Understanding group behavior is essential for improving collaboration and productivity. While research on group behavior in virtual reality (VR) is significantly advanced, understanding group dynamics in mixed reality (MR) remains understudied. Understanding MR group dynamics will enable designing systems that optimize collaboration, enhance productivity, and improve user experiences. This work outlines how MR headsets sensory systems can effectively capture group behavior, devise algorithms to process and interpret the data and demonstrate the correlation between group behavior and task-related performance metrics. We propose a framework for group behavior analysis in MR, or GroupBeaMR for short, to capture and analyze group behavior in MR. Using the rich sensory capabilities of MR headsets, GroupBeaMR passively collects data on conversation, shared attention, and proximity. This data is processed using social network analysis techniques to identify patterns of interaction and assess group behavior. Our evaluation, involving 44 participants in 11 groups, demonstrates the effectiveness of GroupBeaMR in capturing and analyzing group behavior in collaborative MR tasks. An example of insight from GroupBeaMR is that balanced participation in different types of interaction leads to higher group cohesion. These findings enable real-time assessments of group behavior in MR that can enhance collaborative experiences.

翻译：理解群体行为对于提升协作与生产力至关重要。尽管虚拟现实（VR）中的群体行为研究已取得显著进展，但对混合现实（MR）中群体动态的理解仍相对不足。理解MR中的群体动态将有助于设计能够优化协作、提高生产力并改善用户体验的系统。本研究阐述了如何利用MR头显的传感系统有效捕获群体行为，设计算法处理并解读数据，并论证群体行为与任务相关绩效指标之间的关联。我们提出了一个用于MR中群体行为分析的框架（简称GroupBeaMR），以捕获和分析MR中的群体行为。GroupBeaMR利用MR头显丰富的传感能力，被动采集对话、共享注意力和空间邻近性等数据。这些数据通过社会网络分析技术进行处理，以识别交互模式并评估群体行为。我们在包含11个小组、共计44名参与者的实验中评估了GroupBeaMR，结果证明了其在协作MR任务中捕获和分析群体行为的有效性。GroupBeaMR的一个分析示例表明：在不同类型的互动中保持均衡的参与度会带来更高的群体凝聚力。这些发现使得对MR中群体行为的实时评估成为可能，从而能够提升协作体验。