A System for Distributed Context Reasoning

Abstract

Context aware systems use context information to adapt their behaviour accordingly. In order to derive high level context information from low level context, such as sensor values, context reasoning methods that correlate observable context information, are necessary. Several context reasoning mechanisms have been proposed in the literature. Usually these mechanisms are centralized, leading to suboptimal utilization of network resources and poor system performance in case of largescale scenarios. Therefore to increase the scalability of context reasoning systems the development of methods that distribute the reasoning process is necessary. Existing distributed approaches are method specific and do not provide a generic formalization for distributed reasoning. In this paper we introduce a novel system which enables distributed context reasoning in a generic way that is independent of the reasoning algorithm.

Categorization of the Context Within the Medical Domain

“…Pascoe et al [9], Chang et al [27], Hofer et al [17] and Prekop et al [20] divided the context into two main categories, the physical context that can be measured by the physical sensors and the logical context that contains abstract information about the environment or the interaction such as the user's emotional state, goals, etc. In a similar way Wang et al [22], Guan et al [30] and Rizou et al [38] distinguished between two categories of context. First, we have low level or observabale context which represents the information that can be directly obtained from sensors or other sources, and secondly there is high level context or non-observable information that must be inferred from the first kind of context.…”

Providing QoS Guarantees in Large-Scale Operator Networks

“…For instance, the operators of a large-scale sensor network [3,5] might implement functions such as data selection and aggregation in order to jointly filter and process streams of sensor data. Other application areas include complex event processing [13] and distributed reasoning systems [11].…”

Exact convex formulations of network-oriented optimal operator placement

“…Many data-processing tasks like context reasoning [1], estimation using Kalman filters in wireless sensor networks [2] or data aggregation in global sensor networks [3], require the processing of data received at spatially distributed sources for the use at specific locations. Rather than gathering all data at a central location, it is favourable to process the data already in the underlying communication network.…”