In recent years, flying robots such as quadcopters have gained increased interest in robotics and computer vision research. For navigating safely, these robots need the ability to localize themselves autonomously using their onboard sensors. Potential applications of such systems include the autonomous 3D reconstruction of buildings, inspection and simple maintenance tasks, surveillance of public places as well as in search and rescue systems.

(6 ECTS, changed on 12.7.2012)

December 07, 2012: This course has been distinguished with the Teach Inf Award 2011/12 for the best lecture in computer science in the summer term 2012.

In this course, we will provide an overview of current techniques for 3D localization, mapping and navigation that are suitable for quadcopters. This course will cover the following topics:

- necessary background on robot hardware, sensors, 3D transformations

- motion estimation from images (including interest point detection, feature descriptors, robust estimation, visual odometry, iteratively closest point)

- filtering techniques and data fusion

- non-linear minimization, bundle adjustment, place recognition, 3D reconstruction

- autonomous navigation and exploration of unknown environments

The lecture will be accompanied by a lab course where the students will implement their own visual navigation system. This course is an excellent preparation for a master thesis project in this area.

Exercise sheets will be passed out every other week, containing both theoretical problems and programming exercises (in C++). In an exercise group every other week, we will discuss the solutions to the theory problems and the programming problems. Active participation in the exercises is the requirement for participation in the final exam. This will be written or oral, depending on the number of attendees. The questions will cover all material presented in class.

The practical exercises will be implemented directly on a Parrot Ardrone quadrocopter, so we expect a lot of fun (and broken propellors).

The lecture will be given by Jürgen Sturm.

Lecture: Tuesday, 10:15-11:45, room 02.09.23 (FMI, Boltzmannstrasse 3)

Teaching assistant: Nikolas Engelhard

Lab course/practice: Thursday: 14:15-15:45, room 02.09.23 or 02.09.38 (lab) (FMI, Boltzmannstrasse 3)

Registration: via TUM campus

The oral exam takes place in room 02.09.59. Sign up for a time slot on the list in front of the secretary (room 02.09.52). On the examination day: Please take a seat in the sofa corner in front of room 02.09.52 until we pick you up.

A printer ready version (two sides, six pages per sheet) of the lecture notes can be found here: pdf

Recordings can also be found on the TUM TeleTeachingTool website. The raw recordings (before processing) can usually be found a few days before they have been post-processed as MP4 here.

The participation in the exercises is obligatory. Participation in the robot lab is recommended, but not mandatory.

The video-taped talks from the final presentations are now available online (password required, same as for ICRA proceedings).

• Probabilistic Robotics. Sebastian Thrun, Wolfram Burgard and Dieter Fox. MIT Press, 2005.
• Computer Vision: Algorithms and Applications. Richard Szeliski. Springer, 2010.

• ICRA 2012 Proceedings (password required, ask Jürgen Sturm if lost)
• Lehrevaluation (in German)