Autonomic body sensor networks

Recent advances in Body Sensor Networks (BSNs) and ambient assistive technologies have facilitated the realisation of technology enabled assisted living. Integrating ambient sensors with wearable sensors will not only enable comprehensive health monitoring of the user, but by utilising network resources from ambient sensors, it could also greatly prolong the battery lifetime of wearable sensors, extend the coverage of BSNs and enable context aware sensing. The challenge lies in the complex management of these two platforms. Autonomic sensing provides a solution by allowing the sensor network to manage itself. This paper proposes a self-organising approach for an integrated ambient and wearable sensor network. A multi-cluster network design is adopted in the proposed approach, and the network structure is limited to having a maximum of two hops for real-time requirements. In addition, to maximise the lifetime and reduce signal losses of the network, a cluster-head selection mechanism using a Maximum A Posteriori (MAP) estimation technique is designed based on sensors´ remaining battery, connectivity and distance to the base station. This approach enables full connections of a two-hop sensor network in a data collecting scenario, and it has been shown to outperform the conventional sensor network approach, the Low-Energy Adaptive Clustering Hierarchy (LEACH).