UNIVERSITY PARK, Pa. — Taking an interdisciplinary approach — and mixing in a little do-it-yourself initiative — a team of Penn State Berks researchers developed a wearable device that can tell the difference between indoor and outdoor lighting. The team adds that the device could help scientists better understand the health benefits of outdoor lighting and, in the future, lead to wearables that could nudge users to get more outdoor time.
The researchers describe the portable device in a recent issue of Hardware X. It includes sensors that can collect data on light sources, including wavelength and frequency data. They used a machine learning method called an artificial neural network — or ANN — to examine that data to determine whether the device was placed indoors or outdoors.
According to Matthew Rhudy, assistant professor of engineering, outdoor and indoor lighting have distinct differences that sensors can detect. The sun gives off vast quantities of light in the ultraviolet spectrum, or UV light, he said. Indoor lighting tends to be in the visible spectrum.
“We collected a lot of data with the sensor package,” said Rhudy. “We would set the sensors in various places and let it log data for an hour or so in all sorts of different situations and in different lighting conditions. The UV index alone is pretty good at detecting it. If it’s detecting significant amount of UV index, there’s a pretty good chance you’re outside.”
Light intensity, even if it’s cloudy, and differences in color temperature, are also signals for exposure to outdoor lighting. Current research suggests that exposure to natural light outdoors can have both mental and physical health benefits, according to the researchers. While scientists currently study these effects, they mainly rely on participants to record their own observations of when they are indoor or outdoor, which scientists refer to as self-reports.
According to the Penn State Berks team, self-report may not be as accurate as a wearable device that can precisely note the changes in lighting, as well as the amount of exposure to the different light sources.
Initial testing of the device was done by placing the device in fixed locations to collect light data for approximately one hour. Once the data was logged, the researchers used an ANN to classify the data as either indoor or outdoor lighting. To validate the device, the researchers placed the sensor package in a fixed indoor or outdoor location and collected data in a variety of locations, times and weather conditions. A total of 3640 indoor and 1368 outdoor samples were collected.