Quantization of Deep MIMO architectures

Betreuer: Edgar Beck
Art der Arbeit: Masterarbeit (MSc)
Ausgabe: 02/2020
Bearbeiter: Sven Haesloop
Status: in Arbeit
Kurzfassung:

Motivation:

Owing to the great success of Deep Learning based pattern recognition in this decade, Machine Learning has recently gained much attention from the communication system research community. It offers several benefits for handling algorithm and model deficits. For example in massive MIMO systems with lots antennas, computations become complex. Therefore, we have an algorithm deficit. Recently, several equalizer for massive MIMO based on simple neural network structures were proposed which can achieve near-ML performance. To further reduce complexity, there is the possibility to quantize the neural networks.

Goal:

The aim of this thesis is first to implement the proposed MIMO detectors and then to quantize them with minimum loss in performance.

Requirements:

In order to work on this thesis, a successful participation in the lectures Wireless Communications and Communication Technologies is required. Also, a solid understanding in linear algebra, optimization theory and stochastics is needed. Programming skills in Python and participation in the lecture Advanced Topics in Digital Communications are advantageous but not necessary.

Zuletzt aktualisiert am 26.02.2020 von E. Beck
AIT ieee tzi ith Fachbereich 1
© Arbeitsbereich Nachrichtentechnik - Universität BremenImpressum / Kontakt