Pedestrian Dynamics Data Archive

Paul Geoerg , Jette Schumann , Stefan Holl, Maik Boltes

DOI: 10.34735/ped.2017.1
Short URL: http://ped.fz-juelich.de/da/2017sime

The presented studies on the impact of individual capabilities on the characteristics of pedestrian dynamics were performed within the joint research project ”Safety for people with physical, mental or age-related disabilities (SiME)”. We kindly ask the user of the data to refer to the cited publications at the end of this page.

Emergency exits as bottlenecks in escape routes are important for designing traffic facilities. The assessment of performance of egress routes is commonly a result of methods based on studies conducted under laboratory conditions with homogeneous crowds. However, this assumption is unlikely, as many crowds consists of individuals with widely distributed skills, prowess, motivation and abilities. We analysed the influence of presence of participants with different types of disabilities on movement through a corridor and a bottleneck.

The results presented at the end of this page suggests that the presence of participants with recognisable disability will influence the characteristics of the fundamental diagram which is largely due to anticipation of movement speeds caused by considerate behaviour. Surprisingly, there are differences in reproducing this effect by consideration of different crowd compositions and their implications on the validity of the specific flow concept.

The movement studies were conducted on two days in an industrial hall in Wermelskirchen-Dabringhausen (Germany) in June 2017.

A more detailed description of the data can be found in the full documentation.

Twelve studies with more than 145 single runs and overall 252 participants with populations composing of people with and without disabilities were performed. In addition studies without any people with disabilities have been conducted. Two geometric configurations with different widths of 0.9 m, 1.0 m, 1.1 m and 1.2 m were investigated: a corridor and a bottleneck as seen in the figures. The length of the bottleneck was constant with 2.4 m. The boundaries were built from wooden three-layer panels with a height of 2.0 m. A waiting zone of around 30 m² was located at x = [−18, −12] with an initial density of around 3/m².

We differ the population of the studies into two basic subpopulations: the participants with disabilities (PWDs) and non-disabled participants (NDP). The mean age of PWD was 47.57 ± 6.99, the mean age of NDP was 35.93 ± 16.26. The mean age of the reference population was 32.07 ± 15.50. The heights of all participants range from 1.45 to 2.04 m with a mean of 1.74 ± 0.1 m.

A more detailed description can be found in the full documentation

All data is provided in ZIP archives (two camera views, extracted trajectories of each person and collected IMU data of wheelchair users). The beginning of the filename is chosen based on the study as given in Table 3 of the full documentation followed by a number identifying the run configuration ([00-08] for bottleneck studies, [01-08] for corridor studies).

• Metadata: Metadata for the experiment in json form. [55 KB]
• Trajectories: Trajectories as txt files in 25 fps. They are provided as combined trajectories of three overlapping cameras perspectives to cover the movement over the entire geometry. [133 MB]
• Trajectories: Trajectories as h5 files in 25 fps. They are provided as combined trajectories of three overlapping cameras perspectives to cover the movement over the entire geometry. [198 MB]
• Camera 1 (X3000): Overview camera capturing the central experiment area at 25 fps. [1.8 GB]
• Camera 2 (GoPro3): Overview camera capturing the starting area at 25 fps. [1.9 GB]
• IMU files: Naming of the follows the pattern: [Study Prefix][Run] [PWD ID][a/b] with a indicating the upper sensor and b the lower sensor on the backrest. The second sensor for the person with PWD ID=38 could not be synchronized with the camera data and therefore IMU data for only one person are provided.

A more detailed description of the data including the population can be found in the full documentation.

Data from this repository was used for Geoerg 2021 and Schumann 2021. The raw trajectory data is provided in this repository. To extract the data for the analysis presented, please follow the description provided in the methods section of Geoerg et al., 2022. The raw data can be used to derive the statistical measures for the individual speed and distance to neighbors characteristics. The raw data are provided for each study sorted by persID and measurement area.

An additional exploratory data set is provided that includes distinct behavioral actions (e.g., waiting, turning, hand gestures) observed in a subset (see table below) of the raw video data. A defined hand-coding procedure was used to detect and quantify individual behaviours as they would be objectively perceptible to an external observer (e.g., hand gestures, turning). Distinct behaviors were coded for the following conditions and trials.

Coding of observable behaviour was performed by one rater using a coding manual. Each objectifiable distinct behaviour in a measurement area was coded with a specific character (e.g., ‚A‘ for any occurrence of observed conversations, ‚B‘ for any occurrence of contact between upper parts of the body, etc.). Please note that not all coded behaviours of the coding manual were considered for analysis Geoerg et al., 2022.

This work has been performed within the research project “Safety for people with physical, mental or age-related disabilities (SiME)” funded by the German Federal Ministry of Education and Research (BMBF) (grant numbers 13N13946 and 13N13950) within the programme ”Research for Civil Security”. We thank the team and employees of the Lebenshilfe Bergisches Land and the anonymous participants for the committed and patient support during planning and conduction of the movement studies.

Paul Geoerg, Jette Schumann, Stefan Holl and Anja Hofmann. The Influence of Wheelchair Users on Movement in a Bottleneck and a Corridor. In: Journal of Advanced Transportation, pp. 1–25. ISSN: 0197-6729 (2019). doi:10.1155/2019/9717208.

Jette Schumann, Maik Boltes and Armin Seyfried. Hybrid Tracking System for Pedestrians in Dense Crowds. In: Traffic and Granular Flow 2017. Ed. by Hamdar S. H. (2019). doi:10.1007/978-3-030-11440-4_23.

Paul Geoerg, Jette Schumann, Stefan Holl, Maik Boltes and Anja Hofmann. The influence of individual impairments in crowd dynamics. In: Fire and Materials, vol. 2018, no. 2, p. 1, (2020). doi:10.1002/fam.2789.

Paul Geoerg. The Influence of Individual Characteristics on Crowd Dynamics. PhD thesis. (2021)

Jette Schumann. Utilizing Inertial Sensors as an Extension of a Camera Tracking System for Gathering Movement Data in Dense Crowds. PhD thesis. (2022)

Paul Geoerg, Jette Schumann, Maik Boltes and Max Kinateder. How people with disabilities influence crowd dynamics of pedestrian movement through bottlenecks. In: Scientific reports 12(1), 14273 (2022). doi:10.1038/s41598-022-18142-7.

A complete list of publications can be found in the full documentation of the data.