Mathematical frameworks, state-of-the-art algorithms at the intersection of astrodynamics, control theory, estimation theory and numerical optimization to advance capabilities in cislunar and deep space.
About the Lab
In the ASCLab, we aim to study and propose solutions to various multidisciplinary challenges. We study multi-body dynamical systems and autonomous systems design methodologies to improve mission architecture, Space Domain Awareness (SDA), observability, adaptability, and robustness.
We also research sensor-fusion, 3D reconstruction, and learning-based algorithms to enable autonomous guidance and control of planetary landers and rovers.
Active Projects
Five concurrent threads spanning SDA design, trajectory optimization, maneuver detection, ML guidance, and sensor fusion.
Submodular D-optimality optimization of lunar surface sensor placements for robust cislunar RSO custody. Integrates terrain-aware bright-body exclusion with IEKF-based FIM tracking over full RSO catalogs using GSDE framework for near-real-time placement.
POD and EDMD reduced-order modal bases from CR3BP trajectory ensembles. Maneuver identification through in-subspace vs. out-of-subspace innovation energy decomposition — separates geometric degeneracy from genuine anomalous signatures. NRHO targets, L2 Southern Halo observers.
GP surrogates with ARD kernels over the optimal cost landscape for low-thrust transfers across NHL/SHL orbit families. Probabilistic reachability level-sets quantify achievable destinations under navigation error without repeated TPBVP solves. Manifold-seeded initialization.
Hybrid gradient-boosting + temporal convolutional network architecture trained with navigation-error-augmented state distributions. Learned guidance policy accounts for realistic navigation noise at inference without real-time replanning. Cislunar CR3BP transfer scenarios.
Blender · Singh · AAS 25-524
Develops intuition and quantitative metrics relating observer constellation geometry to RSO trackability. Multi-fidelity bridge from CR3BP analytical tools to full-ephemeris propagation. Evaluates constellation adaptability under degraded coverage and node-loss scenarios.
Computationally efficient multi-view stereo pipeline for resource-constrained onboard planetary lander systems. Sparse plane hypotheses seed dense depth recovery via iterative photometric consistency refinement, enabling high-fidelity 3D terrain reconstruction under tight latency budgets.
Expand reachability framework to leverage quasi-periodic invariant structures seeded from periodic orbits in the CR3BP. Develop cost functions that define novel and relevant metrics and compare solutions in torus and phase spaces.
Prior Work
Completed projects that established the lab's core methodological foundations.
Alvarado · Earl · Singh
Chan · Singh
Diehl · Singh
Shayda · Singh
Alvarado · Singh
Kim · Singh
Kim · Singh
Kim · Singh
Alvarado · Singh