Behavioural Animation Demos at VRlab

Real-Time Crowd Motion Planning: Group Behavior

We introduce an improved short-term collision avoidance algorithm and simple efficient group behaviors for real-time crowd simulations. Our approach allows the use of several motion planning algorithms of different precision for regions of varied interest. Pedestrian motion continuity is ensured when switching between such algorithms.

Barbara Yersin, Jonathan Maïm


>top

POPULATING ANCIENT POMPEII WITH CROWDS OF VIRTUAL HUMANS

Pompeii was a Roman city, destroyed and completely buried during a catastrophic eruption of the volcano Mount Vesuvius.
In collaboration with the Computer Vision Laboratory of ETHZ, we have revived its glorious past by creating a 3D model of its previous appearance and populated it with crowds of virtual Romans.

VRlab EPFL: Barbara Yersin, Jonathan Maïm, Daniel Thalmann
CVlab ETHZ: Simon Haegler, Pascal Mueller, Luc Van Gool

>top

Real-Time Scalable Motion Planning for Crowds

We developped a hybrid architecture to handle path planning
of thousands of pedestrians in real time, while ensuring dyna-
mic collision avoidance. The scalability of our approach allows
to interactively create and distribute regions of varied interest:
zones af highest interest are governed by a long-term potential
field-based approach; while other zones exploit a graph of the
environment and short-term avoidance techniques.
Our method also ensures pedestrian motion continuity when
switching between motion planning algorithms.

Fiorenzo Morini, Barbara Yersin, Jonathan Maïm

>top

Unique Instances for Crowds

We present an algorithm to allow for each individual in the crowd to be unique. Our contribution is twofold: first, a simple and efficient method for attaching accessories to individuals so as to modify their look. Second, a new and generic technique based on segmentation maps to add detailed color variety and patterns to human meshes as well as accessories.
Both methods are scalable to suit all human levels of detail exploited in crowd simulations. Our approach is adapted for real-time simulations of thousands of unique characters.

Barbara Yersin, Jonathan Maïm

>top

Real-Time Navigating Crowds: Scalable Simulation and Rendering

Navigation Graphs allow efficient navigation planning and simulation for crowds. They can be automatically computed from the mesh of the environment where the crowd is to be set. In real-time, our solution is able to simulate and render tens of thousands of people. The rendering engine exploits the 'Levels of Detail' principle, with models ranging from skinned articulated bodies to billboards. The crowd setup is eased by introducing navigation flows: the designer defines some destinations of interest in the environment, and the number of individuals aiming at going there, in order to populate its favorite virtual world.

>top

Riot in the City - multi-agent simulation framework

We introduce a new approach to virtual reality simulations with virtual characters. Using a high-level simulation framework, the user is able to collaborate with the agents on solving high-level problems, such as simulated crowd control, social interaction in a restaurant or training scenarios. The facilitator-based framework enables the user to work in multiple modes, such as first-person (direct) interaction or indirect, order-based interaction. Included action planner transparently supports teamwork and high level goals, letting the user to focus on the task being solved and not at the animation details.

>top

EPOCH

European research network on Excellence in Processing Open Cultural Heritage, EPOCH showcase 2.4.2 shows a reconstruction of the abbey of Ename and its monks around 1050 A.D.Our real-time crowd Engine allows to set up a simulation and fine-tune only global parameters. The resulting animations are captured, integrated into the abbey (provided by Ename-Center) and rendered with 3DSmax. Compared to the traditional approaches, our animation pipeline provides a quick and efficient way to control a large amount of virtual humans following a complex scenario.

link to the project EPOCH

>top

CAhrISMA

This realtime simulation of a virtual crowd inside a mosque has been developped for CAHRISMA project (Conservation of the Acoustical Heritage by the Revival and Identification of the Sinan's Mosques' Acoustics). While the virtual humans are praying in realtime, the user can visit the virtual reconstruction of Sokullu mosque and enjoy the prayer’s sound (auralised according to the mosque’s acoustics).

link to the project CAHRISMA

>top

a motivational model of action selection

This model of action selection based on motivations is sufficiently robust and flexible for modeling adaptive autonomous virtual humans.
In this example we simulate a realistic behavior for one day life of a virtual humans.
At each moment in time, the virtual human chooses the most appropriate action with respect to his motivations and his environment perceptions

>top

real time crowd simulation

Our crowd system aims to reproduce realistic scenarios involving large number of virtual human agents. It allows both scripted and autonomous behaviors of the agents as well as interactions with them. The agents react to other agents, to a virtual environment, and to real humans interacting with the simulation according to a set of behavioral rules.
Possible application of our system could be any virtual environment system requiring real-time execution of complex autonomous behaviors of many agents, such as various training systems, computer games or educational applications.

>top

CROWD BRUSH

We introduce a novel approach to create complex scenes involving thousands of animated individuals in a simple and intuitive way. By employing a brush metaphor, analogous to the tools used in image manipulation programs, we can distribute, modify and control crowd members in real-time with immediate visual feedback. We use an efficient technique allowing to render up to several thousands of fully three-dimensional polygonal characters with key-framed animations at interactive framerates.

>top

ERATO

In this cultural heritage project we provide the restitution of a Roman odeon (Aphrodisias) and its audience. The Odeon model is created with commercial software according to maps and archeological documentation. To create the audience crowd, we are using an intuitive authoring tool: characters are placed, modified and animated using a brush metaphor. A real-time crowd rendering engine is developed that handles 1000 humans or more.

link to the project ERATO

>top