Context-Aware Mobile Power Management Using Fuzzy Inference as a Service

As smartphones become ubiquitous, their energy consumption remains one of the most important issues. Mobile devices operate in a dynamically changing context, and their embedded sensors can be used to extract the relevant context needed for resource optimization. In this paper, we present a context-aware power management system implemented as a widelyapplicable middleware application. Fuzzy inference is used to represent a highlevel description of context, which is provided as a service. We test our approach using actual periodic and streaming applications on a mobile phone. Our results show energy reduction of 13-50% for periodic applications, and 1836% for streaming applications.