- 8 Aug 2016
KickoffStart of 2016 Autonomous Drone Elective
The course was kicked off with a lecture from Patrick Christ giving an introduction to the project and the Python libraries available for drone control. The technical challenges of the first day were to get familiar with the drones, implement manual controls and extract camera frames. In the afternoon Laura Bechthold lead a thought provoking case study tackling ethical questions of drone usage and decision making.
Deep Learning SessionDeep Learning lecture paired with hands on exercises
On the second day Dr.Seyed-Ahmad Ahmadi held a lecture on Image Processing and Object Detection, which was complemented by an introduction to Deep Learning by Patrick Christ. After the theoretical input sessions, the students had to train a network to recognize hand written numbers using Nvidia's DIGIT System, before starting to implement their own classifier.
Autonomous Flying AttemptsHackathon and first autonomous flight attempts
After a swift session on PID Controllers by Sebastian Schlecht the students spent their time putting the concepts they learned the previous days into practice. First autonomous flying attempts were made - with varied success.
Ideation WorkshopDesign Thinking workshop to ideate on business cases
After half a day of further development and coding the students engaged in a Design Thinking workshop lead by Dr. Felix von Held and Felix Werle from the Institute for Innovation and Change Methodologies. In a structured approach the students ideated creative business cases related to autonomous drones and presented them in front of the class. As the deadline kept getting closer, the students optimized their algorithms in an all night long hackathon.
Race DayLast minute fixes with the finish line in sight
The final day of 2016's drone elective was kicked off by a keynote speech presenting a current drone-related research project. Following that the entire team went to TUM's Informatics department in Garching to test the developed solutions by competing in a race through a small obstacle course.
About This Course
The ambitious goal in the 'Autonomous Drones Elective' was to program drones that fly autonomously within five days and compete in a final drone race. In order to meet this extraordinary challenge, students from various disciplines – ranging from liberal arts to engineering – acquired basic concepts of Computer Vision, Object Recognition and Deep Learning.
The course was provided by excellent lecturers who supported the course participants with their technical know-how and addressed the bigger picture behind autonomous drone flight. Furthermore, the course participants ideated on drone related business cases in a Design Thinking workshop and evaluated advantages and risks of autonomous drones in a thought-provoking ethic session.