Marcio Lobo Netto, Marcos Antonio Cavalhieri, Luciene Cristina Rinaldi Rodrigues
Cognitive Sciences Research Group, Polytechnic School of the University of São Paulo
05508-900, Av. Prof. Luciano Gualberto, trav. 3, 158 - São Paulo, Brazil

{lobonett,mac,luciene}@lsi.usp.br

The main goal of this tutorial is to introduce general artificial life concepts, and to present some deeper details on those aspects that may be of interest for the computer graphics community, particularly to those people involved with the design of virtual characters or with computer animation.
The tutorial intends to motivate new researches in this area. It provides an overview of this field, starting with analysis of life concepts, aiming to provide the support required to propose mathematical models that can reproduce life aspects appropriately in computer simulations. Then, it presents different approaches used to study different aspects of life, from very simple unicellular beings, to more complex multi-cellular ones, containing a large variety of functions and specific organs. Virtual creatures can be developed and used in a large diversity of scenarios, and for each of them different simulation approaches may apply. Therefore this tutorial uses some case studies to describe how these approaches can be effectively used on these different scenarios. For instance, we can be interested, as scientists, on the observation of the evolution of virtual creatures representing beings from some species, studying their evolution, adaptation capacity, and so for. Or we may be interested, as graphic animators, on the production of real looking and behaving virtual creatures that play their act based on some movie script. The tutorial presents diverse perspectives to analyze and to design virtual beings, describing their internal architectures and external social relationships. The final results emerge from the dynamics associated to these models, and are strongly dependent from adaptive concepts. We introduce some models to represent the subsystems contained in these simulations, showing how they can evolve and adapt, and the emergence of nice natural features, recognizable as similar to those found in live-systems.

Behavioral and cognitive animation can provide an efficient framework and tools to develop virtual creatures. Therefore we will present some techniques and concepts that can be used to implement these features,providing the virtual actors with the required capacity to decide how to play their role, or how to react in different circumstances. We are interested to show how these concepts and techniques can be applied in order to provide means to support auto adaptation, based on strategies as genetic evolution and learning. We also discuss how mental models can be built to describe the personal vision of the world, and how to use basic language structures to communicate knowledge and as basis for reasoning and decision taking.