| Description |
Wireless networks have become an indispensable part of our daily lives, and constant wireless connectivity of some kind is almost ubiquitous. This constant connectivity along with the availability of cheaper and smaller wireless devices has opened up new applications that were not possible before. As a result of these applications, more data is being collected at the edge of the network, such as sensor readings, pictures and videos from smartphone cameras, activity tracking data, and health data. While we have devices that can take advantage of this connectivity, there is a need for efficient ways of communicating this information reliably, while still taking advantage of the specific application needs. In this dissertation, we discuss two such applications. For the first application, we focus on the needs of an Internet of Things (IoT) framework for Epidemiological research. We present a framework that makes deploying sensors in homes easier and reliable. We share lessons learned from deploying our system in several homes of participants throughout the Salt Lake area. Within the context of IoT sensor deployments, we address two specifics challenges we faced when deploying sensors in homes. The first problem is improving the range ofWiFi so thatWiFi devices on the edge of connectivity can still send small amounts of data even if they are beyond the range of normal WiFi. We implement our solution and show that it works well beyond the range of standard WiFi. The second problem we faced is quickly bootstrapping WiFi connectivity of devices that are being deployed to a home's WiFi network. We develop a solution that securely broadcasts the WiFi security credentials, allowing multiple WiFi devices to receive this information and connect to a home'sWiFi. In the second application focused on in this dissertation, we look at live video upstreaming using an ad-hoc smartphone network to upload parts of the data in parallel. We develop a novel algorithm for distributing data between smartphones that adapts to mobility and changes in the network. We implement our system on real hardware and test it in realistic conditions. |
