The Role of High Definition 3D LiDAR in Self Driving Cars
summary
Velodyne describes building rotating multi-channel time-of-flight LiDAR that produces instantaneous 3D point clouds—400,000 pulses per blink enabling real‑time SLAM and mapping—and shows fielded hardware from DARPA runs to mining trucks. Key specs given: 200 meters range, 32 channels in an 800 g automotive form factor, target <$500 volume cost. Claimed system-level impact: fleets using HD 3D LiDAR report far fewer disengagements (Google ~2,000 km vs typical 40–50 km; Cruise climbing into 100–200 km), and potholes, lane markers and small debris become measurable.
AI-written summary · about
chapters (21)
- 0:00Introduction
- 0:35LiDAR is here to stay
- 2:51Lessons Learned
- 3:56DARPA Challenge
- 5:07Map Building
- 5:35Urban Challenge
- 6:08Lane Keeping
- 8:08Lane Markings
- 9:03Poor Ambient Lighting
- 10:18skewed intersections
- 11:09encroachment radar
- 11:32unmapped objects
- 12:07LiDAR performance
- 13:23Driving complexity
- 15:26The market is changing
- 16:48Only one company left
- 17:03Nontraditional automakers using LiDAR
- 17:36Where is LiDAR going next
- 21:01Machine learning and reflectivity
- 23:38AI infrastructure
- 24:56LiDAR data
related talks
16m"Putting a New Spin on
21mThe race to replace human drivers
31mDeep learning for self-driving vehicle technology
42mProduct Management For Autonomous Cars
51mTeaching Autonomous Vehicles to Drive with Visible and Non-Visible Light Simulation
27mDriving Autonomous Vehicles Forward w/Real World Applications
20mCarMap: Fast 3D Feature Map Updates for Automobiles
26mFrom Panoramic Vision to Lightweight 3D Worlds for Semantic Visual Tracking