license: mit
tags:
- autonomous-driving
- lidar
- sensor-fusion
- raw-data
Multisensor Unaligned Streams (Test Fleet Alpha)
Overview
This dataset contains ongoing, unprocessed raw binary streams captured directly from the integrated sensor suites of our test vehicle fleet.
To facilitate independent parsing, the raw buffers are segregated by hardware source:
lidar_pointcloud_*- Raw 3D spatial returns.radar_doppler_*- Unfiltered RF velocity data.camera_rgb_*- Raw optical binary buffers (non-debayered).
Important Note: These are unaligned binary blobs extracted directly from the hardware buffer. They do not have standard file headers.
Usage
To parse these streams into a synchronized simulation environment, you must use our internal offline decoding tool. You will also need the static calibration matrix.
- The offline decoder is archived in
offline_decoder_environment.dat. - The multi-sensor calibration keys are stored in
sensor_calibration_matrix.dat.
Do not attempt to parse the stream files without the decoder and matrix, as the output will be meaningless noise.