Chance-Constrained, Drift-Safe Guidance for Spacecraft Rendezvous
Andrew W. Berning, Ethan R. Burnett, Stefan Bieniawski
TL;DR
This work tackles robust, drift-safe spacecraft rendezvous guidance under uncertainty by formulating a direct-collocation, chance-constrained optimization problem that enforces passive safety via keep-out constraints. It integrates a closed-loop uncertainty quantification (UQ) framework—employing exponential correlated random variables and the Gates maneuver model—with a linear covariance propagation scheme to efficiently assess dispersion effects. The optimization is solved through successive convexification, recasting the nonconvex problem as a sequence of SOCPs and using a Lambert-style correction for burns guided by current navigation estimates. Demonstrations in the LEO regime show drift-safe, 3-sigma containment of dispersed trajectories and substantial improvements in propellant or time when comparing minimum-propellant and minimum-time objectives, with potential applicability to GEO and NRHO missions.
Abstract
A robust drift-safe rendezvous trajectory optimization tool is developed in this work, with applications to orbital rendezvous and proximity operations. The method is based on direct collocation and utilizes a sequential convex programming framework, and is extended from previous work to include passive safety constraints. The tool is then paired with a dispersion analysis framework to allow trajectories to be optimized subject to plant, navigation, and actuator uncertainties. The timing, direction, and magnitude of orbital maneuvers are optimized subject to the expected propellant usage, for a given navigation system performance. Representative trajectories are presented for the LEO flight regime, but the approach can also be applied to GEO and NRHO with minimal modification.