Aerodynamic parameter identification for aeroassisted vehicle via cubature smoother with expectation-maximization

Abstract

Aeroassisted orbital maneuvers are recognized as a potentially efficient mode of trajectory adjustment due to their significant reductions in fuel consumption and flexible path correction capabilities. The aerodynamic parameter identification during atmospheric flight, which are characterized by trans-regime and hypersonic speeds, is crucial for enhancing the precision and safety of the aerodynamic-assist systems, given their sensitivity to minor perturbations in aerodynamic parameters. This paper introduces a high-precision aerodynamic parameter identification method based on state normalization and sample point generation with a tuning factor. A normalized dynamic and measurement model for aeroassisted maneuvers is established to mitigate the high dynamism of angular states such as heading angle due to wide-ranging positional and velocity changes. An initial data update strategy that accounts for global state errors is proposed, along with an Expectation-Maximization and Cubature Kalman Smoother (EM-CKS) designed to reduce the sensitivity of aerodynamic parameter identification to initial errors. To enhance the robustness of the algorithm to various mission scenarios, a sample generation method incorporating a tuning factor is also proposed, which broadens the adaptability and transferability of the presented method. Simulation analysis demonstrates the method's efficacy in aerodynamic parameter identification across different initial state variations and lift-to-drag ratios, using two typical Martian aeroassisted maneuver scenarios: extensive aerodynamic descent and aerodynamic capture. The results confirm the method's convergence and robustness.