An edX MOOC that introduces basic concepts for autonomous navigation of quadrotors started on May 5, 2015. While it is an advanced level course, the first week was a gentle, general introduction so there's time to catch up if you join now.  

IProgrammer has a fondness for quadrotors and you can find plenty of examples on this by following the links at the end of this article.

The 8-week course Autonomous Navigation for Flying Robots goes further however and engages students in flight experiments. You don't need your own quadcopter as the course provides a browser-based simulator for which you'll write code snippets in Python. The code is however compatible with the Parrot Ardrone quadrotor. So, to quote the course description, if you happen to have a Parrot Ardrone you can try out your solutions for real!

The MOOC, which ran for the first time last year comes from the Computer Vision group at the Technische Universität München and is based on an award-winning lecture. 

In this video Dr Jugen Stürm sets the scene and explains what students will learn:

Requiring around 4 hours per week, it consists of lecture videos, interactive quizzes and hands-on programming tasks. The topics covered are:

• 3D geometry
• probabilistic state estimation
• visual odometry, SLAM, 3D mapping
• linear control 

By the end of the course students should be able to: 

• Understand the flight principles of quadrotors and their application potential
• Specify the pose of objects in 3D space and to perform calculations between them (e.g., compute the relative motion)
• Explain the principles of Bayesian state estimation
• Implement and apply an extended Kalman filter (EKF), and to select appropriate parameters for it
• Implement and apply a PID controller for state control, and to fine tune its parameters
• Understand and explain the principles of visual motion estimation and 3D mapping

Like other edX MOOCs students are encouraged to pursue a Verified Certificate at a cost of $50 but you get full access to all the courseware for free.

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info