From Modular Spacecraft to AI-Driven Engineering
Speaker: [CEO of Planetary Utilities]
Event: Professional Event / Conference Presentation
Overview of Key Themes
The speaker, the CEO of Planetary Utilities and a former JPL employee, presented a compelling vision for the future of spacecraft and systems engineering. The talk centered on two interconnected themes: the necessity of a modular, open-source approach to hardware design for large-scale space systems, and the transformative potential of integrating AI and web technologies into the engineering workflow. The presentation detailed the journey from developing a physical, modular spacecraft prototype to creating a sophisticated software architecture that leverages AI as an accessibility layer for complex digital twin models, ultimately aiming to create a more efficient, collaborative, and intuitive engineering environment.
Key Points of the Presentation
- The Need for a New Paradigm in Spacecraft Manufacturing: Current spacecraft are built like bespoke items in a workshop. To industrialize space and deploy hardware faster, a shift is needed towards large-scale, modular systems. The speaker’s company, Planetary Utilities, developed a modular space system architecture featuring a central “backbone” to which various functional modules (power, propulsion, etc.) can connect, exchange power, and communicate. This allows for flexible and scalable mission design using standardized building blocks.
- A Story of Resilience: The first prototype of this modular system, built in the speaker’s front yard, survived the devastating Eaton fire that destroyed his home. Seeing the prototype still standing amidst the destruction served as a powerful confirmation to the team that they were on the right path. The fire-tested prototype now stands in their new office entryway.
- Shifting Focus to Software and AI: Recognizing the industry’s current state, the company pivoted to focus more on the software architecture needed to design and manage such complex systems. The core problem identified is the immense cost and complexity of current engineering simulations (e.g., a single spacecraft docking can cost $1 million in simulations). The solution proposed is to bring modern software engineering practices, like CI/CD frameworks, into hardware engineering.
- The Role of AI as an Accessibility Layer: A key insight is using AI not to replace deterministic engineering models, but to act as an “accessibility layer.” Engineers often forget the intricate details of models they created years prior. An AI agent can be taught how to use these models, allowing users to interact with them through natural language, ask questions, and retrieve information without needing to recall every technical detail. This democratizes access to complex engineering data.
- Leveraging OpenMBEE and Web Technologies: The architecture is built upon OpenMBEE (Open Model-Based Engineering Environment) principles, inspired by the federated and collaborative nature of the internet. By using web services, REST APIs, and SysML v2 as a common communication layer, the system can integrate with various domain-specific, web-based tools like Onshape (for CAD), NVIDIA Omniverse (for visualization), and Jama (for requirements management). This creates a seamless “digital thread” where data flows between different stages of the engineering lifecycle.
- Demonstration of the AI-Powered Workflow: The speaker demonstrated a chat-based interface where a user could design a spacecraft using conversational commands. The system would build a SysML v2 model in the background, generate CAD geometry, run simulations for power and thermal analysis, and export the results to different platforms. This interface allows for rapid iteration and makes complex design tasks accessible even to non-experts.
Key Quotes
- “My hunch is that we will do space very differently in the future… we also need to start building large spacecraft.”
- “We came to that place, and we saw this prototype still standing there, and that gave us a big confirmation that we’re on the right path and that this is something that we have to do.”
- “We need to go and bring software engineering practices into the hardware engineering. There is no other way.”
- “You can teach an AI agent very easily how to use [a model] and so you can just go back and say, ‘hey, tell me how to use this model’ or, ‘you know, can you give me this answer to my question?'”
- “The internet is already the greatest example of how a federated, collaborative system should work. And so internet technologies and web services are basically the main example of how we can build great engineering.”
Action Points for the Audience
- Embrace Modular Design Principles: When developing complex hardware, think in terms of standardized, interoperable modules with clearly defined interfaces. Finalizing interfaces early in the design process is critical to enabling parallel work and accelerating project timelines.
- Integrate Modern Software Practices: Adopt software engineering best practices, such as version control, CI/CD, and automated testing, for your hardware simulation and modeling workflows. This is essential for managing complexity and ensuring reliability.
- Explore AI as an “Accessibility Layer”: Consider how Large Language Models (LLMs) and AI agents can be used to create intuitive, conversational interfaces for your complex engineering models and data. This can drastically lower the barrier to entry and improve collaboration across teams.
- Advocate for Web-Based, API-First Tools: Prioritize engineering tools that are web-based and offer robust REST APIs. This architecture is key to building a truly integrated, federated, and collaborative digital engineering ecosystem that is not tied to a specific machine or location.
- Contribute to and Leverage Open Standards: Support and participate in open-source communities like OpenMBEE. Contributing to and adopting shared standards, protocols, and integrations benefits the entire industry by preventing redundant work and solving common problems collectively.
