Chapter
Converging towards a Fusion Problem with Large-Scale ML in Autonomous Vehicles
1:42:51 - 1:54:48 (11:57)
Autonomous vehicles are solving complex problems with a fusion of LIDAR, camera, and radar data. The optimization for long-range sensors and the ability to handle end-to-end machine learning are key for solving long-tail challenges unique to trucks.
Clips
Trucks are now equipped with new sensors to optimize sensing horizon that is unique to them with a built-in redundancy for LIDAR cameras and radar.
1:42:51 - 1:46:23 (03:31)
Summary
Trucks are now equipped with new sensors to optimize sensing horizon that is unique to them with a built-in redundancy for LIDAR cameras and radar.
ChapterConverging towards a Fusion Problem with Large-Scale ML in Autonomous Vehicles
Episode#241 – Boris Sofman: Waymo, Cozmo, Self-Driving Cars, and the Future of Robotics
PodcastLex Fridman Podcast
Autonomous vehicles' near range perception systems use LIDAR, cameras, and radar.
1:46:23 - 1:49:47 (03:24)
Summary
Autonomous vehicles' near range perception systems use LIDAR, cameras, and radar. LIDAR is a prominent sensor for redundancy but has shortcomings for long-range sensing due to occlusions, weather conditions, and limited range.
ChapterConverging towards a Fusion Problem with Large-Scale ML in Autonomous Vehicles
Episode#241 – Boris Sofman: Waymo, Cozmo, Self-Driving Cars, and the Future of Robotics
PodcastLex Fridman Podcast
More parts of systems have a property where you want to put more data into it and it gets better.
1:49:48 - 1:54:48 (05:00)
Summary
More parts of systems have a property where you want to put more data into it and it gets better. This has led to a convergence towards thinking of problems as a fusion problem in machine learning.