Multi Sensor Context Awareness

Context awareness is one of the holy grails in ubiquitous computing, envisioning many embedded computing devices that serve their user while being perceptive enough of events and situations around them. This extended abstract will give a brief overview of an approach for resolving this key concern in ubiquitous and wearable computing, which uses techniques and ideas from statistics, machine learning, and signal processing. The issue of context awareness is treated as a traditional classification problem where raw sensor signals have to be fused and converted into a description that makes sense for both application and user.