Human Activity Recognition based on Real Life Scenarios

In Active and Assisted Living (AAL) systems, a major task is to support old people who suffer from diseases such as Dementia or Alzheimer. To provide required support, it is essential to know their Activities of Daily Living (ADL) and support them accordingly. Thus, the accurate recognition of human activities is the foremost task of such an AAL system, especially when non-video/audio sensors are used. It is common that one or more sensors could share or represent a unique activity, and consequently, finding out the most optimal window size among them to represent such an activity is challenging. This paper proposes a Recurrent Neural Networks (RNN) based on a windowing approach for subject-independent human activity recognition. The proposed RNN model is trained based on dynamic systems perspectives on weight initialization process. In order to check the overall performance, this approach was tested using the popular CASAS dataset and the newly collected HBMS dataset. The results show high performance based on different evaluation metrics.