Nonlinear Geometric and Differential Geometric Guidance of UAVs for Reactive Collision Avoidance

摘要

Local collision avoidance for safe autonomous Unmanned Aerial Vehicles in an urban environment has been investigated. A nonlinear differential geometric guidance law based on 'collision cone approach' and 'dynamic inversion' was developed and successfully simulated. Both linear 'aiming point guidance' (APG) and nonlinear APG algorithms have been developed and validated from 3-D simulation studies. Finally a first-order autopilot was incorporated to provide satisfactory guidance with up to 0.2 second delay.