Recognising Activities at Home

Authors:

Abstract
What activities take place at home? When do they occur, for how\ud long do they last and who is involved? Asking such questions is\ud important in social research on households, e.g., to study energyrelated\ud practices, assisted living arrangements and various aspects\ud of family and home life. Common ways of seeking the answers\ud rest on self-reporting which is provoked by researchers (interviews,\ud questionnaires, surveys) or non-provoked (time use diaries). Longitudinal\ud observations are also common, but all of these methods\ud are expensive and time-consuming for both the participants and\ud the researchers. The advances of digital sensors may provide an\ud alternative. For example, temperature, humidity and light sensors\ud report on the physical environment where activities occur, while\ud energy monitors report information on the electrical devices that\ud are used to assist the activities. Using sensor-generated data for\ud the purposes of activity recognition is potentially a very powerful\ud means to study activities at home. However, how can we quantify\ud the agreement between what we detect in sensor-generated\ud data and what we know from self-reported data, especially nonprovoked\ud data? To give a partial answer, we conduct a trial in a\ud household in which we collect data from a suite of sensors, as well\ud as from a time use diary completed by one of the two occupants.\ud For activity recognition using sensor-generated data, we investigate\ud the application of mean shift clustering and change points\ud detection for constructing features that are used to train a Hidden\ud Markov Model. Furthermore, we propose a method for agreement\ud evaluation between the activities detected in the sensor data and\ud that reported by the participants based on the Levenshtein distance.\ud Finally, we analyse the use of different features for recognising\ud different types of activities
Funding Information
  • ESRC National Centre for Research Methods, University of Southampton (ES/N011589/1)

This publication has 39 references indexed in Scilit: