Autonomous navigation through complex dynamic real-world scenarios is a demanding challenge for today’s mobile robots. Even for humans without map knowledge or GPS this constitutes an ambitious task in unknown environments. Present-day cognitive robotics research is still far away from enabling robots to accomplish even a basic task like navigating to designated goal locations only given symbolic identification labels of the locations. Humans have the compelling ability to consult other humans regarding required but missing directional information in order to fill such identified knowledge gaps, such as the way to a particular location “X”. Particularly, in unknown environments there will always be knowledge gaps for the robot as not everything can be pre-programmed and online learning on its own is not always a feasible or efficient solution. The ability to assess gaps in its own knowledge and to retrieve missing information from other agents like humans whenever possible is, thus, a highly desirable, yet, missing feature of today’s robots. Information exchange between humans and robots must not only be limited to robots serving information to humans, but also extended to humans supporting robots by providing specific missing information.
The goal of the Interactive Urban Robot (IURO) project is to develop and implement methods and technologies enabling robots to navigate and interact in densely populated, unknown human-centred environments and retrieve information from human partners in order to achieve a given navigation or interaction goal. With this goal and relevance for real-life scenarios in mind, IURO takes robots out of the well-structured and well-known environments that are generally found in state-of-the-art robotic experiments, perfectly matching the EU Workprogramme by contributing to improved perception and action capabilities of autonomous robots. Robots with capabilities investigated in IURO will act more robustly in changing and unforeseen situations, with higher autonomy and will be more dependable as they are able to proactively improve their knowledge deficiencies through interaction with humans.