Skip to Main Content
The use of laser scanners (LS) is investigated for assured navigation of autonomous vehicles, such as unmanned aerial vehicles (UAVs), in structured urban environments. The navigation solution is based on features (e.g., lines) that are extracted from laser scan images. The ultimate goal of this research effort is to develop feature-based techniques that will enable assured navigation capabilities in Global Positioning System (GPS)-denied environments. As a first step for achieving the set goal, detection and isolation of feature motion is investigated in order to exclude moving features from the navigation estimation problem. Relatively fast-moving features are removed using the integration with inertial sensors. This approach allows for the removal of features whose velocities exceed the inertial velocity drift. Autonomous integrity monitoring techniques that exploit redundant feature geometries are applied to detect and isolate slowly moving features (i.e., features with velocities below the inertial drift) for those cases where a single slowly moving feature is present in the laser scanner (LS) field-of-view (FOV) for any given scan. Minimum detectable feature velocity is evaluated based on the actual quality of features observed in scan images. A simulation scenario is implemented to demonstrate that the proposed detection and isolation method allows for reliable exclusion of even a very slowly moving feature (velocities at a cm/s level). Validity of the simulation setup is verified using experimental data.