Learning sensors usage patterns in mobile context-aware systems

Context-aware mobile systems have gained a remarkable popularity in recent years. Mobile devices are equipped with a variety of sensors and become computationally powerful, which allows for real-time fusion and processing of data gathered by them. However, most of existing frameworks for context-aware systems, are usually dedicated to static, centralized architectures, and those that were designed for mobile devices, focus mainly on limited resources in terms of CPU and memory, which in nowadays world is no longer a big issue. Mobile platforms require from the context modelling language and inference engine to be simple and lightweight, but on the other hand - to be powerful enough to allow not only solving simple context identification tasks but also more complex reasoning. These, with combination of a large number of sensors and CPU power available on mobile devices result in high energy consumption of a system. The original contribution of this paper is a proposal of an intelligent middleware for mobile context-aware frameworks, that is able to learn sensor usage habits, and minimize energy consumption of the system.