ROBUST HIGH-FIDELITY GRAVITY-ASSIST TRAJECTORY GENERATION USING FORWARD/BACKWARD MULTIPLE SHOOTING

A common need among trajectory designs is the conversion of a trajectory from low-fidelity models (e.g. patched conic or zero sphere-of-influence methods) to high-fidelity (i.e. higher order acceleration) models. One favorable solution is to approach the problem in terms of control variables, which iteratively satisfy a set of feasibility constraints at match points. The control variables are associated with key events in the trajectory topology, such as gravitational assists. The match points are associated with points in deep space between events. This procedure, involving forward/backward numerical integration and multiple shooting, enables robust convergence by maintaining the trajectory topology during the differential correction process. This paper assesses this benefit in comparison to single shooting approaches and presents a generalizable implementation using the commercial package Systems Tool Kit.