CIF:Small: Many-user Information Theory: A New Paradigm

Project: Research project

Project Details

Description

Classical multiuser information theory studies the fundamental limits of models with �xed (often
small) number of users as the coding blocklength goes to in�nity. This research proposes a new
paradigm, referred to as many-user information theory, where the number of users is allowed to
grow with the blocklength. This paradigm is motivated by emerging systems with a large number of
users, such as machine-to-machine communication. Speci�c examples of many-user models include
the many-access channel (MnAC), the many-broadcast channel, the many-relay channel, as well
as models with many correlated sources. Moreover, each transmitter or receiver in the many-user
system may be active with certain probability in a given block.

Since the conventional notion of capacity as the supremum of the data rate measured in bits per
channel use may not be the right metric for the system performance (e.g., per user rate may vanish
as the number of users grows). An important question in this investigation is how to generalize the
notion of capacity to many-user systems. Another question of great interest, is whether separate
detection of user on-o� activities and decoding of user messages achieve the capacity in the case
where users' on-o� activities are random.

Intellectual Merit: This project pushes the envelop of communications, information theory
and signal processing by studying models with a massive number of users, beyond existing theory for multiuser systems. The new theory will provide crucial guidance for the design of future machine-
to-machine systems, where the key concepts will also be evaluated by the experimental studies
proposed here.

Broader Impacts: Successful design of many-user systems will facilitate applications such
as machine-to-machine communication, which may a�ect many aspects of people's life, including
healthcare, transportation, smart grid, smart home, smart city, and public safety, to name a few.
Ideally, this research will lead to e�cient and e�ective architecture and supporting technologies for
implementing future many-user systems. The project also provides many opportunities for under-
graduate students, including those from underrepresented groups, to gain experience in research.
StatusFinished
Effective start/end date9/1/148/31/18

Funding

  • National Science Foundation (CCF-1423040)

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.