Dr.-Ing. Tobias Fehenberger
I received my Dipl.-Ing. and Dr.-Ing. in Electrical Engineering and Information Technology from TUM in 2011 and 2017, respectively. I also have a B.Sc. in Management and Technology from TUM. I studied abroad at the American University of Beirut, Lebanon, and did my Master Thesis at the Optical Networks Group at University College London (UCL), UK. From 2012 to 2017, I was a Research and Teaching Assistant at the TUM Institute for Communications Engineering. Until August 2018, I was a Senior Researcher at TUM. Currently, I am on parental leave.
In my free time, I enjoy nature: hunting, hiking, and fishing.
I investigate optical communication systems, which have become the backbone of the digital age. Virtually all IP-based traffic of the Internet is transmitted over optical fibers, enabling high data rate services such as HD video streaming and cloud storage. The low loss of optical fibers over a huge spectrum of several THz and optical wideband amplifiers are key technologies that enable transmission of several trillion bits (Terabits) per second over thousands of kilometers of single-mode optical fiber with a core thinner than a human hair. Although these capacities seem huge, a steady demand for increased throughput has been observed for the past decades, with no end in sight. High-order modulation formats and advanced forward error correction schemes are potential options to achieve larger data rates. My research addresses these two entities jointly, a field known as coded modulation.Google Scholar profile
In this section, you find some Matlab™ source code that I found useful for my research. If you find an error or have a question, please let me know via email.
For a given input and a memoryless channel, the MI is the largest achievable rate. If you do not want to make restrictions on the receiver and its decoder, use MI. If you consider a receiver with binary decoding and iterations between demapper and decoder are not allowed, GMI is an achievable rate (MI is in general not!).
Calculate (O)SNR according to the GN model: calcOSNR_GNmodel.m (April 28, 2015)