Signal Processing Challenges for Communications beyond 5G EURASIP - - PowerPoint PPT Presentation

Signal Processing Challenges for Communications beyond 5G

EURASIP Special Area Team Signal Processing for Communications

EURASIP was founded in 1978 by Prof. Murat Kunt (EPFL, Lausanne) and registered as non-profit association in 2006 by a group of “Working Members”. EURASIP is governed by its Board of Directors (BoD). There are eight BoD members, seven of which are elected by the members for a period of four

• years. EURASIP provides a framework for European collaboration in Signal

Processing:

• EUSIPCOs, co-sponsored conferences & workshops
• EURASIP Journals
• EURASIP Open Library
• EURASIP PhD Thesis Repository
• EURASIP Awards
• EURASIP Local Liaison Programme
• EURASIP SATs
• EURASIP Seasonal PhD Schools

www.eurasip.org

EURASIP in a Nutshell

EURASIP SAT SPC

• Members (as of Jan. 2016)

Christoph Mecklenbräuker (Chair) Raymond Knopp (Vice-Chair) Mark A. Beach Hanna Bogucka Andreas Burg Giuseppe Caire Andrea Conti Merouane Debbah Thomas Eriksson Gerhard Fettweis Lajos Hanzo Geet Leus Marco Luise Marius Pesavento Ana Perez-Neira Bjorn Ottersten Markus Rupp Mikko Valkama

EURASIP Position Paper 3

Outline

• We collected topics for challenging research

directions over the next 3-5 years time- horizon, beyond “mainstream” 5G.

• Focus on open-ended, academic-oriented,

challenging, and fundamental issues

• For each of the topics, we point out:

1) Relevance 2) Is it “beyond 5G” ? 3) What are the novelty and main challenges?

EURASIP Position Paper 4

Topics

• A. Processing of side-information for improving

network performance

• B. Satellite networks as part of beyond 5G
• C. Quantum-assisted design of wireless systems
• D. Signal processing for network-assisted

cyberphysical systems

• E. Inferential networks
• F. Hybrid visible light communications and radio

Networks

• G. Networked signal processing and wireless

fronthaul/backhaul

• H. Open-source tools for radio network innovation

EURASIP Position Paper 5

• A. Processing of side-information for

improving network performance

1) Relevance: Side-information (about propagation environment traffic and mobility patterns is rarely exploited today, but promises large benefits. 2) Is it “beyond 5G” ? It seems that the rationale for the air interface design remains the classical per-frame pilot- based approach. 3) What are the novelty and main challenges? Side- information from sensors may help in harsh communication environments. Significant side-information is expected to be available in the internet-of-things, enabled by geographical information services and sensor data fusion. Connection between channel estimation, collaborative crowdsouring and machine learning.

EURASIP Position Paper 6

• B. Satellite networks as part of

beyond 5G

1) Relevance: Cost effective solutions for ubiquitous, broad

coverage and disaster resilient communications need to go beyond the 5G mainstream

2) Is it “beyond 5G” ? 5G has not yet considered a strong

integration with satellite networks

3) What are the novelty and main challenges? Seamless

Hybrid terrestrial-satellite backhauling, Integration in content delivery networks, Ubiquitous services for connected vehicles (aero, maritime etc), new network architectures and constellations (LEO, MEO, HAPS,...)

EURASIP Position Paper 7

• C. Quantum-assisted design of wireless

systems

1) Relevance: the powerful parallel processing capability of Quantum Search Algorithms (QSA) can be exploited for solving large-scale wireless optimization problems. 2) Is it “beyond 5G” ? This topic is highly exploratory and definitely goes beyond the boundaries of 5G mainstream R&D. 3) What are the novelty and main challenges? Classically the mutual information between the channel’s input and output has to be maximized. For quantum channels the capacity has to be redefined. Quantum channels are capable of increasing the achievable capacity. As a stunning result, redundancy-free error correction is possible over noisy

• channels. Quantum computers may enable exponentially

complex combinatorial ML decoding in classical systems.

EURASIP Position Paper 8

• D. Signal Processing for Network-

Assisted Cyberphysical Systems

1) Relevance: we envisage cyberphysical cloud-based systems where a large number of coordinating and moving objects (cars, drones) are jointly controlled via wireless networks. 2) Is it “beyond 5G” ? The latency and delay jitter introduced by a layered protocol stack, even in the presence of „new waveforms” at the physical layer, will not be able to meet the constraints of distributed control of large cyberphysical systems over wireless networks. 3) What are the novelty and main challenges? Joint-source channel coding & compressed sensing: leveraging the fact that linear maps are well conditioned with respect to noise (graceful degradation, no packet losses or retransmissions). Efficient design of plant state estimators/trackers from compressed sensor signals. Beyond

EURASIP Position Paper 9

• E. Inferential networks

1) Relevance: Beyond 5G networks can definitely extend the typical role of wireless networks devoted to secure and efficient

• communication. In particular, we foresee that they can serve as

inferential networks to infer various types of processes. 2) Is it “beyond 5G” ? Yes, connected to IoT applications at large (unfortunately IoT ``research‘‘ has been hijacked by CS, and focuses mainly on standardization of interfaces and interoperability). 3) What are the novelty and main challenges? Multi-dimensional random sampling generalizes the reconstruction of a stationary random process in one dimension. For irregular sampling, was described by a Levinsons theorem establishing the condition for perfect reconstruction. Open question remains: multidimensional random sampling and optimal interpolation in multiple dimension when nodes are randomly scattered.

EURASIP Position Paper 10

• F. Hybrid visible light communications

and radio networks

1) Relevance: Although bandwidth is plentiful in millimeter wave bands, communication is restricted to small areas and mainly line-of-sight links especially due to power-limitations in mobile terminals. 2) Is it “beyond 5G” ? Current proposals in most of the worlds for the 5G patchwork are ignoring VLC. 3) What are the novelty and main challenges? The use of common baseband signal processing units for VLC and radio

• transceivers. Highly-selective adaptive beamforming

techniques using feedback on the radio-link. The capacity of the free-space optical link (even in the single-user case) is still unknown, although some high-SNR bounds have been found.

EURASIP Position Paper 11

• G. Networked signal processing and

wireless fronthaul/backhaul

1) Relevance: The evolution to 5G+ networks includes a paradigm shift in signal processing: distributed and dynamic processing of radio signals in different parts of the network. Processing will be shared between remote radio-heads and centralized data centers for joint processing of the information from many RRH. 2) Is it “beyond 5G” ? Evolutionary with respect to the standrad C-RAN approach. 3) What are the novelty and main challenges Aggressive frequency reuse in fronthaul/backhaul, dynamic network planning/reconfiguration, low-complexity wideband transceivers for large antenna arrays (e.g., 1-bit ADC, HDA MU-MIMO precoding)

EURASIP Position Paper 12

• H. Open-source tools for radio

network innovation

1) Relevance: Evolution to 5G+, there is the need for open-source tools to ensure a common R&D and prototyping for rapid proof-

• f-concept designs. The majority of the development will be

collaborative and include university, research centers and a more general population of hackers. 2) Is it “beyond 5G” ? Yes, 5G is still driven by a small set of very large companies (fighting for IP). 3) What are the novelty and main challenges? Creation of a global

• pen-source ecosystem between academia and industry through

collaboration with standards-bodies (ETSI, 3GPP, IEEE). Developing a library of signal processing and upper layer protocol functions with controlled execution delay on general purpose machine providing dependable performance.

EURASIP Position Paper 13

Recommendation

We would like to attract attention on more open- ended, academic-oriented, challenging, and fundamental research topics, more properly adapted to a thriving scientific community that represents a major strength and a main asset for the sake of EU leadership in InformationTechnology, Markedly different from the current 5GPPP approach.

EURASIP Position Paper 14