If you have made an application or movie based PedSim and want to have screenshots or video sequences of it featured here, let me know!
Example Movies
For the following short movie sequences, pedestrian trajectories generated by PedSim were used.
|
Pedsim T15 - 5000 vs 5000 from Christian Gloor on Vimeo. This is an example scene from
a pedestrian simulation generated 2004 by using PEDSIM. The output was
sent to povray to produce single pictures as well as a movie. Note
that this is a early version of pedsim. Also the focus is on the
simulation dynamics and not on the output quality. |
Pedsim/Alpsim T4 - 1 hiker: peak to hotel from Christian Gloor on Vimeo. This is an
example scene from a pedestrian simulation generated 2005 by using
PEDSIM/Alpsim. The output was sent to povray to produce single
pictures as well as a movie. Note that this is a early version of
PEDSIM/Alpsim. The goal was to validate decisions of the agents based
on visual stimuli. |
|
Epic Space Battle with 300 Ships from Christian Gloor on Vimeo. This is an attempt to
visualize the output of "SpaceSim", a 3D version of "PedSim". There
are 300 ships of two kinds fighting each other. The behavior of the
ships is defined by the agent logic, and all the movements are
calculated automatically. |
Schatzalp from Christian Gloor on Vimeo. This is an example scene from a pedestrian simulation generated 2005 by using PEDSIM/Alpsim. The output was sent to povray to produce single pictures as well as a movie. The goal was to validate decisions of the agents based on visual stimuli, in this case a planned building on a mountain in Davos, Switzerland. |
More Examples
| Older Movie Sequences, downloadable |
| Movie 1 (divx, 15M) | This is an example scene from a pedestrian simulation generated by using PEDSIM. The output was sent to povray to produce single pictures as well as a movie. Note that this is a early version of pedsim. Also the focus is on the simulation dynamics and not on the output quality. |
| Movie 2 (divx, 7.6M) | This is another example scene from a pedestrian simulation generated by using PEDSIM. This video is a bit smaller. It is a simpler scenario containig of about 200 warriors. |
| Movie 3 (divx, 15M) | This is a very early version of PEDSIM. However, you can clearly see the pedestrian dynamics here. The pedestrians acted more like particles, a fact that has been improved since then. This is a simulation of approx. 1200 pedestrians. |
| Movie 4 (divx, 17M) | "Even if it is breached, it will take a number beyond reckoning - thousands to storm the keep." "Tens of thousands." "But, my lord, there is no such force." Yes, there is. Actually, two of them. |
| Movie 5 (divx, 80M) | A movie taken from the perspective of the eyes of a hiker in the Swiss Alps. Using pedsim, this project tried to determine what kind of visual stimuli are important to hikers. |
| Still frames from the movie sequences above |
|
This is a single frame taken from the 1st movie. |
|
This is the same scenario as the 1st movie, but 'filmed' from a different location. 3 frames were merged together to create some kind of motion blur. |
|
This is a frame taken from the 3rd movie. On a pentium IV 2GHz, the practical upper limit is at about 10'000 pedestrians. In this picture, there are about 1200. |
|
This is a frame taken from the 4th movie. In this picture, there are about 10'000 pedestrians. |
| Application Screenshots |
|
Overview screenshot. Waypoint display is enabled. |
|
Zoomed in screenshot. The various forces that affect the trajectory of the pedestrian are marked: Blue is the force that accelerates the pedestrians away from the obstacles, green away from the other agents. The "Look Ahead" beahavior from the mental layer will add a magenta-colored force, which directs pedestrians in a way that there will be no head-to-head collisions. Red is the direction they would want to walk, and yellow the direction they actually do walk. |
|
This is a screenshot of the 2-dimensional visualizer. The simulation shown here is the one used for the example movie. You can see the two armies approaching each other. |
|
This is a screenshot of the 2-dimensional visualizer. It is possible to zoom in and display the agents' trajectories. This is useful for analyzing the behaviour of the agents. |
|
This is a screenshot of the 2-dimensional visualizer. The simulation shown here is a pedestrian simulation based on PEDSIM, but modified in order to make pedestrians follow hiking paths in the Alps. |
|
This is a screenshot of a 3-dimensional visualizer. The simulation shown here is a pedestrian simulation based on PEDSIM, but modified in order to make pedestrians follow hiking paths in the Alps. However, this visualizer is not part of the PEDSIM package. It listens to the xml data stream sent by the simulation and visualizes the progress in real time. (c) by Duncan Cavens, ETH Zürich |