Project number
27060
Organization
General Dynamics Mission Systems
Offering
ENGR498-F2026-S2027
1. Project description: The United States Coast Guard (USCG) faces a persistent and growing capability gap in detecting semi-submersible and low-profile vessels (LPVs) used by transnational criminal organizations for narcotics trafficking. These vessels are deliberately engineered to minimize radar and visual signatures, rendering conventional maritime surveillance methods largely ineffective. Distributed underwater acoustic sensor networks — commonly known as hydrophone arrays — offer a compelling solution by exploiting the unavoidable acoustic signatures generated by vessel propulsion systems. This capstone project challenges students to design, build, and demonstrate a scaled proof-of-concept (PoC) hydrophone array system that integrates distributed sensing, real-time signal processing, machine learning-based classification, and a command-and-control (C2) dashboard interface.
At full operational scale, such a system would consist of networks of seafloor-mounted hydrophone nodes spanning hundreds of nautical miles, feeding acoustic data into AI-powered fusion centers that cue intercept assets in near-real time. For this capstone effort, students will demonstrate the same fundamental technical concepts at small scale — using a network of 2 to 5 prototype sensor nodes deployed in pool and/or open-water lake environments — targeting small, motorized watercraft as acoustic sources in place of full-sized vessels. The core technical challenge is identical to the operational problem: detect, classify, and localize an acoustic source of interest using a distributed sensor network, with results presented to an operator through an intuitive command interface.
2. Students will gain hands-on, portfolio-quality experience across a broad, in-demand skill set:
• Underwater Acoustics & Signal Processing:
- Fundamentals of underwater acoustic propagation and the acoustic environment
- Digital signal processing techniques: sampling, filtering, FFT, spectrogram generation
- Time Difference of Arrival (TDOA) and GCC-PHAT cross-correlation algorithms
• Machine Learning & Data Science:
- End-to-end ML development lifecycle: data collection, labeling, augmentation, training, evaluation
- Convolutional Neural Network (CNN) architecture design and training using PyTorch or TensorFlow
- Audio classification using spectrogram-based deep learning approaches
- Deploying a trained model for real-time inference in an embedded/edge environment
• Embedded Systems & Hardware Engineering:
- Embedded Linux operation on single-board computers (Raspberry Pi / NVIDIA Jetson)
- Audio interface integration and audio capture programming
- Waterproof enclosure design and fabrication
- Sensor node power management and power system design
- Wireless communication module integration (WiFi / LoRa)
• Software Engineering & Full-Stack Development
- Python application development for sensor data acquisition and processing
- Real-time web application development
- Interactive map integration
3. Deliverables:
Project deliverables are phased through the two-semester sequence and include:
- Weekly verbal or written progress reports on work completed and problems encountered
- Trade study report with technology recommendations
- Preliminary Design Review
- Testing and validation report
- Final implementation of the system
- Comprehensive project documentation and user manual
- Project presentation and demonstration
4. Required background: Background information required to successfully execute the project includes the following. Note that not all students are expected to possess all skills — the team should collectively cover these areas, and the faculty advisor and industry mentor will provide supplemental guidance
- Programming Proficiency (Python) —Python programming skills; the majority of signal processing, machine learning, and backend development will be implemented in Python. Familiarity with NumPy, SciPy, and Pandas is highly desirable.
- Machine Learning Fundamentals — Prior coursework or experience in machine learning, including familiarity with neural network concepts, training procedures, and model evaluation metrics. Experience with PyTorch or TensorFlow is preferred but not required.
- Digital Signal Processing (DSP) — Coursework or familiarity with DSP concepts including sampling theory (Nyquist), Fourier transforms (FFT), filtering, and spectral analysis. Prior audio or communications signal processing experience is a plus.
- Electrical Engineering Fundamentals — Basic understanding of electronics, circuit design, power systems, and sensor interfaces sufficient to assemble and troubleshoot prototype hardware nodes.
- Web Development (Basic) — Familiarity with web application development concepts; experience with React.js, JavaScript, or Python web frameworks (Flask, FastAPI, Django) is desirable for the C2 dashboard development workstream
5. Supplies/Equipment Required: The following materials and equipment are required for project execution: underwater microphone elements, edge compute, audio capture interfaces, enclosures, power source, communications equipment, target acoustic sources, GPS modules, waterproofing hardware, and other miscellaneous electronics and hardware dependent on student design.
Category Estimated Cost
Sensor Node Hardware (×3 nodes) $1350
Target Vessels & Acoustic Sources $325
Compute and Communications Equipment (x3 nodes) $750
Miscellaneous electronics $75
Miscellaneous hardware $150
Documentation & Miscellaneous $175
Subtotal $2,825
2 Additional nodes $1400
TOTAL ESTIMATED BUDGET $2,825-$4,225
At full operational scale, such a system would consist of networks of seafloor-mounted hydrophone nodes spanning hundreds of nautical miles, feeding acoustic data into AI-powered fusion centers that cue intercept assets in near-real time. For this capstone effort, students will demonstrate the same fundamental technical concepts at small scale — using a network of 2 to 5 prototype sensor nodes deployed in pool and/or open-water lake environments — targeting small, motorized watercraft as acoustic sources in place of full-sized vessels. The core technical challenge is identical to the operational problem: detect, classify, and localize an acoustic source of interest using a distributed sensor network, with results presented to an operator through an intuitive command interface.
2. Students will gain hands-on, portfolio-quality experience across a broad, in-demand skill set:
• Underwater Acoustics & Signal Processing:
- Fundamentals of underwater acoustic propagation and the acoustic environment
- Digital signal processing techniques: sampling, filtering, FFT, spectrogram generation
- Time Difference of Arrival (TDOA) and GCC-PHAT cross-correlation algorithms
• Machine Learning & Data Science:
- End-to-end ML development lifecycle: data collection, labeling, augmentation, training, evaluation
- Convolutional Neural Network (CNN) architecture design and training using PyTorch or TensorFlow
- Audio classification using spectrogram-based deep learning approaches
- Deploying a trained model for real-time inference in an embedded/edge environment
• Embedded Systems & Hardware Engineering:
- Embedded Linux operation on single-board computers (Raspberry Pi / NVIDIA Jetson)
- Audio interface integration and audio capture programming
- Waterproof enclosure design and fabrication
- Sensor node power management and power system design
- Wireless communication module integration (WiFi / LoRa)
• Software Engineering & Full-Stack Development
- Python application development for sensor data acquisition and processing
- Real-time web application development
- Interactive map integration
3. Deliverables:
Project deliverables are phased through the two-semester sequence and include:
- Weekly verbal or written progress reports on work completed and problems encountered
- Trade study report with technology recommendations
- Preliminary Design Review
- Testing and validation report
- Final implementation of the system
- Comprehensive project documentation and user manual
- Project presentation and demonstration
4. Required background: Background information required to successfully execute the project includes the following. Note that not all students are expected to possess all skills — the team should collectively cover these areas, and the faculty advisor and industry mentor will provide supplemental guidance
- Programming Proficiency (Python) —Python programming skills; the majority of signal processing, machine learning, and backend development will be implemented in Python. Familiarity with NumPy, SciPy, and Pandas is highly desirable.
- Machine Learning Fundamentals — Prior coursework or experience in machine learning, including familiarity with neural network concepts, training procedures, and model evaluation metrics. Experience with PyTorch or TensorFlow is preferred but not required.
- Digital Signal Processing (DSP) — Coursework or familiarity with DSP concepts including sampling theory (Nyquist), Fourier transforms (FFT), filtering, and spectral analysis. Prior audio or communications signal processing experience is a plus.
- Electrical Engineering Fundamentals — Basic understanding of electronics, circuit design, power systems, and sensor interfaces sufficient to assemble and troubleshoot prototype hardware nodes.
- Web Development (Basic) — Familiarity with web application development concepts; experience with React.js, JavaScript, or Python web frameworks (Flask, FastAPI, Django) is desirable for the C2 dashboard development workstream
5. Supplies/Equipment Required: The following materials and equipment are required for project execution: underwater microphone elements, edge compute, audio capture interfaces, enclosures, power source, communications equipment, target acoustic sources, GPS modules, waterproofing hardware, and other miscellaneous electronics and hardware dependent on student design.
Category Estimated Cost
Sensor Node Hardware (×3 nodes) $1350
Target Vessels & Acoustic Sources $325
Compute and Communications Equipment (x3 nodes) $750
Miscellaneous electronics $75
Miscellaneous hardware $150
Documentation & Miscellaneous $175
Subtotal $2,825
2 Additional nodes $1400
TOTAL ESTIMATED BUDGET $2,825-$4,225