Abstract
In the near future, robots are not seen as automated machines, but intelligent and powerful assistants for human beings. They are able to assist people in common places such as offices and houses. This paper presents a scenario of a housekeeping robot. The robot is assumed to receive instructions from a human user and perform tasks in a kitchen. A human user is expected to communicate with a robot by using a natural language. Thus, such kind of robots requires a natural and intelligent human-machine interface. This can be facilitated through a natural language interface. The major challenging issue in developing a natural language interface for the robot is ambiguity and uncertainty in user instructions. This paper presents a technique to resolve the problem. The technique is created by applying possibility theory, fuzzy sets, and knowledge about a user and robot to the user instructions.