Reactive obstacle avoidance for Rotorcraft UAVs

Abstract

This article presents a new technique for obtaining a flexible trajectory based on the deformation of a Bézier curve through a field of vectors. This new technique is called Bézier Trajectory Deformation (BTD). The trajectory deformation is computed with a constrained optimization method (Lagrange Multipliers Theorem). A linear system is solved to achieve the result. As a consequence, the deformed trajectory is computed in a few milliseconds. In addition, the linear system can be solved offline if the Bézier curve order is maintained constant during the movement of the robot, which is the common case. This technique can be combined with any collision avoidance algorithm that produces a field of vectors. In particular, it has been developed for artificial potential field methods. BTD is combined with a recently proposed PF method, the Potential Field Projection method (PFP). The resulting technique is tested in dynamic environment showing its real-time performance and its efficacy for obstacle avoidance.

Conflict Detection and Resolution System Architecture for Unmanned Aerial Vehicles in Civil Airspace

“…On-board sensors are the most popular way to provide surveillance for UAVs. This type of surveillances has been assessed, including for cameras (visual light and infra-red), 11 acoustics sensors, 12 and even miniaturized versions of active-sensors like laser-range-finder 13 and RADAR.…”

Multiple-Scenario Unmanned Aerial System Control: A Systems Engineering Approach and Review of Existing Control Methods

“…Kobilarov [116] utilized a cross-entropy method in conjunction with rapidly expanding random trees for object avoidance path generation. Hrabar [117] proposed using an expanding elliptical search method to determine paths for object avoidance.…”

Performance evaluation in obstacle avoidance

“…In this method an unsafe region is established, part of this unsafe region is a diamond shaped region resulting form the error of the stereo system. Other methods use the stereo vision to build a map , (Hrabar, 2011), (Jongho & Youdan, 2012) but do not consider the detection performance.…”

“…Which results in a larger safety zone for large obstacles. In a paper from Hrabar (2011) an elliptical shape is fit around the obstacle and a circle safety region around the UAV. This circle is aligned perpendicular to the flight direction which is propagated linearly.…”