A universal navigability map building approach for improving Terrain-Aided-Navigation accuracy

An interesting way to improve Terrain-Aided- Navigation (TAN) accuracy may consist in finding the best trajectory which gathers the maximum of information coming from the terrain sensor. This paper proposes a universal approach for improving the TAN accuracy based on a new criterion derived from the fundamental Cramer Rao Lower Bound (CRLB). Under few hypotheses the local contribution of the terrain to the navigation accuracy can be extracted from a recursive expression of the CRLB interpreted in term of information. A navigability map can also be defined in computing this criterion at each node of a regularly spaced grid. This map can be computed for every kind of vehicle using any kind of geophysical sensor. Any classical path planning algorithm can thus be used to find trajectories maximizing the cumulated navigability score, even under mission constraints. This paper demonstrates the main capabilities of this criterion through three different applications. A precision airdrop application from an aircraft equipped with a multi-beam scanner laser and an erroneous embedded Digital Elevation Map shows the ability to take advantage of an error model associated to the map. A marine vessel navigation in a GPS-denied environment illustrates that this criterion performs well for any kind of TAN, in particular for a navigation with a gravity anomaly map (gravimetry). Finally an altimetry terrain following application exhibits this new criterion's ability to outperform the well-known and widely used roughness.