Collision Free Path Planning based on Local 2D Point-Clouds for MAV Navigation

Abstract

The usage of Micro Aerial Vehicles (MAVs) in different applications is gaining attention, however one of the main challenges is to provide collision free paths, despite the uncertainties in localization, mapping, or path planning. This article proposes a novel collision-free path planner for MAVs navigation in confined environments, while not being dependent on the information of the localization, only relying on 2D local point-cloud data. The proposed backup path planner generates velocity commands for a trajectory-following controller, while guaranteeing a safety distance from all points in the local-point-cloud. The proposed method considers the kinematics of the MAV and can be extended to any robotics application, such as ground vehicles. The proposed method is evaluated in a Gazebo simulation environment and successfully provides a collision-free navigation.