Modelling the Dynamics of Team Behaviour for Multiagent Human-Machine Systems
Abstract: Successful team performance requires individuals effectively coordinate their movements and actions with each other to achieve task success. This includes effectively deciding who, how and when to act, with robust decision-making often differentiating expert from novice performance. I will present recent research demonstrating how a combination and dynamical motor primitives and cutting-edge machine learning techniques can not only be employed to model and predict human perceptual-motor behaviour and decision-making during team action but can also help identify the information that best explicates expert task performance. Motivated by the increasing need to develop artificial systems capable of safe and robust human interaction, I will also detail how these models can be employed to control the movement and decision-making dynamics of interactive artificial agents and AI systems.