The repository contains the source code of my master's thesis "End-to-end optimisation of MIMO systems using autoencoder based on the deep learning". It is a learning algorithm which generates a MIMO transmitter and receiver realisation simultenously. This includes digital modulation and mimo encoding. The MIMO channel is modelled as a MIMO-Rayleigh-Block-Fading-Channel. For further inquieries or questions just contact me.

First make sure CUDA Toolkit is installed to use GPU resources.

Most important Python Packages needed:

• numpy
• matplotlib
• tensorflow
• pickle
• cudNN

• Thank you Felix Wunsch and Dr. Holger Jaekel for guiding me through the thesis
• Source code is based on Dr.-Ing. Jakob Hoydis' code for the SISO model from the paper "An Introduction to Deep Learning for the Physical Layer" (Timothy O’Shea; Jakob Hoydis, 02 October 2017)