Skip to content

Geometric Controller with Artificial Potential method for trajectory planning of a quadrotor UAV drone

Notifications You must be signed in to change notification settings

gaetanotorella/FSR_FinalProject

Repository files navigation

Geometric Controller with APF method for UAV

This project presents a geometric controller to manage the position and attitude of an Unmanned Aerial Vehicle (UAV). The developed control system aims to ensure high stability and reliability during flight. The Artificial Potential Field (APF) method was implemented for obstacle avoidance, providing a solution both in offline mode, for preemptive path planning and in online mode, for real-time adaptation using exteroceptive sensors. The simulations were conducted in two distinct development environments: Simulink and ROS/Gazebo. Simulink was used to model and simulate dynamic controls, allowing for a detailed analysis of system performance under controlled conditions. ROS/Gazebo provided a realistic simulation environment in which drone behavior was tested in complex scenarios that included various obstacles. The simulation results indicate that the proposed system can maintain safe and precise navigation even in the presence of unexpected obstacles. The combined approach of geometric control and APF method has proven effective in avoiding collisions, significantly enhancing the drone’s capability to operate autonomously in dynamic and potentially complex environments. This study confirms the validity of the proposed method and suggests its applicability for future implementations in autonomous drones intended for critical missions such as surveillance, inspection, and goods delivery.

Installation Instruction - MATLAB/Simulink


To clone the Matlab branch of the repository

git clone https://github.com/gaetanotorella/FSR_FinalProject.git

Installation Instructions - Ubuntu 20.04 with ROS Noetic


To clone the ROS branch of the repository

cd ~/catkin_ws/src
git clone -b ROS --recursive –submodules https://github.com/gaetanotorella/FSR_FinalProject.git
cd ..
catkin build
source devel/setup.bash

ROS simulation


  1. Launch Gazebo and spawn the UAV

    roslaunch geometric_controller UAV.launch world_name:=APF_world
    

    default world is "APF_world", other worlds are in "/geometric_controller/worlds"

  2. Launch geometric controller

    roslaunch geometric_controller geometric_controller_apf.launch
    
  3. Launch APF method

    roslaunch geometric_controller apf_method.launch version:=offline x_goal:=0 y_goal:=0 z_goal:=2
    

    available version are "offline" (only for "APF_world") and "online". It's possible to choose desired position by passing x y z coordinates.

About

Geometric Controller with Artificial Potential method for trajectory planning of a quadrotor UAV drone

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published