Information Theoretic Limits of Multi-user Channels with State

Open Access
Li, Min
Graduate Program:
Electrical Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
June 04, 2012
Committee Members:
  • Aylin Yener, Dissertation Advisor
  • Aylin Yener, Committee Chair
  • William Kenneth Jenkins, Committee Member
  • David Jonathan Miller, Committee Member
  • Guohong Cao, Committee Member
  • Osvaldo Simeone, Special Member
  • state-dependent channels
  • strictly causal state
  • non-causal state
  • multiple-access channels
  • relay channels
  • GP binning
  • feedback
The communication channel with a random parameter called state was first introduced by Shannon and has received considerable attention in recent years because of its usefulness to model relevant phenomena occurring in wireless communication links, such as fading and interference. In many scenarios, nodes for communication, either transmitters or receivers, are informed of knowledge about the state or are capable of measuring the state. Hence, a fundamental problem is to understand methods of leveraging state information available at nodes to enhance communication reliability over channels. This dissertation makes progress along this line of study and concentrates on exploring new problems in multi-user communication setups, in order to provide fresh insights on different channel models with state from an information theoretic point of view. Two themes are of particular interest in the dissertation. The first theme focuses on the merits of strictly causal state information at encoders. Within this theme, two models have been considered. First, multiple access channels (MACs) with independent states, each known strictly causally to one encoder, are studied. An achievable strategy that generalizes and improves upon previous studies is identified. Capacity results are found for a class of channels. Moreover, the proposed scheme is extended to state-dependent MACs with an arbitrary number of users. Secondly, a state-dependent relay channel is studied in which strictly causal channel state information is available at the relay and no state information is available at the source and destination. The source and the relay are connected via two unidirectional out-of-band orthogonal links of finite capacity, and a state-dependent memoryless channel connects the source and the relay, on one side, and the destination, on the other. Two achievable schemes are proposed that exploit both message and state cooperation, possible due to the orthogonal links and the availability of state information at the relay. Capacity results are identified for some special cases. The second theme involves hop-by-hop communication channels with state. In particular, a state-dependent parallel-relay diamond channel is studied, where the source-to-relays cut is modeled with two noiseless, finite-capacity digital links with a degraded broadcasting structure, while the relays-to-destination cut is a general multiple access channel controlled by a random state. It is assumed that the source has non-causal channel state information and the relays have no state information. In this model, first, the capacity is characterized for the case where the destination has access to the state sequence. It is demonstrated that in this case, a joint message and state transmission scheme via binning is optimal. Next, the case with state information at the source only is considered. For this scenario, lower and upper bounds on the capacity are derived for the general discrete memoryless model. In general, this dissertation points to the advantages of strictly causal state information in multi-user communication channels. The results are in contrast to the single-user channel, in which strictly causal state information is not beneficial. Moreover, this dissertation provides methods of leveraging state information available at the source node in hop-by-hop communication channels.