Machine-induced coordination behavior in human-machine interaction

The dynamics of human-machine interactions can vary from simple to complex. It spans a range that go from basic rule-based serial stimulus-response exchanges to more realistic bi-directional interactions that evolve in real-time according to (possibly) complex laws. Increasingly, a subset of such systems, those between humans and humanoid robots has been drawing more scrutiny because of their potential as surrogate systems for live multi-agent social interaction. Such systems may provide insights into the mechanisms of human cooperation and competition as well as uncover novel behaviors with judicious extension of the parameter space in which to investigate them. We draw on a recently proposed human-machine interaction system, the virtual partner interaction (VPI) to focus on the fundamental issue of machine-induced coordination behavior on a human partner. This study uses coordination dynamics, an empirical/theoretical framework that has found numerous applications in studies of coordination at the levels of both brain and behavior. Specifically, we study VPI inducing a coordination behavior opposite to that which a human is required to execute during a coordination task. Its importance and scalability to more general situations is discussed.