The dataset consists of tracking data of over 23,000 vehicles travelling though five different roundabouts in Sydney, Australia. This data was collected by a vehicle outfitted with a ibeo.HAD Feature Fusion detection and tracking system. This system uses 6 ibeo LUX 4 beam, 25 Hz Lidar scanners to identify road users at a range of up to 200m, and has an on-board computer for classification and tracking, in real time.
These tracks are sorted by origin and destination for each vehicle. Data about each detected vehicle includes: X/Y relative positioning (metres), velocity (metres/second), heading (radians), size (width/height metres), classification [bike, car, heavy vehicle, pedestrian], and classification confidence.
When using this dataset please cite:
A. Zyner, S. Worrall, and E. Nebot, “ACFR Five Roundabouts Dataset: NaturalisticDriving at Unsignalised Intersections,” IEEE Intelligent Transportation Systems Magazine, Early access:https://doi.org/10.1109/MITS.2019.2907676
A. Zyner, S. Worrall, and E. Nebot, “Naturalistic driver intention and path prediction using recurrent neural networks,” IEEE Transactions on Intelligent Transport Systems, early access: https://doi.org/10.1109/TITS.2019.2913166