This paper considers the problem of selecting a subset of nodes in a two-hop wireless network to act as relays in aiding the communication between the source-destination pair. Optimal relay subset selection with the objective of maximizing the overall throughput is a difficult problem that depends on multiple factors including node locations, queue lengths and power consumption. A partial decode-and-forward strategy is applied in this paper to improve the tractability of the relay selection problem and performance of the overall network. Note that the number of relays selected ultimately determines the performance of the network. This paper benchmarks this performance by determining the net diversity achieved using the relays selected and the partial decode-and-forward strategy. This framework is subsequently used to further transform relay selection into a simpler relay placement problem, and two proximity-based approximation algorithms are developed to determine the appropriate set of relays to be selected in the network. Other selection strategies such as random relay selection and a greedy algorithm that relies on channel state information are also presented. This paper concludes by showing that the proposed proximity-based relay selection strategies yield near-optimal expected rates for a small number of selected relays.
Relay-assisted communication is a promising strategy for both centralized and decentralized communication networks [1,2]. Two-hop relay-based communication is having a considerable influence on emerging standards both in local area networks, IEEE 802.11s [1] and broadband wireless access networks, IEEE 802.16j [2]. Two-hop relay systems consist of a source, a destination and one or more relays where the relay nodes work together as a single set of intermediaries between the source and the destination [3]. Direct transmission occurs between the source and the destination, and the relays assist the source only if the destination cannot decode the direct transmission. There are multiple concrete benefits of introducing these intermediate relays, which include improved system throughput and greater coverage [2]. Multihop relaying [4,5] is a key enabling technology for networks of the future, but before the performance tradeoffs of multihop relaying can be characterized, it is critical that the issues facing two-hop relaying be fully understood.
Given that the source can enlist multiple nodes to simultaneously act as relays, two questions naturally arise. First, how many relays must the source enlist to aid its transmission to gain the maximum advantage for the resources consumed? Second, which of the nodes in the pre-existing network must be enlisted to act as relays? When multiple-relay selection is allowed, there are numerous tradeoffs that govern system performance [5][6][7]. While selecting a large number of relays offers the benefit of coherent combining, resulting in increased throughput and thus higher overall quality of service, it suffers from drawbacks as well. Firstly, system resources are drained faster when multiple relays are selected. Second, there are complexity and implementation issues -it is difficult to synchronize the transmissions from multiple disparate relays [8][9][10][11],
and receiver complexity increases with the number of relays. A single relay can be selected to assist the source transmission [12][13][14][15][16][17][18][19], which offers lower gains in terms of total diversity and rate but is simpler to implement and consumes less power over the entire network. This paper has two goals. One goal is to understand the fundamental limits of multiple-relay selection to benchmark various relay selection algorithms. To this end, we focus on minimizing relay power consumption and treat implementation issues and complexity as a secondary concern.
Regardless of the number of relays selected, it is difficult to determine which node(s) in the network must act as relays to aid the source transmission. For example, selecting the relay with the best channel to the destination may not be an optimal strategy, as this relay may be heavily loaded with traffic and running low on resources. Thus, relay selection is a very difficult problem, as selecting the “optimal” subset from the set of candidate relay nodes is affected by the presence of multiple parameters that govern system performance. In Paper: J3-TVT, First Revision, First Draft, October 23, 2018 particular, relay node selection often translates to a combinatorial optimization problem [31], which currently does not have an elegant polynomial-time algorithmic solution.
The second goal of this paper is to provide algorithms for relay node selection that serve as a good approximation to the problem of optimal relay selection from the point of view of throughput maximization with power allocation. Moreover, we desire the algorithms to have low complexity and be highly intuitive in terms of design. Note that any selection algorithm is closely coupled with the transmission strategy employed in the network (such as decode/amplify/compress-and-forward). Thus, we discuss the transmission strategy employed in this paper and then delve into the details of the algorithms.
In our paper, we use a partial decode-and-forward transmission strategy proposed in [21] 1 . Partial decode-and-forward as described in [21] relies on a two-level superposition coding strategy introduced by T. Cover for broadcast channels [26] and further studied in [27][28][29]. Under this setting, the transmitter employs a layered coding strategy, allowing the receiver to decode the transmitter’s message partially if it is incapable of determining it in its entirety. Note that the conventional decode-and-forward strategy as in [22] is a special case of the partial decode-and-forward strategy, and therefore, partial decode-and-forward is a useful tool that has all the properties of decode and forward incorporated into it. In particular, partial decode-and-forward offers both the diversity advantages of amplify-and-forward and the inherent robustness to noise of decode-and-forward [20].
The other main advantage of partial decode-and-forward is the tractability it lends to the relay selection problem. While multiple-relay selection based on partial decode-and-forward transmission
