Fig 1.
Snapshot of an experiment with bidirectional pedestrian streams.
Test persons with black shirts are moving from left to right (red shirts from right to left). Lanes unstable in time and space are formed. For a movie of this experiment we refer to supporting information S1 Movie.
Fig 2.
Left: Entry without guiding barriers (semicircle setup). Test subjects are positioned in a semicircle in front of two entrances 0.5 m in width. Right: Entry with guiding barriers (corridor setup). From top to bottom t = 0 and 6 s after the command to start entering. The densities in front of the barriers at t = 0 are comparable. With guiding barriers, the density in front of the entrances is significantly smaller at t = 6 s. For movies of the experiments we refer to supporting information S2 Movie (semicircle setup) and S3 Movie (corridor setup).
Fig 3.
Cumulative trajectories for the semicircle (left) and corridor (right) setups.
The red rectangles show the measurement areas for the time series of the density in front of the entrances. Trajectories in the corridor setup show four lanes in the straight part and merging at the entrances.
Fig 4.
Time series for density in front of the entrances.
After the start command, the constriction effect leads to an increase of density in front of the entrances. For the semicircle setup the density fluctuates around 8 m-2 after the constriction, while for the corridor setup it is around 5 m-2.
Fig 5.
Density profiles as mean values for time interval t = 20s to 80s, see Fig 4.
In the semicircle setup (top) a heterogeneous structure with hot spots up to 11 m-2 appears.
Fig 6.
Relation between time and distance to entrance for the semicircle setup (left) and corridor setup (right).
Light blue dots correspond to the initial position while black lines show the cumulative data for the relation during entering.
Table 1.
Results of a linear regression of the relation between time and distance to entrances.
While the slope as an indicator of the speed of the entering progress is different the correlation coefficient as an indicator of the justness of the process shows negligible variations.
Fig 7.
Semicircle (left) and corridor (right) setups.
Participants are instructed to imagine that they are standing in the designated areas (white ellipses).
Fig 8.
Perceived justness: mean, standard deviation (left). Corridor setup is perceived as significantly juster than the semicircle setup. After watching the video, the difference in perception is confirmed and increases. Level of comfort: mean, standard deviation (right). Participants feel more comfortable in the corridor setup than in the semicircle setup. Again, the difference is confirmed after watching the videos and increases.
Table 2.
People observed showing inappropriate behavior.
Table 3.
Forms of inappropriate behavior (frequency of occurrence, three most frequently mentioned only).
Table 4.
Percentage of participants who think that they can contribute to faster access.
Table 5.
Strategies for faster access (frequency of occurrence, three most frequently mentioned only).
Table 6.
Social norms (frequency of occurrence; three most frequently mentioned only).