Project Daedalus: Autonomy & Controls for In-Space Manufacturing
Project Daedalus team at the COSMIC Capstone Challenge finals, NASA Goddard Space Flight Center (2025)
Background
From August 2024 to May 2025, I worked on Project Daedalus, a year-long senior design project at Virginia Tech. The mission was part of the 2024–2025 Consortium for Space Mobility and In-Space Servicing, Assembly, and Manufacturing Capabilities (COSMIC) Capstone Challenge. Our task was to design a payload for a small satellite bus that could autonomously perform additive manufacturing, inspection, and assembly in low Earth orbit (LEO).
The payload was built for the Blue Canyon Technologies X-Sat Venus Class bus, which limited us to about 60 W of orbit-average power, a payload envelope of 17 × 16 × 27 inches, and a maximum mass near 78 kg. These bus characteristics created constraints that shaped our decision-making process. We limited the design to 74 kg to leave margin for integration and shielding. The bus also provided 13.6 Ah of onboard storage and S-band up/downlink under 2 Mbps, enough for health checks and telemetry.
To handle these limits, we used a structured systems engineering approach. Trade studies, value system design, and iterative reviews guided choices about power, mass, and risk. The final payload integrated three main subsystems: the Directed Acoustic Energy Deposition (DAED) aluminum printing system, a laser welding system for the assembly process, and an AI-based non-destructive evaluation (NDE) system for defect detection and fault recovery with minimal human intervention.
The project was guided by Dr. Kevin Shinpaugh from Virginia Tech and Dr. Randy Spicer from Northrop Grumman, whose mentorship emphasized technical rigor and systems engineering discipline.
Project Summary
Payload subsystem integration — DAED printer, NDE system, laser welding, and power management
Our payload combined three main systems: DAED for in-space aluminum printing, a laser welding unit for part assembly, and an AI-powered Non-Destructive Evaluation (NDE) system for defect detection. Together, these demonstrated a practical approach to in-space additive manufacturing and inspection.
Systems engineering was central to the project. Every decision — from power allocation to component placement — had to be validated against mass, volume, and communication limits. The integration process forced us to constantly balance subsystem performance with the overall mission design.
My Role & Contribution
- Led the Autonomy & Controls subteam and contributed to systems engineering trade studies.
- Researched machine-learning-based inspection methods for 3D-printing defect detection.
- Designed an NDE (non-destructive evaluation) system that controls, inspects, and monitors the fabrication subsystem by integrating an AI computer with convolutional neural networks (CNNs).
- Built the FDIR (Fault Detection, Isolation, and Recovery) logic to keep the payload operational.
- Defined the command and data handling architecture, linking the AI computer, onboard computer (OBC), satellite bus, and ground station.
- Helped refine the system architecture and ConOps flow, ensuring smooth integration across subsystems.
High-level AI inspection flowchart — part of autonomy and controls design
Integrated system architecture — subsystems connected through systems engineering process
As the Autonomy & Controls lead, I focused on making the payload operate independently in microgravity and vacuum conditions. That required real-time inspection, FDIR, and communication handling, all within strict constraints on power, volume, and mass.
I designed the NDE by combining optical and infrared cameras with a machine-learning model to detect defects layer by layer during the 3D metal-printing process. To make this viable for space, I applied transfer-learning methods and conducted trade studies that led us to select the NVIDIA Jetson Xavier NX. After running an STK Total Ionizing Dose (TID) analysis in MATLAB, we added 2 mm of aluminum shielding. The pipeline classified defects such as flash formation, voids, and surface roughness by comparing pre-trained defect models with real-time data from optical and IR cameras during operation.
Demonstration of the CNN-based AI inspection pipeline — optical and IR cameras for real-time defect detection
I also developed the logic for our FDIR system, modeled after NASA’s High-Performance Spaceflight Computing concepts, and mapped out the command and data handling flow. This required me to:
- Understand and define fault levels from 0 to 4.
- Design a system with minimal human intervention from the ground.
- Use a systems-engineering process to decide when the core computer or AI computer should make autonomous decisions based on fault level.
- Define an architecture where NDE implements FDIR and distributes each subcomponent’s role in printing, inspection, and monitoring.
Command and data handling architecture — linking autonomy logic to bus and ground systems
Challenges
The main challenge was working within the BCT X-Sat Venus-class bus constraints: ~60 W of power, strict volume, and capped mass. Every subsystem had to be justified, and trade-offs often forced redesigns.
We also had to ensure the design was practical beyond paper. Concepts had to survive launch loads, operate in microgravity, and remain reliable for two years. Systems engineering reviews and trade studies helped scale ambitious ideas into achievable designs.
BCT X-Sat Venus-class bus — constraints that defined our design space
What I Learned
Because this senior project came right after my military service, I initially led with a directive, top-down style. That approach works in organizations with a strict chain of command, but it was less effective in a peer-based capstone team.
I switched to a collaborative style: align first on mission goals, assign ownership based on each teammate’s strengths, and resolve technical tradeoffs through open discussion. That shift improved trust, execution speed, and subsystem integration across the team.
I also learned the value of collaboration. As an international student, I used to try to prove myself by working harder alone. This project showed me that spacecraft are built through teamwork, not isolation. I could not have reached this outcome without my team’s diverse expertise and shared effort.
I had to constantly switch between mission-level goals and detailed design questions, from autonomy requirements to shielding thickness, thermal margins, and data flow. Systems engineering gave me the framework to connect those levels and ensure every subsystem decision supported the mission as a whole. The most important lesson was that in complex system design, every creative idea must be validated through rigorous feasibility analysis. Creative concepts without engineering validation cannot be used in this industry.
Project Results
Official coverage: Virginia Tech claims 2nd place at COSMIC Capstone Challenge
Publication: AIAA SciTech Forum 2026
- 1st place at Virginia Tech Space Vehicle Design Challenge
- 2nd place nationally at COSMIC Capstone Challenge (presented at NASA Goddard Space Flight Center)
COSMIC Capstone Challenge national result — Team H.A.D.E.S., 2nd place
Certificate of Excellence in Space Vehicle Design — First Place at Virginia Tech