Intelligent Virtual Humans Behavior Daniel Thalmann, EPFL-VRlab In Behavioral Animation, virtual humans acquire the capabilities of perceiving their environment and are able to react and make decisions, depending on this input. The question that we are facing is: how to populate virtual environments with virtual humans so that they can behave autonomously and with some individuality. In particular, there is still a great challenge to select the right laws and rules to implement individual but believable behaviors.It takes a very flexible motion control and only computational models which are controlled by a few parameters can bring it. One of the major problem is to find such models and to compose them to create complex motion. In this talk, we will address case studies in perception-based game strategy in VR and AR, crowd behavior, and interaction with "smart" objects.