Tensor Halo is the world’s highest-resolution automotive-grade lidar, emitting an unprecedented 25.6 million beams per second. Powered by advanced AMD and Broadcom chips, Tensor Halo processes and transmits more than ten times the resolution of most lidars found on current robotaxis.
Tensor Sentinel blind-spot-lidars are the four near-range, ultra-wide-view blind-spot lidars, each guarding a side of the Robocar.
Utilizing exceptional chips from ams OSRAM and Hamamatsu Photonics, the Halo hyper-lidar detects objects with just 10% reflectance up to an impressive 1,000 feet away, effective for high-speed highway driving scenarios.
The Tensor Halo hyper-lidar delivers unmatched redundancy and functional safety. Every stage of its pipeline — from measurement and computation to transmission and power — is fully backed by redundant systems, ensuring continuous, safe operation even if a component fails.
Halo stands as the most advanced and safest automotive lidar solution in the world today.
The Tensor Sentinel blind-spot lidar gives no compromise in field of view for safety, covering 180° horizontally and 110° vertically at the same time, capable of seeing directly beneath the vehicle down to -90° towards the gravity direction.
Unlike off-the-shelf lidars which degrade significantly within 20 in, it detects objects as close as 4 in without precision loss, yet achieving over 2.5 times higher resolution.
Each laser beam from Halo and Sentinel lidars captures multiple laser-returns from objects, a critical capability that allows them to see through challenging conditions such as rain, fog, snow, or dust.
Unlike conventional lidar systems that discard valuable signal data through basic Time-to-Digital Converters, Halo and Sentinel leverage advanced Analog-to-Digital Converters from Texas Instruments.
This preserves the original signals, allowing advanced algorithms to more effectively distinguish true signals from noise, while also reliably measuring reflectance — information that helps the AI accurately identify the material properties of objects.