Recognising the Activities of Low Entropy Mobile People Using Wireless Proximity Data

To improve the functional decline among elderly people, understanding the human behaviour is a substantial approach. Human behaviour is complex in nature and a challenging task to understand by monitoring the daily life activities. A tiered approach is adopted and investigated in this research work to recognise the high level activities and behaviour of the target users in order to alleviate their health related activities. An algorithm is devised to detect low level tasks by using the contextual information (e.g. Bluetooth and Wi-Fi proximity data). These tasks are further utilized for recognition of high level activities. Different scenarios and experiments are conducted to investigate the problem in hand thoroughly. Using wireless proximity data for activity and behaviour detection also improves the understanding of important structures in social relationships.