A new method for performing entry, descent, and landing trajectory reconstructionis presented as an extension of the standard inertial navigation approach ofintegrating inertial navigation unit data. The method, Inertial Navigation StatisticalTrajectory and Atmosphere Reconstruction (INSTAR), provides statisticaluncertainties for the reconstructed trajectory parameters by incorporating inertialnavigation and Monte Carlo dispersion techniques. It also permits the inclusionof redundant data by refining the dispersed trajectories with those thatsatisfy the redundant observations to within specified tolerances. The primaryadvantage of inertial navigation over typical existing filtering techniques is theindependence from aerodynamic and atmospheric models. The INSTAR conceptalong with a demonstration using flight data from the recent Mars ScienceLaboratory mission is presented. Initial conditions and acceleration biases aredispersed from nominal using Monte Carlo techniques, and the subset of trajectoriesthat land near the reference landing site are used to update initial conditionsand acceleration biases and to obtain trajectory statistics. These biases arethen compared to the acceleration biases observed in the flight data. Eventually,the process will extend to other redundant data types and to atmospherereconstruction, a process that relies on trajectory parameters and their uncertainties.
展开▼
