Division of Electromobility

Research

Electric, autonomous, connected!

Trajectories of Vehicles and Persons

Description

The dataset consists of trajectories of vehicles and persons created from video recordings over a time of twenty hours at four different junctions. In the dataset the trajectories of 14998 cars, 783 trucks, 380 buses, 235 motorcycles, 498 bicycles and 1459 persons are contained

The junctions are shown in the following figure.

Dataset

The dataset consists of five folders containing the csv files (trajectory_logx.csv) with the trajectories. The folders are named after the respective junction shown in the preceding figure. The csv files for each junction are stored in several numbered logs, each corresponding to a period of about 20 min. The trajectories are stored in the logs for each frame in the images in which the respective object was detected and tracked. The trajectories are stored in the following format.

FieldContent
trackID Unique ID for each recording to identify individual trajectories for subsequent analysis
observation_frame

Number of the frame in the recording for the current point of the trajectory, all other information refers to this point in time only, corresponds to the timeline of the video

classClass of the object (car, truck, bus, motorcycle, cyclist, pdestrian)
position_xx-coordinate in the top view of the junction (can be used to display the trajectories)
position_yy-coordinate in the top view of the junction (can be used to display the trajectories)
distanceDistance traveled between last and current image (in meters (m))
velocityDetermined speed of the road user (in meters per second (m/s))
yawDirection of movement of the road user in the top view (in degree (°))
x1

x-coordinate of the upper-left corner of the bounding box (can be used to display the results of the detection)

y1y-coordinate of the upper-left corner of the bounding box (can be used to display the results of the detection)
x2

x-coordinate of the lower-right corner of the bounding box (can be used to display the results of the detection)

y2y-coordinate of the lower-right corner of the bounding box (can be used to display the results of the detection)

 

The positions position_x and position_y correspond to the positions in the top view of the junctions and can be used to display the trajectories as shown in the following figure.

The coordinate system for the top view is shown in the following figure.

Besides the dataset a Python script is included which can be used to display the trajectories.

Download

The Python script and the dataset (in folder Data_in) as well as related information (e.g. readme, installations instructions) are available under

Download Dataset with Phython Script

A version which only contains the csv files with the trajectories is available under

Download Datadet without Python Script

Acknowlegdement

The dataset was created as part of the project

Concepts for Integration of Bicycle Traffic in Future Urban Traffic Structures with Autonomous Vehicles

This project was funded by the German Federal Ministry of Transport and Digital Infrastructure from resoruces of the the National Cycling Plan 2020 under grant VB1905.

Go to top