- Safety software for Volvo Cars
- In-car and cloud software services for autonomous driving
Zenseact Open Dataset (ZOD)
Zenseact Open Dataset (ZOD) → Multimodal AD dataset Dataset components
- Frames - 100k curated camera-LiDAR pairs (images) with 2 secs of surrounding LiDAR data and GNSS
- Sequences - 1473 20-second video sequences with complete sensor data for 20 seconds
- Drives - 29 multi-minute drive sequences with full sensor coverage for a few minutes
Sensor setup
- Hi-res camera - 120 deg FOV RGB camera
- LiDAR - 1x Velodyne VLS128 and 2x Velodyne VLP16
(timestamp, x, y, z, intensity, diode index)← Point clouds
- High precision GNSS/IMU - OxTS
-
(timestamp, lat, long, alti, head, pitch, roll, vel, acc, ang_rate, poses) - Other consumer-grade sensors
- Vehicle data - Steering wheel angle, acceleration/brake pedal ratio, turn indicator status, satellite positioning data etc.
Research Publications
- NeuRAD
- ZOD
Questions and Doubts
- Pitch and roll of a self-driving car?
- Check point clouds
- Difference between camera and LiDAR?
- What is
VulnerableVehiclein the ZOD dataset?