Optimal Information Filtering for Robust Aerocapture Trajectory Generation and Guidance

Abstract

Entry flight is a critical mission phase of planetary aeroassist problems. During atmospheric flight, aleatory-epistemic uncertainty and environmental factors reduce the accuracy of predicted future states for precision targeting. This problem has been approached historically with closed-loop guidance rooted in certainty equivalence. This property separates estimation and control problems, allowing each to be considered independently. In other concept studies, an observer model is neglected altogether in favor of assuming perfect state knowledge. However, a flight system will inevitably have imprecise state information and variability in its underlying dynamics and measurement models. Systemic uncertainty is a fundamental limitation of existing entry guidance approaches. This work seeks to overcome these challenges by posing aerocapture as a robust optimization problem. The cost objective of the maneuver is reformulated to account for uncertainty in atmospheric structure, vehicle performance parameters, and state estimation accuracy using an observer-based consider filter. An expected value performance cost is developed from anticipated measurement conditioning effects. A rapid solution methodology is illustrated using explicit integration strategies with a parameterized control structure. Results for a Mars aerocapture concept study show improvement in the postcapture orbit accuracy with low computational overhead.