Cowbird: A Flexible Cloud-Based Framework for Combining Smartphone Sensors and IoT

Abstract

Combining sensors from smartphones and Internet of Things (IoT) (e.g., light, sound, UV radiation) enables a new kind of context aware applications in areas like smart cities or smart buildings. Such applications would require continuous evaluation of frequently changing data (smartphone sensors and IoT sensors), and continuous polling of IoT sensor data to the phone. Offloading both the evaluation and the polling of IoT sensor data to the cloud would improve energy efficiency and data transfer costs for the phone. To this end, we built Cowbird, an energy efficient framework to help developers easily build applications that evaluate sensor data both on the phone and in the cloud. To show that Cowbird is flexible and easy to use, we built a prototype app called "Quiet Route" that helps the user to avoid walking through noisy areas. We also compare the performance of Cowbird and the existing Cuckoo framework in terms of energy efficiency, data transfer cost and CPU load and show that Cowbird performs much better in the best case and equally good in the worst case.