Humanoid Cognitive Reasoning

Lecturer (assistant)
  • Gordon Cheng [L]
  • Karinne Ramirez Amaro
Duration4 SWS
TermSommersemester 2019
Language of instructionEnglish
Position within curriculaSee TUMonline
DatesSee TUMonline


Admission information


Understanding difficult problems in the cognitive robotic research and apply possible solutions. Upon successful completion of this lecture, students are able to analyze and apply advanced techniques from Cognitive Intelligence in order to design a flexible imitation reasoning system for a humanoid robot. After the completion of the module, the students are able to evaluate the learned reasoning models which enable humanoid robotic systems to reason about complex situations. Additionally, after this lecture, the students are able to understand the difficult problems in the cognitive robotic research area and they are able to apply the possible solutions. As a results, the students are able to develop and produce models based on real problems using several higher-level reasoning methods.


Foundations of Cognitive Intelligence, Cognitive Semantics and Advanced Reasoning techniques applied to robots. Topics: 1. Cognitive systems 1.1 History and Foundations of Cognitive Intelligence 1.2 Introduction to Embodied Intelligence 1.3 Trajectory-level vs. Symbolic-level techniques 2. Levels of abstraction of the problem space 2.1 Low-level (trajectory-level) 2.2 High-level (semantic-level) 2.3 Inference in the high-level representations 3. Uncertain Reasoning for complex decision making 3.1 Uncertain knowledge and reasoning engines 3.2 Acting under uncertainties considering the context information 4. Cognitive Semantics and Advanced Reasoning techniques applied to Robots 4.1 Knowledge applied in Robot Reasoning 4.2 Semantic-based Learning approaches (OACs, Affordances, ...) 4.3 Hierarchical learning approaches 5. Inferring and understanding human intentions from demonstrations 5.1 Robot cognitive perception- What to observe? 5.2 Robot Execution – How to move? 5.3 Robot Decision Making – Why it is meaningful? 6. Transferring the reasoning models to robots 6.1 Design of a general imitation system 6.2 Brief Introduction to Cognitive Architectures


C/C++ programming skills, strong background in discrete mathematics, basic knowledge on AI (Prolog) and robotics is highly recommended

Teaching and learning methods

Lectures will be held ex cathedra. Laboratory work will consolidate the students' knowledge of the principles of the subject matter and deepen their understanding of robotic learning methods from a higher level perspective.