Motivation: 
Robots are becoming more present in our daily lives. As household robots perform more complex tasks and their behavior becomes context-dependent, robot behavior may deviate from human expectations and confuse users. A particularly fundamental subset of these expectations is commonsense in nature: expectations grounded in everyday understanding of the physical and social world. For example:
Physical commonsense — if a robot picks up an object, people expect it to eventually put it down somewhere.
Social commonsense — if a robot is holding food and a table is nearby, people expect the food to be put on the table rather than placed on the floor.
Task-related commonsense — if a robot begins cleaning a table, people expect it to finish it rather than stop halfway.

Despite their central role, such expectations remain largely unformalized in human-robot interaction (HRI) research. They are typically handled through researcher intuition or scenario-specific description, with no systematic representation that generalizes across tasks. This thesis takes a step toward filling that gap.
 

Suggested Research Questions:
RQ1. What characterizes commonsense expectations of robot actions in household HRI?
RQ2. How can such expectations be represented in a structured, systematic form?

Intended Outcomes:
A working characterization of commonsense expectations of robot actions in household HRI.
A systematic representation of such expectations, applicable across household scenarios.
Instantiations of the representation in a small set of representative kitchen HRI scenarios.
An initial evaluation of the representation.