Web-enabled Robots Robots that use the Web as an Information Resource

As robots are starting to perform everyday manipulation tasks, such as cleaning up, setting a table or preparing simple meals, their control programs must become much more knowledgeable than they are today. Typically, everyday manipulation tasks are specified vaguely and the robot must therefore infer by itself how to do the appropriate actions to the appropriate objects in the appropriate way in order to accomplish these tasks. These inferences can only be done if the robot has access to the necessary knowledge, including knowledge about how objects look, what their properties are, where they can be found, what might happen if particular actions are performed on them, etc. In this article, we describe and discuss the use of information that is available in the world-wide web and intended for human use as a knowledge resource for autonomous service robots. To this end, we introduce several categories of websites that can serve as information sources and explain which kinds of information they provide. We then investigate several information processing methods that can access these websites in order to provide robots with necessary knowledge for performing everyday manipulation tasks. The use of the web as a knowledge resource is a promising alternative to the hard and tedious task of coding comprehensive specific knowledge bases for robots.