Onboard Trajectory Generation for Entry Vehicles via Adaptive Multivariate Pseudospectral Interpolation

Since the beginning of the Apollo space ight program, entry guidance has been widely treated by engineers and researchers.The rst, successful approach, used for several programs (Apollo, Space Transportation System [1]), was based on the planning of an entry trajectory in terms of the drag-velocity plane.The rationale for this choice resides in the fact that the typical environmental constraints (dynamic pressure, heat ux and load factor), as well as the range-to-go, can be eciently represented in this drag-velocity plane.The longitudinal guidance can then be derived in several ways.For instance, assuming the equilibrium-glide approximation [2], extracting the longitudinal states (altitude, speed, ight-path angle) from the drag acceleration and its derivatives [3], or implementing constraints-tracking guidance schemes [4].Similar results can be obtained if the dragvelocity plane is replaced by the drag-energy plane [57].In any case, approximations, disturbances and modeling errors make the use of a feedback controller necessary to track the scheduled nominal