Options
The information bottleneck method in communications
Citation Link: https://doi.org/10.15480/882.2952
Publikationstyp
Doctoral Thesis
Date Issued
2020
Sprache
English
Author(s)
Advisor
Referee
Title Granting Institution
Technische Universität Hamburg
Place of Title Granting Institution
Hamburg
Examination Date
2020-07-21
Institut
TORE-DOI
TORE-URI
Citation
Technische Universität Hamburg (2020)
The Information Bottleneck method is a generic information theoretical framework which aims for the compression of an observed random variable to a compressed random variable. The focal aim in designing this compression is to preserve relevant information. The method originates from machine learning and so far only has a few practical applications in communications. This thesis describes the application of the Information Bottleneck method to problems of receiver-sided signal processing in communications. The complexity of the receiver-sided baseband processing algorithms for demodulation and channel decoding causes a severe bottleneck in modern digital communication receivers. Their implementation complexity is mainly influenced by the bit width used to represent the signals processed in the hardware of the receiver and the arithmetical operations involved in the signal processing algorithms. Practical receiver implementations, therefore, have to be strongly quantized. Moreover, the involved signal processing algorithms have to be as simple as possible.
The Information Bottleneck method provides algorithms which aim to maximize the preserved relevant information for a given bit width, hence motivating to apply the Information Bottleneck method to receiver design. The problems covered in this thesis are the design of scalar channel output quantizers, the decoding of low-density parity-check codes and channel estimation and detection algorithms. It is shown that the Information Bottleneck design principle allows to build signal processing blocks for these problems which allow for very small bit widths, typically around four to five bits per sample. Anyway, performance close to that of signal processing algorithms with double precision can be achieved. A key to achieve this small bit width is the signal representation using only quantization indices instead of real or complex representation values. Moreover, all operations required in the obtained signal processing blocks degenerate to simple lookup operations. As a result, the aforementioned design goals for communication receivers are inherently achieved by the receiver design with the Information Bottleneck method.
The Information Bottleneck method provides algorithms which aim to maximize the preserved relevant information for a given bit width, hence motivating to apply the Information Bottleneck method to receiver design. The problems covered in this thesis are the design of scalar channel output quantizers, the decoding of low-density parity-check codes and channel estimation and detection algorithms. It is shown that the Information Bottleneck design principle allows to build signal processing blocks for these problems which allow for very small bit widths, typically around four to five bits per sample. Anyway, performance close to that of signal processing algorithms with double precision can be achieved. A key to achieve this small bit width is the signal representation using only quantization indices instead of real or complex representation values. Moreover, all operations required in the obtained signal processing blocks degenerate to simple lookup operations. As a result, the aforementioned design goals for communication receivers are inherently achieved by the receiver design with the Information Bottleneck method.
Subjects
Information Bottleneck method
low-density parity-check codes
quantization
mutual information
lookup tables
DDC Class
600: Technik
620: Ingenieurwissenschaften
Loading...
Name
Dissertation_Lewandowsky_A4_RGB.pdf
Size
5.15 MB
Format
Adobe PDF