Mobile Sensor Networks for Independent Living and Safety at Home

It is projected that the senior population in the U.S. will grow to 72 million by 2030 which will represent 20% of the total population [He et al., 2005]. Because the impact of institutional care is extremely high in terms of the monetary cost, health, and happiness of seniors, health care at home using assistive technology has recently received much attention. The use of computer and information technology will improve the quality of living while reducing the overall health care cost. Furthermore, the need of institutional care can be delayed or reduced since seniors can live independently and more safely with the help of assistive technology.

To help a senior to live independently at her home, we need a sophisticated monitoring system that can keep track of the activities of a person. From the collected activities of a senior, we can build a behavior model which can later be used to detect any noticeable changes in her actions. Wireless sensor network (WSN) technology is an attractive approach to build such a monitoring system due to its low cost and ease of installation. However, with a fixed WSN, it is costly to obtain a full coverage of a house. An alternate to a fixed WSN is a mobile sensor network consisting of mobile and stationary sensor nodes.

In this project, researchers will develop a mobile sensor network system to monitor activities of occupants, to detect abnormal behaviors or emergency situations, and to alarm a third-party in the case of emergency in an indoor environment. They dub this mobile sensor network system “Guardian Angel” because it helps and protects the user, but the user will not feel its presence during her normal activities. In order to make this mobile sensor network system affordable to the general population, the reseachers will develop mobile sensor nodes based on an inexpensive off-the-shelf robotic platform. Unlike high-end mobile robots, the platform they intend to use does not provide high-quality sensors and actuators. Hence, the main objective of this project is to develop robust inference algorithms and multi-robot coordination methods to overcome the issues arising from using inexpensive hardware.