Spotlight on Young Researchers: Smarter choices for complex systems
Designing satellites, medical devices, or autonomous machines is a high-stakes puzzle: every choice can ripple through performance, safety, cost, and certification. One young researcher is tackling this complexity head-on by blending the speed of artificial intelligence with the rigor of formal verification—building tools that help engineers reach safer, more reliable designs faster.
From rehab exoskeletons to satellites
The journey began with a PhD focused on dependability in lower-limb exoskeletons for elderly rehabilitation—systems where any failure could directly affect patient safety. That grounding in safety-critical design now fuels postdoctoral work on automated satellite mission planning and design. The thread running through both is clear: make complex systems that perform reliably when real lives, costly assets, or critical services are on the line.
Put simply, the mission is to build “smart” design workflows—ones that suggest, evaluate, and refine options automatically, without losing the assurance engineers need to trust the result.
AI meets formal verification—inside the CORE VARIANCE approach
At the core of this research is a tight integration of AI-driven design space exploration with formal verification. Within the CORE VARIANCE project, the team develops end-to-end workflows that:
- Capture system variability and constraints through robust variability modeling.
- Automatically generate candidate architectures across hardware and software boundaries.
- Run multidisciplinary optimization to balance performance, cost, mass, energy, and more.
- Embed formal analysis directly into the exploration loop to rule out unsafe or non-compliant options early.
This last point is crucial. Traditional optimization hunts for the “best” design, but may overlook corner cases or subtle interactions. By weaving formal methods into the search itself, unsafe solutions are filtered out before they ever look attractive. The result is a scalable process that doesn’t trade away assurance for speed.
While satellite mission planning is a natural testbed—think payload selection, orbit choices, power budgets, and communication links—the methodology generalizes to any cyber-physical system where traceability, explainability, and regulatory compliance are essential from day one.
Why it matters
Modern systems are mosaics of sensors, software, and actuators operating under uncertainty. Decisions made early—selecting an architecture, assigning functions to components, allocating margins—lock in 80% of lifecycle costs and risks. A workflow that can quickly explore thousands of options while proving that only safe, certifiable designs survive offers a tangible competitive edge. It shortens development cycles, cuts rework, and provides evidence engineers and auditors can trust.
Two worlds, one goal: academia and industry
Working with industry reshapes priorities. Product timelines, tight budgets, and certification gates demand practical, explainable methods that scale. “Theoretical elegance alone is not enough,” the researcher notes. If a technique cannot handle real data, real constraints, and real failure modes, it won’t make it past a prototype.
Academia, by contrast, creates space to question assumptions and push boundaries. That freedom is vital for breakthroughs like coupling formal proofs with generative design tools. But the feedback loop with industry—limited data, cost pressures, deployment hurdles—keeps the work grounded. It ensures the outputs are not only novel, but usable under field conditions.
The collaboration dividend
For companies, partnering with public research opens doors to methods and technologies not yet on the market. It brings deeper analysis, broader exploration, and long-term thinking that de-risks innovation before it hits the factory floor.
For researchers, real-world constraints are invaluable. They expose gaps that controlled lab settings won’t reveal and highlight what must be explainable to pass audits, what must be automatable to fit into toolchains, and what must be traceable to satisfy regulators. This back-and-forth shortens the path from idea to impact—and helps both sides build systems that are not merely “smarter,” but genuinely dependable.
What’s next
As AI accelerates, the pressure to balance speed with assurance will only grow. Expect to see more workflows where optimization and verification are inseparable; where design tools can justify every recommendation; and where traceability isn’t an afterthought, but a built-in feature. Whether for satellites, medical robotics, or autonomous platforms, the future of complex systems design will reward approaches that prove safety and performance together—right from the first sketch.
This young researcher’s work points the way: faster design cycles, provable safety, and decisions you can explain and trust.
