Interview
summary
Velodyne describes how customers apply AI for modular perception to extract lanes, obstacles and object classes from high definition 3D LiDAR rather than hard coding rules, reports that such modules work well but an end to end sensory to vehicle stack is not yet considered safe, and says customers consistently ask for more raw LiDAR data with tradeoffs in cost and form factor. Velodyne provides sensors and data but not AI models and focuses on getting higher density and richer point clouds into developers' hands so systems can improve.
AI-written summary · about
chapters (4)
- 0:00Start
- 0:39What aspects of Al do you think work already?
- 1:09What is your key focus in Data, Machine Learning and Al?
- 1:49What do you like about the Al By the Bay
related talks
19mAI Trends in Intelligent Transport Systems
06mIlluminate the way to enterprise AI copilots with Microsoft and Aisera
32mIsolated Artificial Intelligence: Now or Tomorrow?
50mSelf-driving cars - the future is here....
26mPanel: AI Future - Where Investments Are Being Made
21mLegibility by design
15mStreetViewAI: Making Street View Accessible Using Context-Aware Multimodal AI
31mAI Search for Marketplaces: Experiences & Tech