Robust Multi-objective Trajectory Optimization in the CR3BP

Abstract

Orbital mission analysis is an iterative procedure in which several solutions are tested until the one that better fulfils mission objectives is selected. With new missions in the Circular Restricted Three-Body Problem (CR3BP) becoming more frequent, the mission analysis process for trajectories within this model must be improved. So far, most of the research efforts have been devoted to finding the most accurate optimization algorithm for a particular problem, mainly related to the Gateway mission. However, this study presents a novel approach whose aim is to improve the initial phase of mission design by providing a multitude of optimal solutions that cover most of the trajectory possibilities. Moreover, this should be performed without an initial guess input from the user, nor any kind of previously known solution, so that it can be applied to all kinds of problems. To do so, the proposed research performs a multi-objective optimization on direct, manifold and flyby transfers and places the non-dominated solutions in a Pareto front. In this way, the user can easily choose which solution better meets their requirements in terms of ΔV and time of flight. Once the solution is chosen, it can used as a first guess for a further optimization with the Analysis, Simulation and Trajectory Optimization Software (ASTOS), which increases the accuracy and reliability of the results, by being verified in a higher fidelity model. To achieve this purpose, the tool has been carefully designed, selecting the design variables that with the least amount of information, completely define the full trajectories. Then, a mixed integer distributed ant colony optimizer (MIDACO) is used to find the best solutions. The results obtained include two sample cases. First a LEO to L2 southern Halo orbit and then a GTO to 9:2 resonant L2 NRHO, all in the Earth-Moon system. For all cases, a populated Pareto front with dozens of optimal solutions, was obtained and a single trajectory solution was selected for representation purposes.