, such that triggering a step once will not trigger it again unless another step has been triggered. In future work, we would explore other solutions that could give us the pinpoint accuracy of the IR emitter/detector pairs with the range of the PIR sensors.