Date of Award

6-2026

Degree Name

MS in Industrial Engineering

Department/Program

Industrial and Manufacturing Engineering

College

College of Agriculture, Food, and Environmental Sciences

Advisor

Jill Speece

Advisor Department

Industrial and Manufacturing Engineering

Advisor College

College of Engineering

Abstract

Wildfire response depends on how quickly a detection reaches the people who act on it, and the slowest remaining step is often the link that carries an alert from a remote sensing platform to a satellite. This thesis models the latency of that link, the Air-to-Space uplink, for a wildfire-monitoring UAV that carries a Starlink terminal and sends an ALERT packet to a serving Low Earth Orbit satellite. The uplink is difficult to predict because both the UAV and the satellite move, and because the wildfire environment degrades the channel at the moment the data matters most.

The thesis uses Model-Based Systems Engineering to build the uplink as a single action diagram, decomposes its latency into nine sources, and runs the diagram as a Monte Carlo simulation of 10,000 trials per configuration. A baseline model represents the normal uplink and is validated against published Starlink measurements. A wildfire-modified model reuses the same action diagram and changes only the sources the wildfire environment affects, so that the added latency can be traced to the actions that produce it. The models are run across three packet sizes, from a 200-byte alert to a 2-megabyte image.

The results show that the wildfire environment raises uplink latency unevenly, adding most of its delay through reduced data rate, worse pointing geometry, and increased retransmission. The dominant latency source depends on packet size, with beam alignment dominating the smaller packets and transmission dominating the largest. A 10-kilobyte detection summary is selected as the configuration that meets a sub-minute delivery target while carrying useful monitoring information. Applying a faster beam-tracking method cuts the mean uplink latency of this configuration by about 32% under wildfire conditions. The same method applied to the non-wildfire baseline cuts its mean by about 47%, so the comparison holds the mitigation constant across both cases. Measured against a baseline that receives the same mitigation, the wildfire uplink retains a residual penalty of about 80% in the mean, although both mitigated cases remain well within the sub-minute target. The reduction traces to the targeted source, whose beam contribution falls by about 78%, which confirms that the change acted where the model predicted. The result supports the conclusion that beam-tracking mitigation removes the largest fixed latency contributor for the chosen packet, while the latency added by the wildfire environment itself persists and sets the remaining gap.

Download

Share

COinS