Suborbital Reentry Uncertainty Quantification and Stochastic Optimization

Abstract

Suborbital space launch vehicles are used for atmospheric science, microgravity experiments, and soon, space tourism. However, the uncontrolled reentry of these vehicles into the atmosphere is a potential safety concern to both buildings and humans on the ground as well as the valuable payloads onboard. Implementing active downrange and crossrange control of vehicle reentry is costly, and this paper presents a software-only solution for reducing the probability of a landing zone constraint violation. Through a combination of linear covariance uncertainty quantification and clustering optimization, we solve for a small, finite set of optimal ascent reference trajectories for the launch vehicle to follow during the boost mission phase. Each reference trajectory only differs from the nominal by small crossrange and downrange perturbations, allowing the vehicle to still meet all mission performance objectives. The best reference trajectory may be chosen on the day of flight based on current atmospheric conditions. The result is a landing position probability density function that has been ‘shaped’ to avoid dangerous landing zones, increasing the probability of a successful mission with no physical changes to the space vehicle and only minimal changes to its flight controls software.