Active Terrain Relative Navigation with Surface Mapping Sensors

This work presents a three-dimensional surface map-informed Active Terrain Relative Navigation (ATRN) system for spacecraft precision landing and extra-planetary surface mobility applications. Rotationally and scale-invariant 3D feature extraction and identification methods are introduced, facilitating both global and relative localization in real-time. Numerical spatial gradient and Laplacian methods exploit the surface statistics of the local terrain to identify points which exhibit high degrees of slope and curvature. In contrast to these derivative-based approaches, the proposed Zernike Invariant Gradient (ZIG) 3D feature extraction method is fundamentally built on an integral technique and is inherently robust to noise. Reliable detection and description of 3D features enables correspondence between successive digital elevation map (DEMs) and onboard global reference maps. Resultant fused high-resolution maps inform hazard detection/avoidance and safe-site selection activities for precision landing applications and route planning for surface mobility applications. The architecture of a standalone C++ flight software library that implements the proposed methodologies is also presented. Performance of the proposed ATRN system is evaluated via extensive testing with synthetic digital elevation maps in real-time hardware-in-the-loop (HITL) simulations as well as data obtained during helicopter flight testing of the NASA Goddard Hazard Detection LiDAR (HDL).