TU Berlin

Department of Telecommunication SystemsDr.-Ing. Daniyal Amir Awan

Page Content

to Navigation

Dr.-Ing. Daniyal Amir Awan

I have doctorate in electrical engineering from Technical University of Berlin. I work as a research associate at TU Berlin and as a guest researcher at the Wireless Communication and Networks Department, Heinrich Hertz Institute, Berlin. My research revolves around application of optimization theory, function approximation, and machine-learning to problems in wireless communication systems. I am currently working in the following directions:

1. Set-membership & robust function approximation in dynamic wireless networks with a small sample set. 

2. Nonlinear detection for multi-user uplink using the set-membership paradigm.

3. Energy optimization in future wireless networks. 


Adaptive Learning for Symbol Detection
Citation key AwanBoo2020
Author D. A. Awan and R.L.G. Cavalcante and M. Yukawa and S. Stanczak
Title of Book Machine Learning for Future Wireless Communications
Pages 15
Year 2020
DOI 10.1002/9781119562306.ch11
Location New York, United States
Month December
Editor Wiley & IEEE Press
Publisher Wiley & IEEE Press
Chapter 11
Abstract This chapter introduces a novel machine learning algorithm for symbol detection in multiuser environments. It considers a challenging multiuser uplink scenario in which the number of antennas available at the base station may be smaller than the number of active users. More specifically, the proposed method is an adaptive (nonlinear) receive filter that learns to detect symbols from data directly, without performing any intermediate estimation tasks (e.g. channel estimation). Furthermore, the method is robust against abrupt changes of the wireless environment. The proposed algorithms for symbol detection are based on the theory of reproducing kernel Hilbert spaces, which have been extensively used in diverse fields such as statistics, probability, signal processing, and machine learning, among others. It also discusses the adaptive learning method for symbol detection in multiuser environments.
Download Bibtex entry


Quick Access

Schnellnavigation zur Seite über Nummerneingabe