Effective teamwork is crucial in high-stakes environments like surgical operating rooms [] and emergency response scenarios, where rapid team coordination and decision-making directly impact outcomes. This project integrates and analyze multimodal data to uncover how team dynamics, leadership emergence, and diversity-driven interactions contribute to overall performance.

Our approach employs multimodal sensing technologies to capture a broad range of individual and team behaviors, including psychological (stress, workload, emotional contagion), communicational (voice-activity, turn-taking), and spatial (physical activity, coordination) aspects, from team members in time-sensitive tasks. A key aspect is the integration of diverse data streams, including physiological metrics like heart rate variability, audio recordings for speech analysis, and advanced tracking of individuals' positions within the environment through an Error-State Extended Kalman Filter (ES-EKF). This filter combines ultra-wideband (UWB) positioning with inertial measurement units to address common challenges such as electromagnetic occlusion and reduced accuracy in multi-participant setups. By fusing these data sources, the ES-EKF significantly improves both sampling rate and precision, while preserving the cost-effectiveness and portability of UWB systems [].

Funded by a 2024 Grassroots grant, this interdisciplinary effort leverages expertise from the Haptic Intelligence Department, focusing on advanced sensing and computational analysis, and the Organizational Leadership & Diversity Group, which explores leadership emergence and the effects of diversity on team collaboration. We investigate how variations in factors such as gender, age, and educational experience influence communication patterns and group coordination by engaging diverse participants in a portable escape-game scenario at our institute in Stuttgart. Beyond controlled experimental settings, we extend our research to real surgical environments, where we analyze team interactions in authentic high-stakes conditions. This comparison between simulated and real-world scenarios enables us to develop insights to enhance team performance across domains.

By identifying the key drivers of effective collaboration, this project aims to inform the design of enhanced training protocols, real-time feedback systems, and collaborative technologies for high-stakes environments.

This research project involves a collaboration with Mathieu Chollet (University of Glasgow).