Stochastic Foundations of Cyber-Physical Systems

Submitted by webmaster on Fri, 11/22/2019 - 15:57
Course No: 
182763
Course Type: 
VU
Term: 
2019W
Weekly Hours: 
4.0
Lecturer: 
Elahe Ghalebi
Language: 
English
Objective: 
  • Theoretical background for the design of cyber-physical systems (CPS). Models and algorithms to state estimation, control and planning are introduced.
  • Usage of tought paradigms, i.e., implementation of parts of a CPS including an introduction to robotics.
  • Supervised learning with Deep Neural Networks.
Content: 
  • Probabilistic interpretation of uncertainty.
  • Rational agents as smart cyber-physical systems (CPS).
  • Static (sBN) and dynamic (dBN) Bayesian networks (BN).
  • Uncertain environments as sBN and dBN.
  • Exact and approximate inference in BN.
  • Machine learning (supervised) of sBN and dBN.
  • Decision making and optimal control for Markov Decision Processes.
  • Supervised (sML) and reinforcement (rML) learning.
  • Machine learning (sML and rML) with deep neural networks.
  • Speech-recognition and robotics.

Didactic concept: Topics are tought in the lectures and practiced in exercises including programming exercises, simulation and application on real-world mobile robots.

Information: 

Lectures start s.t.
ECTS-Breakdown 3 ECTS = 150 hours:
Lecture part:

  • 0.5h  lecture introduction
  • 54h (18 lectures, 2h per lecture + 1h pre/postprocessing)
  • 20h exam preparation
  • 0.5h  oral exam----75h

Exercise part:

  • 75h exercises
Notes: 

S. Russel and P. Norvig, Artificial Intelligence - A Modern Approach, 3rd ed., Upper Saddle River, New Jersey: Pearson Education, 2010.
R.S. Sutton and A.G. Barto - Reinforcement Learning An Introduction second edition. The MIT Press Cambridge, Massachusetts London,  England, 2018.

Examination: 

<p>Homework/project assignments and oral examination.</p>

Recommendation: 

Mandatory prerequisites: None. The following prerequisites are helpful but not mandatory.
Fachliche und methodische Kompetenzen: Probability theory, stochastic signals, control theory, discrete mathematics.
Kognitive und praktische Kompetenzen: Mathematical reasoning and implementation skills. 
Soziale Kompetenzen und Selbstkompetenzen: Independent work, interest in combining theory and practice. 
These prerequisites are provided in the following modules: Wahrscheinlichkeitstheorie und Stochastische Prozesse, Signale und Systeme, Modellbildung und Regelungstechnik, Discrete Mathematics