Models of Human Behavior for Human-Robot Interaction and Automated Driving: How Accurate Do the Models of Human Behavior Need to Be?

There are many examples of cases where access to improved models human behavior and cognition has allowed creation robots which can better interact with humans, not least in road vehicle automation this is a rapidly growing area research. Human-robot interaction (HRI) therefore provides an important applied setting for modeling - but given the vast complexity behavior, how complete accurate do these need be? Here, we outline some possible ways thinking about problem, starting from suggestion that modelers keep right end goal sight: A successful human-robot interaction, terms safety, performance, satisfaction. Efforts toward model completeness accuracy should be focused on those aspects success most sensitive. We emphasise identifying difficult scientific objective its own right, distinct each HRI context. propose exemplify approach formulating priori hypotheses matter, involved interactions currently take place between such as automated driving. Our perspective also highlights risks overreliance machine-learned HRI, mitigate against risks.