The lecture Robotic Intelligence: Navigation and Mobile Manipulation provides fundamental and advanced concepts for developing intelligent, autonomous robots for complex environments. The focus is on methods for navigation, legged locomotion, and mobile manipulation, as well as their integration into real robotic systems. Covered topics include sensor fusion, localization and mapping (SLAM), decision-making, and collision-free motion and path planning. Additionally, machine learning techniques for controlling legged robots in challenging terrain and for direct physical interaction with the environment are introduced. The course further addresses concepts of mobile manipulation, enabling robots to safely grasp, handle, or interact with objects and their surroundings. Both classical and modern, learning-based approaches are presented and discussed. By combining these methods, robotic systems can achieve a high level of autonomy. The lecture integrates theoretical foundations with application-oriented methods and enables students to develop and critically assess autonomous robotic systems and their methods for real-world scenarios.

An important component of the course is the practical group work conducted in the Machine Intelligence and Robotics Lab (MaiRo), where students solve tasks using real sensors and robots. They develop both technical competencies and teamwork skills to address complex navigation and manipulation challenges of mobile systems.

The topics include:

• Fundamentals of machine learning and sensor fusion
• Simultaneous Localization and Mapping (SLAM) in unknown environments
• Approaches for motion planning and collision avoidance
• Intelligent decision-making and task planning
• Reinforcement learning for the locomotion of legged robots
• Machine learning approaches for traversing and coping with complex obstacles
• Manipulation of objects in static scenarios
• Mobile manipulation in complex environments

By the end of the course, students can design, implement, and integrate methods for navigation and mobile manipulation. These approaches enable mobile systems to autonomously navigate through complex, dynamic environments, apply learned locomotion strategies, and perform various manipulation tasks.