More than a decade ago, pervasive computing and context awareness where envisioned as the future of computing , initial work concentrating on location, typically indoor. Today, small, handheld computers of various forms and purposes are becoming pervasive in the form of PDAs, mobile phones, and increasingly advanced GPS units. However, except for location based services, like knowing your location based on GPS, context awareness has not really materialised yet. In modern agriculture, computers are pervasive, but only in the sense that they are present everywhere. All types of equipment, ranging from feeding- and ventilation systems to tractors have build in computers, and, in most cases, can also be queried or controlled remotely. These systems provide an excellent base for gathering context, which may then be exploited to ease the work of the farmer. Furthermore, additional sensors may collect context about the individual agricultural worker, e.g. in the form of location or current activity. Knowing the context of the farm and the workers may then be utilised for building pervasive computing applications to support the daily work at farms, e.g. by easing access to information which is useful in the current context, or even automatic registration of work carried out. In this paper, we look at the challenges for building a middleware which supports collection and sharing of context from the heterogeneous sources and environments found at a farm as well as context reasoning, and design a middleware to overcome these challenges, using semantic technologies. We also study new ways of designing such middleware, in that we make the end user of the applications built with the middleware participate in the requirement gathering and design phases.
Middleware Conference: Proceedings of the 5th Middleware Doctoral Symposium, 2008, p. 19-24