A dynamic inversion-based three-dimensional nonlinear aiming point guidance law is presented in this paper for reactive collision avoidance of unmanned aerial vehicles. When an obstacle is detected in the close vicinity and collision is predicted, an artificial safety sphere is put around the center of the obstacle. Next, the velocity vector of the vehicle is realigned towards an 'aiming point' on the surface of the sphere in such a way that passing through it can guarantee safe avoidance of the obstacle. The guidance command generation is based on angular correction between the actual and the desired direction of the velocity vector. Note that the velocity vector gets aligned along the selected aiming point quickly (i.e., within a fraction of the available time-to-go), which makes it possible to avoid pop-up obstacles. The guidance algorithm has been verified with simulations carried out both for single obstacles as well as for multiple obstacles on the path and also with different safety sphere sizes around the obstacles. The proposed algorithm has been validated using both kinematic as well as point mass model of a prototype unmanned aerial vehicle. For better confidence, results have also been validated by incorporating a first-order autopilot models for the velocity vector magnitude and directions.
Backstepping and dynamic inversion control techniques for automatic landing of fixed wing unmanned aerial vehicles
