Improving Amplify-and-Forward Relay Networks:
Optimal Power Allocation versus Selection
Yi Zhao, Raviraj Adve and Teng Joon Lim
Dept. of Electrical and Computer Engineering,
University of Toronto,
Toronto, ON, M5S3G4, Canada.
E-mail: zhaoyi, rsadve, firstname.lastname@example.org.
We consider an Amplify-and-Forward cooperative diversity system where a source node communicates with
a destination node with the help of one or more relay nodes. The conventional system model assumes all relay
nodes participate, with the available channel and power resources equally distributed over all nodes. This approach
being clearly sub-optimal, we first present two power allocation schemes to minimize the system outage probability,
based on complete channel state information and channel statistics, respectively. We further show that the proposed
optimal power allocation methods minimize system symbol error rate as well. Next, we propose a selection scheme
where only one "best" relay node is chosen to assist in the transmission. We show that the selection-AF scheme
maintains full diversity order, and at reasonable power levels has significantly better outage behavior and average
throughput than the conventional all-participate scheme or that with optimal power allocation. Finally we combine
power allocation and selection to further improve performance.
Keywords: Cooperative System, Amplify-and-Forward, Optimal Power Allocation, Selection Scheme.