Capstone Design Projects

Faculty Advisors: Prof. Christian Delozier, Mr. M. Kuiper

Background: Recent military conflicts have demonstrated that traditional sniper detection methods struggle to identify well-concealed threats, leaving soldiers vulnerable to long-range attacks. Existing ground and acoustic detection systems are often limited by short ranges, environmental noise, and significant setup risks, as operators must manually place sensors in high-danger areas. To eliminate these risks, modern military forces need autonomous, aerial surveillance solutions that can detect threats before an attack occurs.

Objectives: The goal of this project was to design and develop a drone-based detection system that identifies concealed snipers, vehicles, and thermal anomalies from a safe standoff distance before a shot is fired. By pairing off-the-shelf drones with real-time computer vision software, the system aims to enhance battlefield situational awareness without endangering operators. Additionally, the software was designed to run on a field laptop to provide interactive mapping, target tracking, and tactical data to command units.

Results: The system successfully demonstrated dynamic aerial threat tracking across several testing environments, operating reliably at target ranges between 100 to 200 meters with an overall drone operational range of over 800 meters. Software testing showed that while custom thermal contrast algorithms achieved an impressive 95.6% precision rate on distinct targets, broader environmental terrain called for combined edge-detection filters, which processed live video with minimal hardware latency (<171 milliseconds per frame) and low resource utilization. Although environmental conditions like thermal crossover presented challenges for pure heat-signature detection, field testing verified that the system's enhanced visual overlays significantly improved the speed and clarity with which operators could spot hidden personnel and vehicles.

Faculty Advisor: Prof. Sina Zarrabian

Background: Current drone operations from tactical vehicles like the Military RZR (MRZR) force operators to step outside protected vehicle armor to manually retrieve drones and replace batteries. This manual process is time-intensive and creates a high-risk stationary exposure window that directly endangers personnel in hostile environments. Traditional commercial "drone-in-a-box" solutions fail to solve this problem for mobile combat assets because they are bulky, require AC grid power, lack rugged active securing, and are vulnerable to harsh terrain and vibration.

Objectives: Project ATLAS aimed to design and field a modular conductive power-transfer landing pad and receiver system that enables autonomous, hands-off drone docking and charging. By enabling the drone to land, recharge, and redeploy without manual operator intervention, the system seeks to keep operators safely under armor. Additional design goals included creating a scalable 36×36-inch landing surface powered directly by standard military field batteries, maintaining an orientation-agnostic interface, and ensuring rapid, high-power transfer (100W USB-C PD).

Results: The team successfully constructed a conductive 3x3 modular pad grid paired with a custom drone receiver caddy featuring a 4-way bridge rectifier, enabling a landed drone to receive power regardless of its position or orientation. Empirical testing confirmed a highly efficient, stable 100.2W power transfer using 24V inputs from field battery banks. However, testing also identified key thermal limits: the voltage regulation module exceeded its 85 degree C target, and the Skydio X2D drone's internal thermal protection automatically capped high-current charging cycles at 10–11 minutes.

Faculty Advisors: Prof. H. Ngo, CDR D. Barber, USN

Background: Underwater communication systems traditionally rely on acoustic or radio-frequency (RF) signals, both of which face fundamental physical constraints. Acoustic links offer long ranges but suffer from low bandwidth and high latency, while RF signals attenuate severely within meters of seawater. Underwater Visible Light Communication (UVLC) offers a promising alternative with high bandwidth and low latency, yet most current systems remain bulky, expensive, or restricted to laboratory environments. To support research into secure, low-probability-of-detection underwater links, military researchers require a compact, accessible, and reliable UVLC testing platform.

Objectives: Project Beacon aimed to design, construct, and validate a low-cost, portable underwater visible-light communication prototype using commercial off-the-shelf (COTS) components. The primary performance objective was to achieve reliable digital data transmission at a rate of at least 100 bits per second (bps) over a distance of at least 10 meters through water. The system was designed around an Arduino-controlled LED transmitter using On-Off Keying (OOK) modulation and a photodiode-based receiver paired with signal-processing circuitry to reconstruct binary bitstreams.

Results: The team successfully developed and tested a functional UVLC prototype that achieved error-free digital data transmission using 1200 baud On-Off Keying modulation. During benchtop and aquatic tank evaluations, the system maintained a reliable 0% byte-error rate across simulated underwater depths up to 60.5 meters (achieved via blue-dye absorption modeling). Performance testing revealed that while transmission remained error-free at speeds up to 1200 baud, bit error rates rose sharply at 2400 baud and higher due to hardware filtering limits and water turbulence. Additionally, hydrostatic pressure tests successfully validated a hermetically sealed IP68-capable enclosure down to a 12-foot water depth.

• Hand Geometry Recognition System
• Ghidra Version Tracking Visualizer
• LiDAR Barcodes for Vehicle ID
• PAC: Programmable Arcade Console

• PAC++: Programmable Arcade Consoles
• GTAV: Give Them A Voice
• SWAT-C: Autonomous Weapon System
• RF Source Localization
• Compact Robotic Ordnance (CRO)
• Autonomous Modular Unmanned Ground Vehicle