LiNKS

Energy Efficient Relaying for Wireless Communications

Energy Efficient Relaying for Wireless Communications

Motivation and Project Description:

In the urge to shift from capacity-oriented to energy-oriented network designs, research has taken significant steps in modeling the energy efficiency and new adequate metrics have been proposed. Whereas the capacity and coverage improvements achieved through cooperative networks have been widely investigated, the energy impact of relaying is not yet clearly understood and several questions remain open:

  • What is the minimal energy consumption that can be expected from relaying?
  • In which conditions such minimal consumption can be achieved?
  • How does relay deployment affect the overall energy consumption?
  • Does minimizing the energy consumption necessarily lead to interference reduction?

In this research work, we bring together two different aspects of energy-efficient relaying: the geographical deployment of relay stations and the choice of the relaying coding scheme. While such aspects may seem disconnected at first, we show that information theory has also its role to play in the energy issue and can be considered as a fundamental key to unlock energy and cost efficiency in relay-aided cellular networks.

 

A New Half-Duplex Partial Decode-Forward Relaying Coding Scheme:

Figure 1. Proposed Half-Duplex Partial Decode-Forward Relaying Coding Scheme
Figure 1. Proposed Half-Duplex Partial Decode-Forward Relaying Coding Scheme

As first step, we question the use of maximum-rate coding schemes for the purpose of energy-efficiency analysis and define the minimal energy consumed by decode-forward relaying. To this end, we design the new half-duplex relaying coding scheme and propose for it three power allocations to maintain a desired source rate and minimize the energy consumption of: a) the considered three-node relay network, b) the relay alone and c) the source alone.

From a theoretical aspect, an important result is that minimizing the network energy consumption is not equivalent to maximizing the network capacity as it is often believed. The whole range of source rates achievable with decode-forward is not covered by any of the individual schemes, but only by their combination. The proposed energy- efficiency approach has a certain benefit over the maximum-rate approach since it leads to the closed-form solution of the optimal power allocation and not solely to a solution on the form ”argmax f(x)”. Moreover, it allows a comprehensive description of the optimal coding (full or partial decode-forward, with or without beamforming etc.).

Energy Optimization:

We propose to optimize the above coding scheme for energy efficiency in the Gaussian relay channel at a given source rate. We consider three cases: the network (both source and relay), the relay and the source power consumptions. We look for the optimal power allocation such that the target rate is achieved with minimum power consumption. The three optimization problems are solved analytically based on the KKT conditions.

  • Network Energy Optimal Set of Power Allocation (N-EE)
  • Relay Energy Optimal Set of Power Allocation (R-EE)
  • Source Energy Optimal Set of Power Allocation (S-EE)
    The optimal scheme for source energy efficiency is a function of the source rate and is composed of four sub-schemes: two-hop relaying (D), decode- forward (B), partial decode-forward with beamforming (A), and partial decode-forward without beamforming (C). Note that two-hop relaying and partial decode-forward without beamforming are optimal only for source energy but not for relay or network energy.

Figure . Proposed Half-Duplex Partial Decode-Forward Relaying Coding Scheme
Figure 2. Power consumption for N-EE and allocation set during phase 2.
Figure . Proposed Half-Duplex Partial Decode-Forward Relaying Coding Scheme
Figure 3. Power consumption for R-EE and allocation set during phase 2.
Figure . Proposed Half-Duplex Partial Decode-Forward Relaying Coding Scheme
Figure 4. Applied sub-schemes in S-EE, as function of the SR- and RD- links.

 
 
 
 
 

Impact of Propagation Environment on Energy-Efficient Relay Placement:

Figure 2. Energy-oriented service area for Relays
Figure 5. Energy-oriented service area for Relays.

While advanced coding schemes and resource allocations can overcome the weaknesses of simpler schemes as two-hop relaying, they should be used wisely given their increased hardware complexity. It is thus important to investigate in which conditions relaying can provide worthwhile energy gain, i.e. how the radio propagation environment impacts the relay performance.

As second step, we investigate relay deployment with regards to the channel propagation characteristics and the relaying coding scheme. To propose meaningful analysis without requiring time-consuming extensive simulations, we develop a geometrical model describing the service area of the relay station, accounting for both coverage and energy. Then, we highlight new trade-offs which balance the coverage extension, the energy consumption and the deployment flexibility. We show that advanced coding schemes can easily overcome harsh radio environments due to suboptimal location, and thus provide much flexibility to the relay deployment. Moreover, there exists a relay-to-BS distance for which the coverage can be increased or decreased without harming the energy-efficiency.

 
 

Modeling and Analysis of Energy Efficiency and Interference for Cellular Relay Deployment

Figure 3. Map showing the spatially-optimized utilization of coding schemes as a function of the circuitry consumption
Figure 6. Map showing the spatially-optimized utilization of coding schemes as a function of the circuitry consumption

As the last step, we characterize the impact of omni-directional relays in terms of both the achieved energy gain and the resulted performance loss in neighbourhood cells, due to the additional interference. First, we refine our previous geometrical models to include the effect of shadowing and the extra energy dissipated in the circuitry for signal processing and network maintenance. Second, we propose new definitions for the relaying efficiency. which allow the prediction of the probability of energy-efficient relaying, the overall energy consumption and the generated interference.

The proposed framework has wide application and can be used for resource management (in particular frequency reuse techniques), load-balancing or base-station switch-off. In addition, a new performance metric is proposed to balance the energy gain provided by relay stations within a cell and the additional interference received in the neighbouring cells, based on the actual energy consumption and not on a given fixed power. We thus design a spatially-optimized utilization of coding schemes and draw a map showing the cell areas where to use each coding scheme to maximize such metric.

 
 
 
 
 
 
 
 

Publications:

  1. “Trade-Offs on Energy-Efficient Relay Deployment in Cellular Networks,”
    F. Parzysz, M. Vu, and F. Gagnon, Vehicular Technology Conference (VTC Fall), 2014 IEEE 80th, pp.1-6, 14-17 Sept. 2014.
  2. “Impact of Propagation Environment on Energy-Efficient Relay Placement: Model and Performance Analysis,”
    F. Parzysz, M. Vu, and F. Gagnon, IEEE Transactions on Wireless Communications, vol.13, no.4, pp.2214-2228, April 2014.
  3. “Energy Minimization for the Half-Duplex Relay Channel with Decode-Forward Relaying,”
    F. Parzysz, M. Vu, and F. Gagnon, IEEE Trans. on Comm., vol. 61, no. 6, pp. 2232 – 2247, Jun 2013.
  4. “Optimal Distributed Coding Schemes for Energy Efficiency in the Fading Relay Channel,” 
    F. Parzysz, M.Vu, F. Gagnon, IEEE Int’l Conf. on Comm. (ICC), June 2012.
  5. “A Half-Duplex Relay Coding Scheme Optimized for Energy Efficiency,”
    F. Parzysz, M.Vu, F. Gagnon, IEEE Information Theory Workshop, Oct 2011.
  6. “Energy-Efficient Schemes for On-Demand Relaying,”
    F. Parzysz, M.Vu, F. Gagnon, The 34th IEEE Sarnoff Symposium, May 2011.