NETWORK METHODS FOR MULTIPLE GRAVITY-ASSIST MISSION DESIGN

An innovative network model of the gravity assist problem enables the quick discovery and characterization of candidate trajectories from a small set of search criteria. The network elements encapsulate information about individual gravity assist encounters and connectivity. This organization of the astrodynamical information makes it possible to deploy well-established search methods to find sequences of flyby encounters with reduced human effort and in a fraction of the time previously required. The connectivity encoded in the model considers energy feasibility and scheduling constraints. Therefore, paths found using the network algorithms are feasible from both an energy and phasing perspective.

Current initial-guess methods only identify a sequence of planet names that may form a tour. Broad searches over launch date and launch V-infinity (sometimes requiring months of computation time) are currently required to identify realistic paths from each possible sequence. The network approach provides (in a shorter period of time) more detailed initial guesses that include the approximate V-infinity and date of each encounter. These initial guesses can directly generate a set of patched-conic trajectories or initialize existing grid-search tools. The technique can accept fidelity improvements and may be extended for use on other mission types.

A collection of potential gravity assist encounters serve as the network vertices. Keplerian models for connecting the gravity assists in energy and time translate into network edges. Network models of more sophisticated trajectory concepts such as resonant transfers and V-infinity leveraging extend the approach to include more complex paths.

General network traversal algorithms form the basis for gravity-assist trajectory searches. Problem-specific network filtering reduces network size and search times. A detailed discussion of algorithm complexity and problem size is also provided.

The new search technique successfully rediscovers known trajectories from historical gravity assist missions. The network method also identifies preliminary gravity-assist trajectories to the Trans-Neptunian Objects Haumea and Makemake.