This paper describes the idea and preliminary results for a Remote Elderly Assisted Living (REAL) system. This system is designed to assist monitoring the increasing numbers of elderly and vulnerable people living on their own. The proposed system aims to monitor the normal activities of the elderly without intrusion into a person's normal living environment. If the system detects some deviation in the baseline of the observed habits such as less energy consumption, it could be an indicator that the observed person has difficulties and the system can trigger an event. In this paper we describe the methodology of the system designed to work based on multisensory information for remotely monitoring the activities of elderly and present a set of preliminary collected results from the proof of concept implementation.