While classical approaches to robotics aim for automation of (often repetitive and tedious) tasks, the goal of Cognitive Robotics is to develop systems that can flexibly deal with a broad range of tasks and variable con-texts. This presupposes quite different requirements than usually found in robotic systems. On the one hand, such cognitive systems should be adaptive in order to handle noisy real world conditions and are able to re-spond under changing conditions. On the other hand, these systems should be flexible in the sense that they can deal with distinct contexts and ideally support humans in all sorts of differing tasks. As an example, con-sider a household robot that should help to clean up the kitchen, bake a pancake, grab something from the fridge to just name a few possible tasks.
Importantly, the function of these cognitive systems differs from systems that are narrowly tailored to one task and can be optimized for this task. Structuring the interaction with the environment is crucial in a cognitive sys-tem and is approached through different interacting subsystems, like perception and action, but planning on a higher level as well (involving a specific form or a more general type of representation). Cognitive Robotics is here a current emerging discipline that draws on Robotics, Artificial Intelligence, and Cognitive Science. As it aims at equipping robots with autonomy and cognitive abilities, the goal is to bring us closer to interactive ro-bots that can support us in everyday tasks
The lecture will focus on such Cognitive Robotic Systems, underlying architectures and prominent components found in many of these approaches. It covers the key elements of traditional robotics, building on these to in-troduce the essentials of cognitive robotics. It emphasizes both theory and practice. In particular, the topics will cover (tentative overview of the lecture):
• Overview
• Interaction with the Environment: as an example locomotion;
From Embodiment to Embodied Representation
• Navigation and Mapping
• Planning
• Robot Cognitive Architectures
• Learning, Learning from Demonstration
• Interaction with Objects: Manipulation
• Interaction with Agents: Social Robots
• Towards Higher Level Cognition
• Benchmarking Robot Cognition
• Conceptual Issues
notwendige Vorkenntnisse:
Vorlesung Robobermanipulatoren
Vorlesung Kognitive Robotik
wünschenswerte Vorkenntnisse:
Modul "Neuronale Netze und Lernen"
Rhythmus | Tag | Uhrzeit | Format / Ort | Zeitraum |
---|
Modul | Veranstaltung | Leistungen | |
---|---|---|---|
39-M-Inf-KR Kognitive Robotik | Kognitive Robotik | unbenotete Prüfungsleistung
benotete Prüfungsleistung |
Studieninformation |
Die verbindlichen Modulbeschreibungen enthalten weitere Informationen, auch zu den "Leistungen" und ihren Anforderungen. Sind mehrere "Leistungsformen" möglich, entscheiden die jeweiligen Lehrenden darüber.
Jeder Teilnehmer bearbeitet einen aktuellen Forschungsartikel und trägt die Ergebnisse im Rahmen eines Vortrages vor. Im Praktikumsteil sollen die vorgestellten Verfahren selbständig implementiert werden.
Zu dieser Veranstaltung existiert ein Lernraum im E-Learning System. Lehrende können dort Materialien zu dieser Lehrveranstaltung bereitstellen: