</a>

Autonomous UAV

Path planner, cascaded controller, extended kalman filter...
Explore the docs »

View Demo · Report Bug · Pull Request

Table of Contents

• Table of Contents
• About the Project
  • Introduction / statement of purpose googledocs file
• Setup
• Run
  • Matplotlib planning
  • Control and Estimation Simulator
• Visualization
  • Simulator Control
• Roadmap
• Contribute
  • :wavy_dash: references
  • Contribute on proposed features
• License
• Contact
• Contributors

About the Project

Introduction / statement of purpose googledocs file

My goal during and after this course is to design an autonomous UAS to work under any circumstances for the logistics industry. While this goal is too broad and requires various apprentices in the field of navigation, aeronautics and programming in low-level to higher-level languages, I have focused on different GSC and simulator platforms such as QGC, Gazebo, Matlab, Cleanflight Configurator and FCND among others. I have also used various open-source and proprietary drones (see my hardwares). In this project, I have achieved the following:

• Action pipeline based on bash scripts and webserver to pull and update code changes remotely.
• Python- and ROS-based connection to each nodes
• Use of Kalman and other filters to produce navigatable simulation maps.
• Plotting realtime data

When successfully deployed, I should be able to monitor the vehicle status and react to events remotely through script changes, reducing human involvement in most normal UAS operations.

Setup

overview

lora on drones

Run

Matplotlib planning

as built in the Motion Planner :100:

1. Grid
2. Medial Axis
3. Voronoi Graph
4. Heuristic Graph

---

Control and Estimation Simulator

To use the simulator for control testing, follow these steps :

Visualization

model class

SessionSerializer, MovementSerializer, GlobalPoistionSerializer, GlobalHomeSerializer, LocalPositionSerializer, LocalHomeSerializer, LocalVelocitySerializer, etc

:hammer: I’ve never been a good db person who still loves to dump everything in a single db table, but the data flowing out from the flight data and a flying vehicle can be msasscaring any semblance of solidines of database structues. I expected the db can love me back in selfishless ways, but it turned out that I have to take the bulls by the horn, alas with allmighty and still unsuccesful. Even with the Django’s third-party rest_framework, this project proves that there is a long way to go for me. Hooray!!

Simulator	Details	Front-end integration	Details
	Ardupilot and px4-based simulator		Django, Postgresql, Vue, Axios and Kalman libraries. Live vehicle tracking

Simulator Control

Roadmap

Integrating hardware and simulator is not accomplished so far, due to defective hardware. It took so long to realize most of the issues I got have been direct results of the degrading materials provided, which I had to replace with my own.

And lastly but not singularly, I realized that most of the hard work could have been avoidable once you know what libraries or filters you can use, specific to each and every requirements. I am glad to have discovered these libraries such as kalman.

My next assignment is to mix up and make it work on a real drone and do the transfer learning, with master/slave connectivity on lorawan.

Contribute

Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.

:wavy_dash: references

• Kalman filters, MCL algorithms, Occupancy Grid Mapping algorithm, FastSLAM, GraphSLAM, ROS Navigation Stack 7, Probabilistic Path Planning, Drone FCND, KUKA Path planning, OpenCV, RNN, CNN, Navigator, PID Control
• books referenced in this project:

  Mastering ROS for Robotics Programming Packt
  Learning ROS for Robotic Programming 2nd Edition
  ROS for Beginners, Basics, Motion and OpenCV
  ROS로보틱스 
  Boost.Asio C++ 네트워크 프로그래밍 쿡북]
  

  | | | | | – | – | – | | 프로그래밍 쿡북 | ros for absolute beginners | learning robotics

Contribute on proposed features

To create a PR:

Follow the given link to make a successful and valid PR: https://help.github.com/articles/creating-a-pull-request/

To send a PR, follow these rules carefully, otherwise your PR will be closed:

1. Make PR title in this formats:

   Fixes #IssueNo : Name of Issue
   

   Feature #IssueNo : Name of Issue
   

   Enhancement #IssueNo : Name of Issue
   

According to what type of issue you believe it is.

For any doubts related to the issues, i.e., to understand the issue better etc, comment down your queries on the respective issue.

License

Distributed under the MIT License. See LICENSE for more information.

Contact

Contributors