Obstacle avoidance system development for the Ardrone 2.0 using the tum_ardrone package

Abstract

For micro aerial vehicles (MAV) to operate in indoor environments, several challenges have been identified such as collision avoidance. This paper aims to present a small scale indoor demonstrator of an indoor collision avoidance system using the Parrot Ardrone 2.0 and the tum_ardrone ROS package. In addition, obstacle detection was developed in order to detect obstacles from the point cloud extracted from the Parallel Tracking and Mapping (PTAM) algorithm. Based on the coordinates of the obstacles, the autopilot computes a safe path for the MAV.