Recognizing and recording activities of daily living is a significant problem in elder care. Ubicomp systems targeting ADL recognition have been limited in the number of ADLs they recognize, the detail they recognize, and their robustness. A new paradigm for ADL inferencing focuses on the objects people use during their day. To do this, it leverages three techniques: radio-frequency-identification technology to sense objects being touched, data mining to partially automate model creation, and a probabilistic inference engine. To test the concept, 14 people performed ADLs in a real house containing 108 tagged objects.
Index Terms:
context-aware computing, sensor networks, ADL monitoring, Proactive Activity Toolkit, Proact
Citation:
Matthai Philipose, Kenneth P. Fishkin, Mike Perkowitz, Donald J. Patterson, Dieter Fox, Henry Kautz, Dirk Hahnel, "Inferring Activities from Interactions with Objects," IEEE Pervasive Computing, vol. 3, no. 4, pp. 50-57, Oct.-Dec. 2004, doi:10.1109/MPRV.2004.7