Gaussian Relay Networks Research Articles

Asymmetric Successive Compute-and-Forward and the Capacity Gap for the Gaussian Two-way Relay Channel

Background: The compute-and-forward strategy is one of the outstanding methods which is used for interference management in wireless relay networks where decoding linear combinations of code words is required. Recently, many efforts have been made for decoding integer and noninteger combinations. The difference between the methods is the manner of handling different conditions of networks, such as equal or unequal power constraints and equal or unequal channel gains. Objective: In this work, we present a modified n-step asymmetric successive compute-and-forward strategy for the communication network where we have both unequal power constraints and unequal channel gains conditions. Methods: In the proposed method, we scale channel gains and coefficients with the square root of power constraints. In this way, despite previous methods, without the need for scaling factors in our formulation, it is still able to solve the problem of general Gaussian relay networks with unequal power constraints and unequal channel gains. We also use scaling factors in our method in order to have the ability to divide the rates between users fairly. Results and Conclusion: We evaluate the ability of the modified strategy for the uplink communication of the two-way relay channel, where one relay can help communication between the two users. At the relay, we decode the linear combinations of the messages of the two users and obtain 1/2 bit/sec/Hz per user capacity gap from the cut-set bound. Through some theoretical and simulation results, we show that by appropriately adjusting parameters, different points and areas of rate regions are achievable.

Cut-Set Bound is Loose for Gaussian Relay Networks

The cut-set bound developed by Cover and El Gamal in 1979 has since remained the best known upper bound on the capacity of the Gaussian relay channel. We develop a new upper bound on the capacity of the Gaussian primitive relay channel which is tighter than the cut-set bound. Our proof is based on typicality arguments and concentration of Gaussian measure. Combined with a simple tensorization argument proposed by Courtade and Ozgur in 2015, our result also implies that the current capacity approximations for Gaussian relay networks, which have linear gap to the cut-set bound in the number of nodes, are order-optimal and leads to a lower bound on the pre-constant.

Partial Decode-Forward Relaying for the Gaussian Two-Hop Relay Network

The multicast capacity of the Gaussian two-hop relay network with one source, $N$ relays, and $L$ destinations is studied. It is shown that a careful modification of the partial decode-forward coding scheme, whereby the relays recover and coherently transmit degraded sets of message parts, achieves the cutset upper bound within $(1/2)\log N$ bits regardless of the channel gains and power constraints. This scheme improves upon a previous scheme by Chern and Ozgur, which is also based on partial decode-forward yet has an unbounded gap from the cutset bound for $L \ge 2$ destinations. When restricted to noncoherent transmission among the relays, the proposed partial decode-forward scheme achieves a slightly larger gap of $\log N$ bits from the cutset bound. The computation of this relaxed achievable rate involves evaluating mutual information across $L(N+1)$ cuts out of the total $L 2^{N}$ possible cuts, providing a very simple linear-complexity algorithm to approximate the single-source multicast capacity of the Gaussian two-hop relay network.

Capacity Approximations for Gaussian Relay Networks

Consider a Gaussian relay network where a source node communicates to a destination node with the help of several layers of relays. Recent work has shown that compress-and-forward-based strategies can achieve the capacity of this network within an additive gap. Here, the relays quantize their received signals at the noise level and map them to random Gaussian codebooks. The resultant gap to capacity is independent of the SNRs of the channels in the network and the topology, but is linear in the total number of nodes. In this paper, we provide an improved lower bound on the rate achieved by the compress-and-forward-based strategies (noisy network coding in particular) in arbitrary Gaussian relay networks, whose gap to capacity depends on the network not only through the total number of nodes but also through the degrees of freedom of the min cut of the network. We illustrate that for many networks, this refined lower bound can lead to a better approximation of the capacity. In particular, we demonstrate that it leads to a logarithmic rather than linear capacity gap in the total number of nodes for certain classes of layered networks. The improvement comes from quantizing the received signals of the relays at a resolution decreasing with the total number of nodes in the network. This suggests that the rule-of-thumb in the literature of quantizing the received signals at the noise level can be highly suboptimal.

Approaching the capacity of two‐pair bidirectional Gaussian relay networks

The authors present an improved achievable rate region for two-pair bidirectional Gaussian relay networks based on successive compute-and-forward method. In these networks, one relay helps in the communication between two pairs of users. In their proposed scheme, the authors use nested lattice codes for encoding the messages at the users, and Gaussian random codes for the encoding at the relay. They use the successive compute-and-forward strategy to decode two integer linear combinations of the lattice codewords in the uplink, and successive interference cancellation for decoding the Gaussian codewords in the downlink. The downlink channel can be considered as a broadcast channel with two receiver groups, but within each group, a pair of users is considered as an additive white Gaussian noise channel (instead of a broadcast channel) because each node knows its own transmitted message. It is shown that for all channel gains of downlink channels and all channel gains of symmetric uplink channel pairs, the strategy achieves rates to within constant gaps of 1/2 and 3/4 bit/s/Hz per user of the cut-set upper bound for restricted and non-restricted models, respectively. These gaps are tighter than those previously obtained for this network, which have not exploited the successive compute-and-forward method.

Structured Lattice Codes for Some Two-User Gaussian Networks With Cognition, Coordination, and Two Hops

We study a number of two-user interference networks with multiple-antenna transmitters/receivers (MIMO), transmitter side information in the form of linear combinations (over an appropriate finite-field) of the information messages, and two-hop relaying. We start with a cognitive interference channel (CIC) where one of the transmitters (noncognitive) has knowledge of a rank-1 linear combination of the two information messages, while the other transmitter (cognitive) has access to a rank-2 linear combination of the same messages. This is referred to as the network-coded CIC, since such linear combination may be the consequence of some random linear network coding scheme implemented in the backbone wired network. For such channel we develop an achievable region based on a few novel concepts: precoded compute-and-forward (PCoF) with channel integer alignment (CIA), combined with standard dirty-paper coding. We also develop a capacity region outer bound and find the symmetric generalized degrees of freedom (GDoF) region of the network-coded CIC. Through the GDoF characterization, we show that knowing mixed data (linear combinations of the information messages) provides a unbounded spectral efficiency gain over the classical CIC counterpart, if the ratio (in decibel) of signal-to-noise (SNR) to interference-to-noise is larger than certain threshold. Then, we consider a Gaussian relay network having the two-user MIMO IC as the main building block. We use PCoF with CIA to convert the MIMO IC into a deterministic finite-field IC. Then, we use a linear precoding scheme over the finite-field to eliminate interference in the finite-field domain. Using this unified approach, we derive the symmetric sum rate of the two-user MIMO IC with coordination, cognition, and two-hops. We also provide finite-SNR results (not just DoF) which show that the proposed coding schemes are competitive against state-of-the-art interference avoidance scheme based on orthogonal access, for standard randomly generated Rayleigh fading channels.

Approximate Secrecy Capacity of Gaussian Parallel Relay Networks

We consider a Gaussian relay network, which has to protect the message from a powerful eavesdropper, which has the possibility of listening to the transmitted signals on orthogonal channels. The secrecy capacity of the Gaussian network is approximated by considering the corresponding network in the deterministic discrete superposition model. Upper and lower bounds on the secrecy capacity are found in terms of the deterministic model, which are potentially easier to evaluate. We perform this evaluation for the particular topology of the parallel relay network, where the relays form a parallel layer between the source and the destination. By analyzing the deterministic model, we approximate the secrecy capacity within a constant gap. The approximation can be expressed just in terms of the channel gains of the network. It is therefore possible to deduct that the simple scheme of hiding the transmitted signals in noise is approximately optimal.

Achieving the Capacity of the <inline-formula> <tex-math notation="LaTeX">$N$ </tex-math></inline-formula>-Relay Gaussian Diamond Network Within log <inline-formula> <tex-math notation="LaTeX">$N$ </tex-math></inline-formula> Bits

We consider the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$N$ </tex-math></inline-formula> -relay Gaussian diamond network where a source node communicates to a destination node via <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$N$ </tex-math></inline-formula> parallel relays through a cascade of a Gaussian broadcast (BC) and a multiple access (MAC) channel. Introduced in 2000 by Schein and Gallager, the capacity of this relay network is unknown in general. The best currently available capacity approximation, independent of the coefficients and the SNRs of the constituent channels, is within an additive gap of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$1.3 N$ </tex-math></inline-formula> bits, which follows from the recent capacity approximations for general Gaussian relay networks with arbitrary topology. In this paper, we approximate the capacity of this network within 2 log <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$N$ </tex-math></inline-formula> bits. We show that two strategies can be used to achieve the information-theoretic cut-set upper bound on the capacity of the network up to an additive gap of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$O(\log N)$ </tex-math></inline-formula> bits, independent of the channel configurations and the SNRs. The first of these strategies is simple partial decode-and-forward. Here, the source node uses a superposition codebook to broadcast independent messages to the relays at appropriately chosen rates; each relay decodes its intended message and then forwards it to the destination over the MAC channel. A similar performance can be also achieved with compress-and-forward type strategies (such as quantize-map-and-forward and noisy network coding) that provide the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$1.3 N$ </tex-math></inline-formula> -bit approximation for general Gaussian networks, but only if the relays quantize their observed signals at a resolution inversely proportional to the number of relay nodes <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$N$ </tex-math></inline-formula> . This suggests that the rule-of-thumb to quantize the received signals at the noise level in the current literature can be highly suboptimal.

Computing Half-Duplex Schedules in Gaussian Relay Networks via Min-Cut Approximations

Computing optimal half-duplex schedules in Gaussian relay networks is a challenging problem due to the lack of an exact capacity characterization and the large number of transmit-receive configurations that must be considered. We approach the problem using a constant-gap capacity approximation based on the cut-set bound with independent encoding at the nodes. We formulate an optimization problem to obtain the cut-set optimal half-duplex schedule and find that it is hard to solve in general. This is because it involves an exponential number of variables, since the number of ways to assign each node to either transmitter or receiver mode is exponential in the number of nodes. We present a general technique that takes advantage of specific structures in the topology of a given network and allows us to reduce the complexity of this problem. In certain classes of network topologies, our approach yields polynomial time algorithms for finding half-duplex schedules that achieve capacity within a constant gap. We use simulations to show running time improvements over alternative methods and compare the performance of various half-duplex scheduling approaches in different SNR regimes.

On the Gaussian Half-Duplex Relay Channel

This paper considers Gaussian relay networks where a source transmits a message to a sink terminal with the help of one or more relay nodes. The relays work in half-duplex mode, in the sense that they can not transmit and receive at the same time. For the case of one relay, the generalized Degrees-of-Freedom is characterized first and then it is shown that capacity can be achieved to within a constant gap regardless of the actual value of the channel parameters. Different achievable schemes are presented with either deterministic or random switch for the relay node. It is shown that random switch in general achieves higher rates than deterministic switch. For the case of K relays, it is shown that the generalized Degrees-of-Freedom can be obtained by solving a linear program and that capacity can be achieved to within a constant gap of K/2log(4K). This gap may be further decreased by considering more structured networks such as, for example, the diamond network.

Gaussian Half-Duplex Relay Networks: Improved Constant Gap and Connections With the Assignment Problem

This paper considers a Gaussian relay network where a source transmits a message to a destination with the help of N half-duplex relays. The information theoretic cut-set upper bound to the capacity is shown to be achieved to within 1.96(N+2) bits by noisy network coding, thereby reducing the previously known gap. This gap is obtained as a special case of a more general constant gap result for Gaussian half-duplex multicast networks. It is then shown that the generalized degrees-of-freedom of this network is the solution of a linear program, where the coefficients of the linear inequality constraints are proved to be the solution of several linear programs referred as the assignment problem in graph theory, for which efficient numerical algorithms exist. The optimal schedule, that is, the optimal value of the 2 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</sup> possible transmit-receive configuration states for the relays, is investigated and known results for diamond networks are extended to general relay networks. It is shown, for the case of N=2 relays, that only N+1=3 out of the 2 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</sup> =4 possible states have a strictly positive probability and suffice to characterize the capacity to within a constant gap. Extensive experimental results show that, for a general N -relay network with N≤8 , the optimal schedule has at most N+1 states with a strictly positive probability. As an extension of a conjecture presented for diamond networks, it is conjectured that this result holds for any half-duplex relay network and any number of relays. Finally, a network with N=2 relays is studied in detail to illustrate the channel conditions under which selecting the best relay is not optimal, and to highlight the nature of the rate gain due to multiple relays.

Efficient Capacity Computation and Power Optimization for Relay Networks

The capacity or approximations to capacity of various single-source single-destination relay network models has been characterized in terms of the cut-set upper bound. In principle, a direct computation of this bound requires evaluating the cut capacity over exponentially many cuts. We show that the minimum cut capacity of a relay network under some special assumptions can be cast as a minimization of a submodular function, and as a result, can be computed efficiently. We use this result to show that the capacity, or an approximation to the capacity within a constant gap for the Gaussian, wireless erasure, and Avestimehr-Diggavi-Tse deterministic relay network models can be computed in polynomial time. We present some empirical results showing that computing constant-gap approximations to the capacity of Gaussian relay networks with around 300 nodes can be done in order of minutes. For Gaussian networks, cut-set capacities are also functions of the powers assigned to the nodes. We consider a family of power optimization problems and show that they can be solved in a polynomial time. In particular, we show that the minimization of the sum of powers assigned to the nodes subject to a minimum rate constraint (measured in terms of cut-set bounds) can be computed in the polynomial time. We propose a heuristic algorithm to solve this problem and measure its performance through simulations on random Gaussian networks. We observe that in the optimal allocations, most of the power is assigned to a small subset of relays, which suggests that network simplification may be possible without excessive performance degradation.

Functional Forwarding of Channel State Information

Based on the recent compute-and-forward technique <citerefgrp xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <citeref refid="ref1"/></citerefgrp> , a novel communication strategy is proposed under which functions of the channel state information are forwarded along the network. Those functions are chosen such that on the one hand, they can be efficiently forwarded, and on the other hand, they are maximally useful to the final decoder of the message. It is illustrated that there is generally a tension between these two requirements. The strategy is shown to perform well for certain classes of multilayer networks where channel state information is acquired locally at each receiver. For example, for a two-stage Gaussian relay network with local channel state information, it is shown that the proposed strategy performs optimally in a scaling-law sense, as the number of relays increases.

Rematch-and-Forward: Joint Source–Channel Coding for Parallel Relaying With Spectral Mismatch

The Gaussian parallel relay network, introduced by Schein and Gallager, consists of a concatenation of a Gaussian additive broadcast channel from a single encoder to a layer of relays followed by a Gaussian multiple-access channel from the relays to the final destination (decoder), where all noises are independent. This setup exhibits an inherent conflict between digital and analog relaying; while analog relaying [known as amplify-and-forward (AF however, it is limited to white channels of equal bandwidth. In this paper, we present a generalization of the analog approach to the case of bandwidth mismatch. Our strategy, coined rematch and forward (R&F), is based upon applying joint source-channel coding techniques that belong to a certain class of maximally analog schemes. Using such techniques, R&F converts the bandwidth of the broadcast section to that of the multiple-access section, creating an equivalent matched-bandwidth network over which A&F is applied. It is shown that this strategy exploits the full bandwidth of the individual channels, without sacrificing the coherence gain offered by A&F. Specifically, for given individual-link capacities, R&F remains within a constant gap from the network capacity for any number of relays and any bandwidth ratio between the sections. Finally, the approach is extended to the case of colored channels.

Approximately Achieving Gaussian Relay Network Capacity With Lattice-Based QMF Codes

Recently, a new relaying strategy, quantize-map-and-forward (QMF) scheme, has been demonstrated to approximately achieve (within an additive constant number of bits) the Gaussian relay network capacity, universally, i.e., for arbitrary topologies, channel gains, and SNRs. This was established using Gaussian codebooks for transmission and random mappings at the relays. In this paper, we develop structured lattice codes that implement the QMF strategy. The main result of this paper is that such structured lattice codes can approximately achieve the Gaussian relay network capacity universally, again within an additive constant. In addition, we establish a similar result for half-duplex networks, where we demonstrate that one can approximately achieve the capacity using fixed transmit-receive (TX-RX) schedules for the relays with no transmit power optimization across the different TX-RX states of the network.

An Algebraic Approach to Physical-Layer Network Coding

The problem of designing new physical-layer network coding (PNC) schemes via lattice partitions is considered. Building on a recent work by Nazer and Gastpar, who demonstrated its asymptotic gain using information-theoretic tools, we take an algebraic approach to show its potential in non-asymptotic settings. We first relate Nazer-Gastpar's approach to the fundamental theorem of finitely generated modules over a principle ideal domain. Based on this connection, we generalize their code construction and simplify their encoding and decoding methods. This not only provides a transparent understanding of their approach, but more importantly, it opens up the opportunity to design efficient and practical PNC schemes. Finally, we apply our framework for PNC to a Gaussian relay network and demonstrate its advantage over conventional PNC schemes.

Near-Optimal Quantization and Linear Network Coding for Relay Networks

We introduce a discrete network corresponding to any Gaussian wireless network that is obtained by simply quantizing the received signals and restricting the transmitted signals to a finite precision. Since signals in the discrete network are obtained from those of a Gaussian network, the Gaussian network can be operated on the quantization-based digital interface defined by the discrete network. We prove that this digital interface is near optimal for Gaussian relay networks and the capacities of the Gaussian and the discrete networks are within a bounded gap of <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">O</i> ( <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">M</i> <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ) bits, where <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">M</i> is the number of nodes. We also prove that any near-optimal coding strategy for the discrete network can be naturally transformed into a near-optimal coding strategy for the Gaussian network merely by quantization. We exploit this property by designing a linear coding strategy for the case of layered discrete relay networks. The linear coding strategy is near optimal and achieves all rates within <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">O</i> ( <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">M</i> <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ) bits of the capacity, independent of channel gains or signal-to-noise ratio. The linear code is therefore a near-optimal strategy for layered Gaussian relay networks and can be used as on the Gaussian network after simply quantizing the signals. The relays in the linear code need not know the channel gains on either the incoming or the outgoing links. The transmit and receive signals at all relays are simply quantized to binary tuples of the same length <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> , which is all that the nodes need to know. The linear network code is a particularly simple scheme and requires all the relay nodes to collect the received binary tuples into a long binary vector and apply a linear transformation on the long vector. The resulting binary vector is split into smaller binary tuples for transmission by the relays. The quantization requirements of the linear network code are completely defined by the parameter <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> , which therefore also determines the resolution of the analog-to-digital and digital-to-analog converters that are required for operating the network within a bounded gap of the network's capacity. As evident from the description, the linear network code explicitly connects network coding for wireline networks with codes for Gaussian networks.

Stabilization of Linear Systems Over Gaussian Networks

The problem of remotely stabilizing a noisy linear time invariant plant over a Gaussian relay network is addressed. The network is comprised of a sensor node, a group of relay nodes and a remote controller. The sensor and the relay nodes operate subject to an average transmit power constraint and they can cooperate to communicate the observations of the plant's state to the remote controller. The communication links between all nodes are modeled as Gaussian channels. Necessary as well as sufficient conditions for mean-square stabilization over various network topologies are derived. The sufficient conditions are in general obtained using delay-free linear policies and the necessary conditions are arrived at using information theoretic tools. Different settings where linear policies are optimal, asymptotically optimal (in certain parameters of the system) and suboptimal have been identified. For the case with noisy multidimensional sources controlled over scalar channels, it is shown that linear time varying policies lead to minimum capacity requirements, meeting the fundamental lower bound. For the case with noiseless sources and parallel channels, nonlinear policies which meet the lower bound have been identified.

A Decode and Forward Protocol for Two-Stage Gaussian Relay Networks

We propose a multihopping decode and forward relaying protocol for two-stage Gaussian relay networks with half-duplex nodes. We analytically show that the achievable rates in suitably defined strong and weak interference regimes are close to the cut-set bound.

Divide-and-Conquer: Approaching the Capacity of the Two-Pair Bidirectional Gaussian Relay Network

The capacity region of multi-pair bidirectional relay networks, in which a relay node facilitates the communication between multiple pairs of users, is studied. This problem is first examined in the context of the linear shift deterministic channel model. The capacity region of this network when the relay is operating at either full-duplex mode or half-duplex mode for arbitrary number of pairs is characterized. It is shown that the cut-set upper-bound is tight and the capacity region is achieved by a so called divide-and-conquer relaying strategy. The insights gained from the deterministic network are then used for the Gaussian bidirectional relay network. The strategy in the deterministic channel translates to a specific superposition of lattice codes and random Gaussian codes at the source nodes and successive interference cancelation at the receiving nodes for the Gaussian network. The achievable rate of this scheme with two pairs is analyzed and it is shown that for all channel gains it achieves to within 3 bits/sec/Hz per user of the cut-set upper-bound. Hence, the capacity region of the two-pair bidirectional Gaussian relay network to within 3 bits/sec/Hz per user is characterized.