Location and Activity Detection for Indoor Environments

Location-based services require to process and infer knowledge based on the understanding of people daily activities. This information acquires relevance in cases when elder people live alone since it could be used to infer facts about his/her health. In this paper a platform for getting location and activity data for indoors is presented. Gathered information is processed, first, to locate people at room level and then to identify whether a person is sitting, standing, walking or even running. For classification purposes we used an artificial neural network, which report an accuracy as high as 97.75 %. A web page, which complements the platform, is made available for those persons who want to know about the rate of visits to the rooms and the rate of, for instance, walking exercise. In our next version we would like to extend the identification of activity by detecting the way the inhabitant interacts with artifacts.