Below, we have prepared a list of videos to demonstrate some of the features and philosophies of the Menge framework.
Architecture: Overview
This video details the core architecture of the framework and demonstrates how the various components work together in Menge.
Example: Comparing Crowd Models
This video demonstrates how Menge can be used to compare pedestrian models. We show the same experiment run with four different models.
Example: Authoring Flexibility
Menge's XML specification is easy to use and author. We demonstrate this by translating several examples from the popular SteerBench benchmarks.
Example: Aircraft Loading - Goal Selection
Menge's goal selection mechanism allows for minimal XML changes to create vast behavioral differences. We demonstrate this by changing the goal selection mechanism for loading a commercial aircraft.
Example: Aircraft Unloading - Custom Transitions
Because Menge divides computation into extensible elements, new functionality can be added to the framework with ease. This example demonstrates a custom transition condition requiring space around an agent to be free.
Example: Office Space - Simulating Complex Scenarios
Menge's Behavioral Finite State Machine (BFSM) allows users to represent arbitrarily complex behaviors. In this office, workers work at their desk, make copies, get snacks, visit the restroom, and exit for lunch according to a probability distribution defined in the XML.
Example: Dynamic Agent State
This example demonstrates a stress model implemented in Menge. As stress accumulates, the internal states of the agents change, altering how they behave and interact. This example also demonstrates Menge's interoperability. Trajectories from Menge are output to an external renderer.
Example: Stadium Exit
Menge is not constrained to operating in a 2D environment. Here we demonstrate results of a stadium exit simulation based on an experiment with human pedestrians.
Example: Scalability
This example demonstrates Menge's scalability. This scene demonstrates 32000 agents simulating at rates faster than real time.
Example: Formations
This example demonstrates Menge's support for formation behaviors. In the first demo, a group of agents switches between an X, a line, a smiley face, and no formation as they move. In the second example, a formation navigates around an obstacle. Notice how the formation reacts and reshapes as agents move to avoid the walls.
Example: DenseSense
These results are from DenseSense, a module for density-dependant behaviors based on Menge. DenseSense computes a density field using a task, and allows the agents to make decisions based on density.