Desensitized Ensemble Guidance

A novel stochastic optimal control problem formulation along with a computational framework is developed, based on a new interpretation anchored in probability density functions, that leverage a desensitization strategy with the use of Fokker-Planck equations and recently developed recovery ensemble control. The formulation is applied to derive a desensitized optimal control framework for Zermelo's boat under parametric uncertainty and successfully address a state-constrained re-entry trajectory optimization problem, accounting for navigational, epistemic and aleatoric state-dependent stochastic uncertainties. In order to affirm the efficacy of our approach, we conduct extensive Monte Carlo simulations, offering insights into covariance evolution and the first-order sensitivity of the probability density function.