Fuel-optimal powered-descent guidance (G-FOLD): Acikmese-Blackmore lossless-convexification of Falcon-9-style rocket landing as a second-order-cone program. ~22% less fuel vs baseline, 100% feasible, ~2m landing accuracy, ~9ms solves. Python, cvxpy/ECOS.
Convex MPC for rocket powered-descent guidance (G-FOLD lineage). Open-loop vs closed-loop landing burn, full booster-return profile flown from a real Falcon 9 separation state and validated against public CRS-11/CRS-12 telemetry to within a few percent at every milestone.
To associate your repository with the gfold topic, visit your repo's landing page and select "manage topics."