An algorithm for dynamic obstacle avoidance applied to UAVs

Abstract

This research focuses on developing a navigation method for mobile robots to effectively avoid moving obstacles while accurately tracking a desired path. The approach introduces an enhanced velocity field that incorporates hydrodynamic theory tools. Initially designed for the 2D case, the method is subsequently extended to the 3D scenario by introducing vector field extensions and rotations.

To validate the proposed scheme, experiments are conducted using a UAV model tasked with tracking a circular contour. The control system employs two PD controllers for regulating the vertical position ($z$) and yaw angle ($\psi$), while the roll ($\varphi$) and pitch ($\theta$) angles are controlled using a nested saturation method.

The numerical results demonstrate the successful achievement of the tracking objective, even when a moving obstacle crosses the reference path. Notably, this study considers the scenario where an obstacle approaches the vehicle from behind, which is often overlooked in similar investigations. This aspect is examined in both the 2D and 3D cases.

Subsequently, the proposed navigation method is tested on a quadrotor vehicle, yielding favorable results.