Andreas Molisch

Research Summary

Dr. Molisch's current research interests concentrate on the following areas:

WIRELESS PROPAGATION: As new wireless technologies are introduced, it is vital to understand and model the properties of the wireless propagation channel that are relevant for those channels. The well-known conventional channel models (Rayleigh fading, Hata pathloss model) are therefore not sufficient for designing and analyzing modern broadband systems. Dr. Molisch is interested in measurement and fundamental modeling techniques for propagation channels for wireless broadband systems – from massive MIMO channels, to vehicle-to-vehicle propagation, ultrawideband channels, to mm-wave and THz channels. Hallmark of his work in the area is a holistic approach that pays equal attention to conducting measurement campaigns (including construction of required equipment), advanced signal processing for evaluating the measurement results, and innovative channel modeling.

LOCALIZATION BASED ON TIME-OF-ARRIVAL: Location-based services are of increasing importance for consumer services, emergency responders, and internet of things. Dr. Molisch is investigating localization in GPS-denied environments via time-of-arrival techniques, employing both deterministic localization methods and machine learning. Realistic effects such as multipath, blocked line of sight, etc. are all taken into account.

MACHINE LEARNING: this is one of the hottest topics in communications research. The work of Dr. Molisch aims to see ML not as a black box, but rather take the special properties of wireless channels and wireless systems into account to obtain ML structures and algorithms that not only are more accurate, but also require less training data. Particular applications are in channel prediction/extrapolation and localization.

WIRELESS VIDEO DISTRIBUTION: the emergence of wireless video services puts an extra strain on cellular networks. A highly promising method for increasing capacity is caching of popular video files either in dedicated helper nodes (femtocaching) that replace backhaul with storage, and/or caching in the mobile devices themselves, and exchanging them via device-to-device communication when requested. Aspects ranging from physical-layer device-to-device communications to application-layer video considerations are being investigated.

EDGE COMPUTING AND AUGMENTED INFORMATION SERVICES: these research activities deal generally with the question of how to jointly assign resources for communications and computing. In edge computing, the results of the computation are sent back to the source of the data, while in augmented information services, they are sent onward to a different receiver. We exploit techniques ranging from Lyapunov drift to machine learning to tackle these problems.

MULTIPLE ANTENNA TECHNIQUES: Multiple-antenna (MIMO) systems are the best and most efficient way to providing the high data rates and high spectral efficiencies that are required for next-generation cellular systems and wireless internet. Dr. Molisch's research concentrates on
(i) reducing costs through hybrid beamforming, a technique he invented in 2003, and which is now widely used especially for massive MIMO.
(ii) improving spectral efficiency and robustness through cooperation between base stations and cell-free MIMO.

NOVEL MODULATION/MULTIPLE ACCESS METHODS: While MIMO/OFDM/OFDMA has become the standard modulation/multiple access method for modern wireless systems, and they are optimum under certain conditions, there are situations in which alternatives are preferable. Dr. Molisch investigates novel approaches ranging from novel multi-carrier schemes to multiplexing via Orbital Angular Momenta.