Basic Approach of Timing in Context Aware Architectures Verified by Concrete Advantages

Context aware applications are incorporating data from various sensors with different physical characteristics and therefore potentially different sampling rates. It is intuitively clear that reasoning in context aware systems should take into account differences in the time when sensor data was captured from different sensors before using this data. So far the timing issues have not been considered in the context awareness research and there are no frameworks that can handle these issues. We propose two architectural approaches to handle the synchronization issues along with a detailed analysis of the problems. To be aware of these issues and how to handle them can improve the accuracy of context reasoning and save network bandwidth and energy. To prove our claims we carried out an evaluation in a real world scenario of sensing with accelerometers whether a person is standing or sitting. This evaluation shows a potential to save a significant amount of battery power in our sensor nodes.