Abstract

Successfully recognizing people’s activities enables a wide range of pervasive computing applications. Recent research on activity recognition, particularly for elder care and health applications, has demonstrated that it is possible to recognize a variety of activities such as driving, walking, and using stairs and elevators [e.g. 2,5,6,9,10,13]. While we are inspired by existing activity recognition research, we believe as a community there are steps we can take together to enable future breakthroughs that are robust and reproducible.  Our general interest in this workshop stems from our research in ubiquitous and pervasive computing. A.J. Brush’s main research interest is technology for homes and families, in particular supporting sharing, sustainability, and helping with everyday problems such as scheduling and coordination. John Krumm focuses on techniques for measuring a person’s location and for using location data to benefit the user. He has worked on predicting driving routes and destinations and looked for routines in logs of activity data. James Scott conducts research on sensors and devices, mobile interaction, energy management, security and privacy.  Two current projects led to our particular interest in this workshop. First, as part of a study to understand arrival and departure prediction for households, we have been collecting GPS data from 12 households which we would like to provide to other researchers. This has required additional effort to collect and anonymize data to address privacy and legal concerns, which we would be interested in discussing with other members of the community.  Second, in a project that involves recognizing activities using sensors on mobile devices, we are more interested in building experiences based on recognized activities than conducting foundational research on activity recognition ourselves. Ideally we would be able to build on previous activity recognition research, but we have not found this easy to do. In the rest of this paper, we describe what we think the community is doing well, places we believe improvement would be beneficial, our recommendations for reporting on activity recognition research, and next steps we feel could benefit the community.