Uncertainty Quantification for Mars Entry, Descent, and Landing Reconstruction Using Adaptive Filtering

Abstract

Mars entry, descent, and landing (EDL) trajectories are highly dependent on the vehicle's aerodynamics and the planet's atmospheric properties during the day-of-flight. A majority of previous EDL trajectory and atmosphere reconstruction analyses do not simultaneously estimate the flight trajectory and the uncertainties in the models. Adaptive filtering techniques, when combined with the traditional trajectory estimation methods, can improve the knowledge of the aerodynamic coefficients and atmospheric properties, while also estimating a realistic confidence interval for these parameters. Simulated datasets with known truth data are used in this study to show the improvement in state and uncertainty estimation by using adaptive filtering techniques. Such a methodology can then be implemented on existing and future EDL datasets to determine the aerodynamic and atmospheric uncertainties and improve engineering design tools.

Statistical Entry, Descent, and Landing Performance Reconstruction of the Mars Science Laboratory

“…The 3σ uncertainty bounds of the reconstructed positions for all three estimates encompass the independently estimated location. The UKF and adaptive filter have tighter bounds than the EKF, corroborating the expected outcomes when simulated data were analyzed by these estimators [3,35].…”

“…The UKF estimates smaller velocity uncertainties than the adaptive filter in this case, but due to a lack of knowledge in the true state, it would be conjecture to attribute a physical rationale for this. Testing with simulated Mars EDL data in the past has shown that the adaptive filter actually had smaller uncertainties than the UKF [35].…”

Mars Atmospheric Characterization with a ChipSat Swarm

“…For trajectory modeling and assessment of atmospheric characterization capabilities, similar entry conditions to the previously flown Mars Exploration Rovers have been used, with uncertainties extracted from [16]. During atmospheric entry, each ChipSat will perform attitude rate and acceleration measurements using an IMU comprising micro-electro-mechanical systems (MEMS) gyroscopes and accelerometers to reconstruct their trajectories.…”

“…The ChipSat mass and reference area are modeled as 65 and 0.02 2 . The uncertainty ranges are taken from [16].…”

Comparison of Several Nonlinear Filters for Mars Entry Navigation Using Radiometric Measurements

“…Future Mars explorations need the vehicles to land precisely at the prescribed locations to complete the missions, such as sample returns, high-mass entry, human explorations and so on. Mars entry phase is the most challenging part of Mars Entry, Descent and Landing (EDL) due to the complex aerodynamics and atmosphere properties (Dutta et al, 2014), which make the Mars entry dynamic model strongly nonlinear and time varying. Also, the entire Mars EDL phase only lasts a few minutes.…”