An Object Model for Sensor Data Integration

One example of large volumes of distributed data is that of sensor networks, where dedicated sensing equipment is used to monitor events and happenings in a wide range of domains, including monitoring human biometrics. In areas such as Personalised Health, large volumes of data are generated to provide an early indication of potential health problems. Our research is focused on sensor data that has been enriched in the form of XML trees, so that each sensor device can be regarded as an XML tree, and in some cases, containing services. While this architecture offers a certain degree of interoperability, current Web Services approaches are not suited to the difficult tasks of integration, restructuring of information and global issues such as updates. While a Web Services approach offers the advantage of loose coupling of data sources, a more formal model is required to underpin the sensor network. In this paper we describe our Object-Reference model (ORef), which extends a standard object-oriented model with references as the sole concept for locating objects (sensor objects) and their properties, and where the basic object-oriented modelling primitives are respecified to incorporate references. References are also used to define transformations present in the model in the sense that they are closed, meaning that each transformation operates on one or more objects, and results in one or more objects. The purpose of this model is to provide a stable bedrock for the loose coupling of data sources and services in an XML sensor network.