Manuel S. Stein

alt text 

Doctoral Institution/Alma Mater:

Technische Universität München
Arcisstr. 21
80333 München
Germany

E-Mail: manuel.stein@tum.de

Biography

I received the Dipl.-Ing. and the Dr.-Ing. degree in electrical engineering and information technology from the Technische Universität München (TUM) in 2010 and 2016, respectively. From 2011 to 2016, I have been a Research Associate at the former Institute for Circuit Theory and Signal Processing (NWS), TUM, while serving as a Teaching Assistant at the Department of Electrical and Computer Engineering, TUM. During the time at the NWS, I have been working on my doctoral thesis about signal parameter estimation from coarsely quantized measurement data under the supervision of Josef A. Nossek. In 2016, I was a Research Associate with the Signal and System Theory Group, Universität Paderborn. From 2016 to 2018, I was a Research Fellow at the Chair for Stochastics, Universität Bayreuth and a Visiting Fellow at the Mathematics Department, Vrije Universiteit Brussel. In 2019 and 2020, I was a Research Fellow with the Circuits and Systems Group at the Department of Microelectronics, Technische Universiteit Delft.

Awards and Honors

For my scientific work, I was awarded the following individual research grants:

Research

The rapid progress in our abilities to store and process numerical data has led to profound technological and societal changes in recent years. Above all, the uninterrupted exponential increase in available storage size and computing power promises the possibility of soon using intelligent algorithms to automate the extraction of complex knowledge from large data sets. Nevertheless, the world around us, which we try to understand with the help of computers, remains a collection of physical quantities. Hitherto largely unnoticed, our capabilities to sense such analog measures and convert them into numerical data have stagnated. This endangers the unrestrained advance in sensing technology and applied data science. Consequently, not only the question “How to extract knowledge from data?” but also “How to get the data into the computer?” becomes a focus of the scientific and entrepreneurial minds of our time.

My research considers sensing, analog-to-digital conversion, and data processing as one joint problem. With this approach, I try to gain a better understanding of the fundamental question, “How to convert physical information efficiently into digital knowledge?”. In particular, I am interested in the probabilistic modeling and statistical processing of data streams in sensing systems with progressive analog-to-digital conversion concepts. The quantitative characterization of the trade-off between information and complexity at the edge between physics and computing systems fascinates me. In the statistical analysis of measurement data obtained by low-complexity analog-to-binary conversion, I am one of the leading experts. This data acquisition technique trades digital amplitude resolution against other sampling options, such as the temporal, spectral, or spatial measurement bandwidth, and makes it possible to design advanced sensing architectures that surpass our today's notion of technically achievable sensitivity, reliability, and latency.

The long-term vision of my scientific work is to contribute to the theoretical foundations in high-performance data processing for low-complexity sensor systems and to push practical implementation by proof of concepts. It is also essential for me to become engaged in the public discourse about the opportunities and challenges of advanced physical-numerical interfaces for individuals and society at large.

Topics

My research interests include:

  • Hardware-aware Signal and Information Processing

  • Low-Complexity Physical-Numerical Interfaces

  • Wireless Sensor Systems (GNSS, Radar, COM)

  • Performance Measures in Signal Processing

  • Low-Latency Statistical Data Processing

  • Large-Scale Array Signal Processing

  • Estimation and Detection Theory

  • Physical-Layer Synchronization

  • Probabilistic Machine Learning