Strategies for autonomous systems to resolve ambiguity in human-system interaction

Knowing when full autonomy will fail and collaboration with others is needed to successfully execute a task is a fundamental ability for humans to ensure efficiency, safety, and even survival. This ability is equally important for autonomous systems, such as robots or self-driving vehicles.

Such systems operate in our public or private spaces where they will often be faced with ill-defined or ambiguous human requests, in particular if the human interaction partner is naïve with respect to the system´s capabilities and inner workings-a situation that will be the norm once household robots or self-driving vehicles will be widely introduced and become part of our everyday lives.

The systems’ success will depend as much on users´ trust and willingness to cooperate as it will on the systems´ technological and engineering capabilities.

A crucial part of this will be the systems´ ability to cope with misunderstandings, ambiguity, and errors in perception of the environment around them, in reasoning about a given situation, and in interaction with human users, termed conflicts in this project.

Accordingly, the aim of this project is to develop mechanisms that allow for automatically identifying such conflicts and to devise strategies that allow for resolving them, and to evaluate these in human-subject studies.

Overall, this project will make significant contributions to a successful, user-friendly interaction with autonomous systems and a satisfying, positive user experience.