Stacon: Self-Stabilizing Context Neighborhood for Mobile IoT Devices

Abstract

Retrieving context information from on-board sensors is critical to many scenarios in the Internet of Things (IoT). The battery capacity and physical size of smart devices can restrict the on-board sensors that can be supported and thereby limit the potential of applications for multi-agent IoT systems. In this work, we propose a novel scheme for building a context neighborhood among nearby devices through infrastructure-less collaboration. A context neighborhood is an ad hoc grouping of sensing devices that can collaborate to sense physical attributes of the environment. In this way, not every device needs to directly sense every context attribute to be able to leverage the information in its applications. Our approach entails a distributed algorithm that dynamically adjusts a device's sensing and sharing strategy based on the heterogeneity of resources in the proximity. We develop a prototype system using off-the-shelf IoT sensor kits; our demonstration exploits the self-stabilization feature to show the benefits this system could bring to higher-layer applications.