Returning scholarship

• Applicant: Dr.-Ing. Alexander Fengler
• Subject Area: Communication Technology and Networks, High-Frequency Technology and Photonic Systems, Signal Processing and Machine Learning for Information Technology
• Term: Funded in 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 543939471

Wireless communication of machine-type devices, i.e. sensors, cars, smart household devices, factory machines, etc, is one of the new central application scenarios of fifth generation mobile networks, known as Massive machine-type communication (mMTC). mMTC is characterized by short message lengths (in the order of 100 bits) and a very large number of devices (10^3 - 10^6 or more), of which only a comparably small number is concurrently active. The state-of-the-art approach for wireless spectrum sharing in human-centric communication is that an active device is first identified by the base-station (BS). The BS then grants each active device access to a dedicated set of resources (i.e. time or frequency blocks). For machine-type communication such a procedure becomes increasingly wasteful because the time needed for identification may exceed the actual transmission duration significantly. For efficient spectrum usage it becomes necessary to use grant-free protocols. A novel type of grant-free communication was recently developed under the name of unsourced random access (U-RA). In U-RA the identification of the devices is completely renounced and instead the BS recovers only a list of transmitted messages up to permutation. This formulation allowed, for the first time, to calculate information theoretic limits of uncoordinated multiple access communication on an additive-white-Gaussian-noise (AWGN) adder channel, even under the constraint of short message lengths and a very large number of devices. The purpose of this project is the development of new coding methods for U-RA, the mathematical analysis and optimization of existing methods, as well as their extension to more complex channel models. To increase the energy-efficiency of the mobile devices the use of multiuser multiple-input-multiple-output (MU-MIMO) is almost inevitable and a lot of effort in the last decades was dedicated to the research of MIMO technology. One important goal of this project is the characterization of information-theoretic limits of U-RA with a MIMO receiver. A special focus in this project is laid on the use of sparse-regression codes (SPARCs), a novel type of channel codes for the AWGN channel. The idea of SPARCs is to use concepts from compressed sensing to encode a sequence of bits as sparse signals, which are subsequently compressed and superposed. This coding construction showed to be very flexible and was proven to be capacity achieving on the AWGN channel. Prior work of the applicant and co-authors showed that the use of SPARCs also provided an energy-efficient solution for the U-RA setting. Another research objective of this project is to further develop, optimize and extend the use of SPARCs in U-RA with a MIMO receiver.

• DFG Programme: WBP Return Grant
• Host: Professor Giuseppe Caire, Ph.D.