UAV, Path Planning, Artificial Potential Field, ROS, Perception System
Abstract
This project addresses the design, development, and implementation of a system for trajectory planning and autonomous navigation of Unmanned Aerial Vehicles (UAVs) capable of avoiding obstacles. Initially, the necessary routines are developed for the vehicle’s perception/mapping system, using the ROS (Robot Operating System) environment, following that, an unusual approach to the Artificial Potential Field (APF) algorithm used in the trajectory planning stage is proposed, implemented in ROS, and being capable of real-time collision avoidance with obstacles. The entire system developed was tested in the RotorS simulator, an UAV Gazebo simulator, with a dynamic model of a quadrotor drone (Hammer X8B). System validation was carried out through simulations based on realistic, yet not complex and controlled scenarios, with the main purpose of verifying the effectiveness in avoiding obstacles and also the behavior of UAV (drone) control for trajectory tracking. As it is developed in the ROS environment, the system is ready to be deployed on a real drone.
