The CTV-Dataset (CTV stands for Cyclist Top-View) is a trajectories dataset for cyclist behaviour in mixed-traffic environments (aka. shared spaces). This dataset is meant to enlarge the available datasets in the community, focusing on cyclists as main road users to help the research in understanding and predicting cyclist behaviour in shared spaces. The dataset results from an experiment conducted in TU Clausthal to extract data from possible interaction scenarios with other road users, such as pedestrians and cars, in shared spaces. The scenarios were captured using a drone with 4K (3840×2160) resolution at 29.97 fps to ensure high-quality results. The trajectories were extracted using an in-house developed computer vision algorithm.
The experiment occurred in the sports institute’s backyard, the site was flat asphalt and seemed appropriate for cycling and also to be used as a shared area. We experimented in two different areas: Area I) rectangle layout with approximately 55 × 20 meters, and Area II) L-shape layout with approximately 30 × 10 and 15 × 15 meters, as shown in the above figures.
The CTV-Dataset has a collection of scenarios for cyclist interactions with other road users; cyclists, pedestrians, and cars. The collected recordings are filtered, processed, and the trajectories were extracted using the OfflineMOT algorithm. The scenarios groups are categorized as follows:
- Cyclist free-flow movement
- Cyclist-pedestrian interaction
- Cyclist-cyclist interaction
- Cyclist-car interaction
- Corner interaction (with obstructed line of sight)
- Cyclist-pedestrian on dirt interaction
Each scenario group consists of different formulations. Here are the overall stats:
|Total duration||1.68hrs (101 minutes) of pure scenarios|
|No. of clips||593|
|Total no. of Trajectories||4078|
The description of the experiment and the specification of the dataset are published in the 26th IEEE ITSC 2023 conference, you can find the paper here. Please cite the dataset using the following paper reference:
We would like to thank Merlin Korth for labeling the training set and Gustav Baier, the drone pilot. Also, thanks to university members who volunteered in the experiment, and the Sports Institute members for their support in experimenting on the university campus.