Transcript Staffer Day Template - Massachusetts Institute of Technology

Information Theory for Mobile Ad-Hoc Networks (ITMANET): The
FLoWS Project
Thrust 2
Layerless Dynamic Networks
Lizhong Zheng, Todd Coleman
MANET Metrics
New Paradigms
for Upper
Bounds
Constraints
Capacity and Fundamental Limits
Capacity
Delay
Upper
Bound
Layerless
Dynamic
Networks
Lower
Bound
Degrees of
Freedom
Power
Application
Metrics and
Network
Performance
Models and
Dynamics
Application and Network
Optimization
Capacity
(C*,D*,E*)
Delay
Utility=U(C,D,E)
Power
Fundamental Limits
of Wireless Systems
Metrics
Models
New MANET Theory
Application Metrics
Layerless Dynamic Networks
• Dynamic : Separation of functionalities by different time scales no
longer optimal.
– Time varying channel/network environments, lack of information, high
overhead costs;
– Data/side information available in a variety of forms, with a wide range of
quality/precision/reliability;
– Broadcasting and interference, beyond point-to-point communications;
• Layerless: New signaling schemes, new metrics
– Network information theory: cooperative/cognitive transmissions, relay and
soft information processing;
– Dynamic information exchange, feedbacks and multi-way communication;
– Heterogeneous data processing, prioritizing data by different levels of
reliability; networking based on new interface to the physical layer;
– The principle of network coding, transmit-collect-combine pieces of
information, generalized form and coordination in dynamic networks;
– Operating with imperfect side information, robustness and universal
designs.
Thrust Areas
• Network information theory
– Multi-terminal communication schemes with cooperative and
cognitive signaling, interference mitigation, broadcast/relay;
• Generalizing network coding
• Structured code designs
– Efficient transmission of heterogeneous data and applications;
• Universal and robust algorithms
• Feedback and dynamic communication problems
Recent Thrust Achievements:
Relaying, forwarding, and combining soft information
• Likelihood forwarding – Koetter
• General relaying for multicast – Goldsmith
• Interference forwarding -- Goldsmith
• Broadcasting Channel with Cognitive Relay – Goldsmith
• Multicasting with relay – Goldsmith
– Structured codes based on the structure of the network;
– Finite field channel / linear codes vs. Gaussian channel / nested
lattice codes
• Sum capacity for deterministic interference channels -- El Gamal
• Broadcasting Channel with 3+ receivers – El Gamal (focus talk)
• Analytical study for zig-zag decoding -- Medard
Recent Thrust Achievements:
Structured codes and Applications
• Broadcasting with layered source codes – Goldsmith
• Generalized capacity/distortion for joint source channel
codes – Effros & Goldsmith
• UEP: performance limits and applications – Zheng
• Joint source channel coding with limited feedback
-- Goldsmith
– Successive refinement with progressive transmissions
– Dynamic of channel blocks between layers
• Distortion-outage tradeoff
-- Medard & Zheng
– Comparison between multiple description and multi-resolution
codes
– New interface to the physical layer
• Low complexity channel coding with Polar codes – Effros
Recent Thrust Achievements:
Generalized Network Coding
• Contents feedback with Network Coding – Effros
• Linear representation in Network Coding – Effos
• When linear coding suffices for all interior nodes?
• Low SNR Relay Networks – Medard
• Peaky signaling with network coding
• Distributed network coding with coded side
information – Effros
• Multiple uni-cast with Network Coding -- Medard
Recent Thrust Achievements:
Feedback, channel memory, and dynamics
• Generalization to finite state broadcast channels (FSBC)
- Goldsmith
• Control principles for feedback channels – Coleman
• Dynamic coding problem interpreted with stochastic control
• Reverse iterated function system (RIFS) decoding algorithms (w/
linear complexity) achieve capacity
• Lyapunov exponents of dynamical systems provide clean way to
demonstrate how posterior matching achieves capacity
• Tilted posterior matching for feedback channels
-- Zheng (focus talk)
• Joint source channel code with outage, with limited
feedback
-- Goldsmith
• Exploiting mobility in relay networks
-- Moulin
Thrust Synergies
Thrust 1
New Paradigm of outer bounds
Provide building blocks for large
networks, translate design constraints
into network modeling assumptions
General Relaying,
interference forwarding
-- Goldsmith
Posterior matching for
feedback channels –
Coleman, Zheng
Thrust 2
Dynamic Network Information theory:
improving performance in presence of
interference, cooperation, and
dynamic environment
Provide achievable performance
region, based on which distributed
algorithms and resource allocation
over large networks are designed
Guide problem formulation by
identifying application constraints
and relevant performance metrics
Network scalability, robust
and distributed algorithms
Performance benchmark and
design justification
Thrust 3
Application Metrics and
Network Performance
Joint source channel
coding with outage
– Goldsmith
Thrust 2 Achievements Overview
Dynamic Network Information Theory
Goldsmith: general relaying, soft combining
Goldsmith: Interference forwarding
Goldsmith: Multicast with
relay; BC with cognitive relay
El Gamal: BC with 3+ receivers
Moulin: exploiting mobility
of relay networks
Effros: distributed network coding
with coded side information
Cover: coordination capacity
Effros: linear representation of network coding
Coleman: Control principle for feedback channels
Goldsmith: Feedback
and Directed Information
Medard, Zheng: Diversity-distortion tradeoff
Goldsmith: Joint source channel coding / outage
Moulin: Universal Decoding in MANETs
Zheng: tilted matching for feedback channels
Zheng: Mismatched receivers
CSI, feedback, and robustness
Effros: two stage polar codes
Goldsmith: Broadcasting with layered code
Structured coding
Thrust Achievement Summary
1. Revolutionize upper bounding techniques (thrust 1)
2. Determine the optimal channel/network “coding” that achieves these capacity upper
bounds when possible, and characterize for which classes of networks gaps still
exist between achievability and upper bounds, and why.
– Coordination capacity – Cover (focus talk);
3. Develop new achievability results for key performance metrics based on networks
designed as a single probabilistic mapping with dynamics over multiple timescales
– General framework for systematic development and comparison of signaling schemes
– New metrics and new interfaces
• Error-erasure, distortion-outage, joint source channel codes;
– New signaling techniques
• Generalized network coding, soft information combining, interference management;
• Message embedding, layered codes;
• Robust and universal algorithms;
• Dynamic communication problems, feedback, mobility;
4. Develop a generalized theory of rate distortion and network utilization (thrust 3)
5. Demonstrate the consummated union between information theory, networks, and
control; and why all three are necessary ingredients in this union.
– Large network results;
– Control view of dynamic information exchange
Thrust Challenges
• Formulations beyond capacity
–
–
–
–
Philosophical alignment of achievability and converse methodologies
Bounds vs. approximations
New metrics focusing on delays and robustness
Hidden dimensions of limits
• Coordination and Networks
– Not all communication has to be reliable: how to measure/utilize soft
information for end-to-end reliability?
– Robustness and imperfect side information
– Coordination overhead
• Coded side info. / unintended helper / confidential messages
• Feedback and dynamic communication problems: are rate and error
exponent the right metrics?
• Impacts and insights to practical problems