Date of Award


Degree Type


Degree Name

Master of Science (MS)


Computer and Information Technology

Committee Chair

Eric T. Matson

Committee Co-Chair

Anthony H. Smith

Committee Member 1

Byung-Cheol Min


The “Internet of Things” (IoT) is a commonplace term in today’s society. Measuring our homes, cities, and farm land, IoT devices and sensor networks help give insight with new amounts and types of data collected. With several billion devices projected to be in use by 2020, an inexpensive, reliable solution for communication is needed. Typically, this would mean Wi-Fi, and in the home that works; but what if the devices are several miles away? LoRaWAN has sought to fll this need, however, appears unreliable given changes in atmospheric conditions. Experiments in Europe have shown a complete communication breakdown when temperatures reach 60◦C, below the 80◦ hardware specifcation. The researcher’s own observations and preliminary test also show inconsistencies in terms of range and received signal strength. This thesis describes the process of designing a physical server-client architecture using a Dragino LoRa client and Raspberry Pis. Using hobbyist-available materials, three LoRa nodes were created to test the effects of cold weather, along with rain and snow. The nodes were deployed from January to the beginning of March, recording communications, weather data, and the received signal strength of packets. This data was then analyzed for factors that affected communication most. The experiment was split into two phases, one for recording the natural environmental conditions, the other, for measuring environmental conditions when heat is applied directly to the LoRa radio chip. Visual representation and statistical correlations were used to determine the relationship between temperature and humidity inside and outside a node, the intensity of rain and snow, and the temperature of the radio chip, compared to the received signal strength (RSS) and received packet ratio (RPR). From the comparisons made, humidity appears to be a leading predictor in LoRa communication reliability. This is followed by temperature, then the amount of rain, and fnally snow. The temperature of the radio chip, from ambient to 60◦C does not seem to affect signal strength or communication in a noticeably impactful way. This shows an indication that communication failure is caused by problems with the antenna or the micro-controller, a distinction other experiments have not made, however the exact distinction between antenna and micro-controller was outside the scope of this study.