How Machines See
UsesCameras only (Tesla Vision). An occupancy network turns the eight-camera video into 3D occupied/free voxels. No lidar, no radar, no HD map.
Unlike othersWaymo fuses lidar + radar + a prior map. Waabi predicts where space will be occupied in 1–3 s.
Uses6th-gen Waymo Driver: lidar + radar + cameras, fused into labeled 3D boxes with track IDs. Localizes against an HD map.
Unlike othersTesla is cameras-only and has no map. Waabi scores future occupancy, not current boxes.
Cameras only → occupancy network + vision boxes
| Variant | Type | Result | TTC |
|---|