The document discusses analyzing group dynamics and interactions during gaming simulations using video-based techniques. It describes an energy policy simulation game involving 3 small groups representing different energy departments. The game aims to create awareness and abstract reality using game elements to maximize learning. Video recordings of the game would allow quantitative and qualitative analysis of group behaviors like discussions, negotiations, and influences between individuals and groups. The document proposes using technologies like audio-visual sensors, signal processing, and machine learning to automatically track behaviors from multimodal data and provide summarized perceptions and feedback on interactions like speaking patterns, eye gaze, and gestures.