Smartphones to be smarter. Will make your commute less stressful.
According to recent research, apps can detect what type of transport the commuters are using and could offer relevant advice.
Researchers at the University of Sussex’s Wearable Technologies Lab believe that the machine learning techniques developed in a global research competition will solve a lot of daily problems. This includes smartphones being able to predict upcoming road conditions and traffic levels and offer route or parking recommendations. Even it will help to detect the food and drink consumed by a phone user while on the move. The study appeared in the Journal of the ACM.
Study author Daniel Roggen says that their study could collect sensor modalities of smartphone and also collect the data with phones placed simultaneously at four locations which include hand, handbag, backpack, and pocket. He added to design robust machine learning algorithms they need to know transport modes, conditions, sensor modalities which are unprecedented.
Roggen and his team collected the equivalent of more than 117 days’ worth of data monitoring aspects of commuters’ journeys in the UK using a variety of transport methods to create the largest publicly available dataset of its kind. The project gathered data from four mobile phones as researchers continue their daily commute for over seven months.
The team launched a global competition challenging teams to develop the most accurate algorithms to recognize eight modes of transport (sitting still, walking, running, cycling or taking the bus, car, train or subway) from the data collected from 15 sensors measuring everything from movement to ambient pressure. The project saw 17 teams take part with two entries achieving results with more than 90 percent accuracy, eight with between 80 and 90 percent, and nine between 50 and 80 percent.
The winning team, JSI-Deep of the Jozef Stefan Institute in Slovenia, achieved the highest score of 93.9 percent through the use of a combination of deep and classical machine learning models. In general, deep learning techniques tended to outperform traditional machine learning approaches, although not to any significant degree. Now it is hoped that the data set will be used for a wide range of studies into electronic logging devices exploring transportation mode recognition, mobility pattern mining, localization, tracking, and sensor fusion.
Roggen said by organizing machine learning competition they can share experiences in the scientific community and set a baseline for future work. He added automatically recognizing modes of transportation is important. For example during video streaming quality should not be hampered despite entering a tunnel or subway, or display information about connection schedules or traffic conditions.