Towards a Model-Driven Approach for Sensor Management in Wireless Body Area Networks

Nowadays, new portable devices are constantly being launched with their ever greater application to an ever growing number of domains. These devices or wearables (from wearable computing) are present in many different areas ranging from healthcare to entertainment, and provide a series of features to enhance the quality of everyday life. When used in conjunction with other wearables, they give rise to wireless body area networks (WBAN) or body area networks (BAN). The large variety of devices along with the lack of standardized services, which means that each designer or engineer must customize the API design, forces developers to implement source code mostly from scratch in order to cope with the heterogeneity of wearables and support their integration on a wider system. The result of these drawbacks is that new device integration is hampered and the time spent on the software development process is increased and these problems are addressed in this paper. We propose a model-driven approach based on a meta-model which has been designed to define and specify interaction with sensors. Our main aim is to distance developers from specific implementation and to cope with heterogeneous designs. The resulting models, which are instances of the proposed meta-model, are specified in a custom language which we call the wearable markup language (WML). We also introduce the coordinator, i.e. component-based software for handling sensor models and improving the integration of new sensors.