SCENTS: Collaborative Sensing in Proximity IoT Networks
Chenguang Liu, Jie Hua, Christine Julien · 2019
Mobile applications commonly use on-device sensors to continuously provide context: temperature, position, sound, etc. By collaborating to sense context, devices can save energy and share rare capabilities with minimal tradeoffs in sensing quality. Further, by leveraging already active communication behaviors, ambient context information can be collected at very little cost. We present a generic collaborative sensing framework, SCENTS, to support collective sensing for mobile IoT applications. SCENTS leverages two truths about IoT networks: (1) devices participate continuously in low-level device discovery mechanisms and (2) nearby devices tend to have similar values for many ambient context properties. We show that SCENTS balances sensing fulfillment and the fairness of energy consumption across devices. We measure the performance of SCENTS using real IoT devices and real world smart-city scenarios.