Racing Robocars
summary
Community builds sub-$400 miniature robocars like the Donkey Car using Raspberry Pi, a Pi camera and cloud training, and trains networks with behavioral cloning or reinforcement learning (AWS DeepRacer appears in events). Events count ~10,000 participants worldwide; top virtual lap here is 18 seconds; two of five finalists are under 16 and one is 13. Low-cost stacks (OpenMV, Jetson Nano, 2D lidar for ~$69) run inference on-board and reliably navigate taped tracks and outdoor kart courses, sometimes outperforming tuned computer-vision rigs.
AI-written summary · about
chapters (8)
- 0:00Intro
- 2:38The next big thing will start out looking like a toy
- 7:50Video demos
- 10:29DIY Robocar vs. Google/Waymo
- 12:59Types of DIY Robocars
- 18:50Breakthrough technologies
- 22:41Video demo
- 25:12Conclusion
related talks
21mAI, Embedded System and Robotics
27mPinchy the bestest boi
35mGhost in the Machine
57mDrones & AI - What's all the buzz about ?
22mA Fully Immersive Driving Simulator for Autonomous Vehicles
01h 02mAutonomous Exploration Robotics with a Raspberry Pi
25mTesting Autonomous Mobile Robots
24mTransforming Driveware into AI Hardware – Unlocking Future of Joy