Robust optimization of impulsive orbit transfers under actuation uncertainties

Abstract

Having some initial and final conditions, it is possible to find a solution for an impulsive orbit transfer, using various trajectory optimization methods. However, there is no guarantee for solution to correspond to the real world condition since there are many possible uncertainties like unpredictable malfunctions in actuation of thrusters that may cause mission failure. The approach of this paper is to develop a method to find a trajectory less sensitive to uncertainties and the motivation is to make a good tradeoff between optimization accuracy and run time. In this paper, novel methods are introduced to make the optimization fast, globally optimized, constraint free and full perturbed. In addition, there is no need to guess an initial guess for the solution because of our use of genetic algorithm. The presented method has two parts for each step of optimization. First part calculates an injection guaranteed trajectory using Lambert's method and second part handles uncertainties over this trajectory. Monte-Carlo method is used for sampling. To reduce the computational cost, a surrogate model is calculated and used in each step of optimization. The result of this robust optimization is a trajectory in which uncertainties make less injection error variance.