Dawit Negussey, professor of civil and environmental engineering in the College of Engineering and Computer Science, has been elected a fellow of the American Society of Civil Engineers (ASCE)—an honor that recognizes the contributions and creative solutions of civil engineers…
Does Your Smartphone Know the Real You?
Ask someone what they use their smartphone for and they will likely provide examples of how they use it to connect with friends, family and work, take photos, listen to music, play games or get directions. Beneath it all, there is an array of sensors that support these useful features, but they could also be used to make our lives more convenient by learning our specific lifestyles.
If you’ve used Google Now or Apple’s latest operating system, you are likely familiar with at least one good example of this. In recent years, like magic, both suddenly began to provide users with directions to their homes and workplaces without the user ever actually providing that information. Although it is somewhat spooky at first, there is actually a simple explanation. Your phone’s sensors know your location and how long you spend in that location. If you travel to a certain location at the same time every day and spend 8 to 16 hours there, it is very likely that it is your home. Similarly, if you travel to the same place five days a week for approximately eight hours a day, it is easy to guess that it is your workplace.
Professor Jian Tang of the College of Engineering and Computer Science is confident that much more can be accomplished using the data that these sensors collect as we use them and carry them around in our pockets every day. At a recent Syracuse Center of Excellence symposium, Tang presented his research on human-centric sensing with smartphones to provide smartphone users with more accurate shopping, dining and entertainment recommendations and provide other information that is specifically useful to them. It could even contribute to the safety of elderly and disabled people.
He explains, “We argue that a smartphone can do much more than just look at a person’s location to provide context. By using all of the sensors available to us, we are able to characterize many elements of their lifestyle. Most current models are equipped with many different sensors, including the GPS, digital compass, accelerometer, gyroscope, camera and a 4G or WiFi interface. Sensors are not limited to things built into the smartphone, either. Any sensor that can be attached to the phone through some network interface, such as Google Glass and FitBit, can provide insight as well.”
Tang and his team developed a mobile application called “Lifestyle Learning by Smartphone Sensing” to use the smartphone to learn a person’s lifestyle. To test it, they selected a group of users in six major cities and monitored their activities through the app for a month and a half and were able to identify places of interest for the test subjects.
Data collected from a businesswoman in Boston showed that she liked to shop, and enjoyed places like the mall, nice restaurants and coffee shops. Amazingly, the lifestyle model and sensors were also able to predict what the participants were going to do in the next two hours with an accuracy rate of approximately 70 percent. They also learned that if the weather is nice and temperature is high, people were more likely to go out to a restaurant. If it was rainy and the temperature was low, people tended to stay in. All of this info could be used to provide more useful advertisements and suggestions.
Tang believes that these sensors could be put to good use to keep disabled or elderly people safe as well. The sensors could alert family members if they detected any aberrant change in the person’s normal behavior, such as a lack of movement or travel to an unusual location at an odd time of day.
With smartphone technology constantly in flux, Tang intends to keep up with the changes. He says, “We have created a unified platform that can request and collect data from the sensors to support different sensing applications. We don’t want to have to build a new platform every time a new use is introduced. Most existing systems are application-specific and we want to build a unified system that can be used for all applications. Also, we want the platform to be able to support new and different kinds of sensors as they are introduced.”