Dissertation in Five Contributions

For readers who want a concise summary, here are five core contributions from my dissertation.

1) Geometry of dynamic output-feedback controllers

I develop a geometric/orbit-manifold viewpoint for controller parameterizations, separating meaningful structure from coordinate redundancy.

2) Intrinsic methods for closed-loop optimization

I derive Riemannian first- and second-order methods for policy design with rigorous convergence behavior.

3) Intrinsic successive convexification (iSCvx)

I extend successive convexification to manifold-valued trajectory optimization in a representation-invariant way.

4) Applications to constrained guidance problems

I demonstrate how these methods apply to practical constrained attitude and trajectory planning settings.

5) Links to distributed geometric optimization

I connect these ideas with consensus and averaging over Lie groups, expanding the distributed optimization perspective.

• 📄 Dissertation PDF
• 🎤 Defense Slides
• 🗂️ Dissertation page