Sensitivity Methods Applied to Orbital Pursuit Evasion
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
A sensitivity method is applied to the orbital pursuit-evasion problem. Techniques are developed to address three particular challenges considered in previous literature: obtaining a saddle-point optimal control solution under nonlinear orbital dynamics; locating barrier trajectories to identify regions of guaranteed interception; and updating a control solution at speeds consistent with real-time control. When applied to a previously published problem, a speedup factor greater than 100 compared to a genetic algorithm is observed. The strengths and limitations of sensitivity methods, applied to the orbital PE problem, are complementary to those of alternative solution techniques such as collocation and random search. To make maximum use of existing sensitivity computations, when driving the barrier indicator to zero, a correction-based method is instead employed to drive the final-time Hamiltonian to zero.
