In the near future there will be more older people and less supporting care personnel. Moreover, people want to live at home longer. How should we solve this problem? One way is to provide remote care through smart technology: sensors that monitor your health and environment, in-home automation and domotics, and enhanced communication facilities. In KSERA (Knowledgeable Service Robots for Aging) we go a step further. A (small humanoid) robot extends a smart home environment and helps older people to live longer in their own homes. Clearly, such a robotic system could help reduce health care costs in the future.

Placing a robot in the homes of older people introduces many unsolved issues, however. The robot should be able to move about in cluttered and unknown environments, it should be able to approach a person, it should communicate, take context into account, respond and interact, make decisions and so on. What do users need in the first place? What are the requirements? How about ethical issues?

Many of these issues are addressed in the KSERA project. In this talk I will focus on the human robot interaction. Not only must the robot be able to interpret a person, it must also provide understandable cues for a person. From cognitive psychology we know that humans interact by internally simulating another person’s behavior. As human beings are, in many ways, similar, the human brain is able to predict and infer the goals of another person. Robots on the other hand are typically unlike human beings as a result of which their behavior becomes unpredictable. They do not interact in an understandable ways causing them to appear clumsy and unintelligent. Here I report how recent insights in cognitive psychology can help build robots that take human-robot interaction to the next level and that, ultimately, can be used to help older persons at home.