Context monitoring framework for sensor-rich mobile environments

Abstract

In this thesis, we propose a context monitoring framework for sensor-rich mobile environments to enable context-aware and proactive services for individual users. Personal mobile devices will be a key platform enabling users to exploit emerging pervasive environments. They will form a PAN-scale dynamic distributed computing platform, dynamically incorporating a number of wearable and space-embedded sensors nearby. On the platform, context-aware applications will be a killer application, providing personalized, situation-aware services. However, it is challenging to develop and run such applications due to the complexity of managing highly constrained and dynamic resources over distributed devices and concurrent applications. While freeing the applications from such complexity, the framework provides a common base for multiple applications to proactively understand users’ contexts and react appropriately. For the purpose, the framework should continuously monitor sensor data, analyze, and figure out a situation of users without any intervention of them on behalf of applications. More important, it should address the problem that the context monitoring imposes heavy workloads on multiple distributed devices such as mobile phones and sensors. The problem is aggravated since resources are highly scarce and dynamically changing. To address the challenges, we develop a novel framework to enable scalable and energy-efficient context monitoring and effective resource use coordination. First, we propose a scalable and energy-efficient context monitoring framework, SeeMon. Running on a personal mobile device, SeeMon effectively performs context monitoring involving numerous sensors and applications. To enable scalable and energy-efficient monitoring, a novel context monitoring approach, i.e., bi-directional approach, is proposed. This approach gives the opportunity to achieve a high degree of efficiency in computation and energy consumption. Such an advantage results f...