(Deep) Learning to Fly
summary
Covers building autonomous drones that perceive their environment and navigate while moving, focusing on using (deep) learning to enable safe drone racing in cluttered or hostile environments. Describes choosing appropriate technologies and walks through all development stages for teaching a drone to race on its own.
AI-written summary · about
chapters (29)
- 0:00Project intro
- 1:02Drone control
- 2:25Art and tech
- 3:31Autonomous controllers
- 5:01Vision positioning
- 6:49Drone testing
- 8:21User experience
- 10:10Optical flow obstacles
- 11:57Flight stability
- 13:03Visual limits
- 14:53Math models
- 15:34Structure from motion
- 17:17Image processing
- 19:04ML basics
- 20:34Neural training
- 23:01Test results
- 25:12DL for control
- 26:39Depth perception
- 28:04Network scaling
- 30:17Network architecture
- 31:41Hardware-software
- 33:13UI functionality
- 35:21Hardware choices
- 36:33Social impact
- 38:05Automation
- 39:12Design and prod
- 40:39User feedback
- 41:40Future vision
- 42:43Q&A
related talks
45m(Deep) Learning to Fly
48m(Deep) Learning to Fly
49mUsing Deep Learning for Perception in Autonomous Systems: A Perspective from Control Theory
29mKickstart Your Machine Learning Career with AWS DeepRacer
48mFlying services with a drone
01h 51mLearn to Drive (and Race!) Autonomous Vehicles
29mDeep Learning, Neuroscience and the future of AI
01h 35mDeep Robotic Learning