As robots become more ubiquitous, they will increasingly interact with humans, especially those who are not robot experts. Applications in education, nursing, aged-care, and rehabilitation will need high-quality human-robot interaction. These interactions will color people's perceptions and profoundly affect the robot's ability to achieve its tasks. Unfortunately, current human-robot interactions are typically brittle and stilted. This is primarily because the robots do not follow the social and cultural norms of humans. To overcome these limitations, this project will provide robots with the ability to use and perceive such social norms, with emphasis on the conventions used in the Arabic world. The result will be robots that are able to interact with humans in a more robust and fluid manner, akin to human-human interaction.
The key to our approach derives from the theory of common ground, developed by Herbert Clark to explain human-human interaction. In his view, participants actively work together to ensure that they share a common mental model of each other's goals and capabilities. Importantly, these mental models include social and cultural norms that dictate how common ground should be achieved. Since these norms differ, somewhat, throughout the world, we will focus on those used in the Middle East.