Transmit Antenna Selection in Underlay Spectrum Sharing: Role of Power Adaptation, Interference Constraint, and Channel State Information

Abstract

Spectrum sharing is essential to support high wireless data rates and a large number of wireless connections by the future technologies in the limited spectrum that is available. In underlay spectrum sharing a secondary user (SU) transmits simultaneously in the same spectrum as the high priority primary user (PU) while ensuring that the interference caused to the PU is constrained. These interference constraints can limit the SU's performance significantly.

Transmit antenna selection (TAS) is a low hardware complexity multiple antenna technique that exploits spatial diversity from multiple antennas to improve the SU's performance. In it, the secondary transmitter (STx) selects one among the multiple antennas and connects it to the one available radio frequency chain to transmit. TAS rule specifies the antenna to transmit as a function of the channel gains from the STx to the primary receiver (PRx) and the channel gains from the STx to the secondary receiver. We study the combined impact of the power adaptation technique employed at the STx, interference constraint, and the channel state information (CSI) available at the STx on TAS for underlay spectrum sharing.

In this thesis, we focus on characterizing optimal TAS rules that minimize the average symbol error probability (SEP) of an underlay secondary system that is subject to a general class of stochastic interference constraints. This includes the average interference constraint and the interference-outage constraint. We first focus on a secondary system that employs on-off power adaptation at the STx and is subject to the interference-outage constraint. For it, we propose a TAS rule that minimizes the average SEP when the STx shares spectrum with a single PRx and has the instantaneous CSI of the link from it to the PRx. We prove it to be SEP-optimal for a general class of fading models with a continuous cumulative distribution function. We derive expressions for its interference-outage probability and average SEP for both perfect and imperfect CSI. We show that it reduces the average SEP significantly compared to the existing selection rules in the literature.

We then generalize the above model to consider the practical scenario in which the SU shares spectrum with multiple PRxs. For it, we develop a TAS rule that minimizes the average SEP when the STx has instantaneous CSI of only a subset of the STx-PRx links. We study the impact of the STx-PRx CSI on the optimal rule by considering the cases when the STx knows channel gains to all the PRxs and when it does not know channel gains to any of the PRxs. We derive expressions for its interference-outage probability and average SEP. We observe that the performance of the optimal rule is insensitive to the interference power threshold when the STx does not know the instantaneous CSI of all the STx-PRx links.

Next, we consider a secondary system that employs continuous power adaptation, is subject to the interference-outage and peak transmit power constraints, and shares the spectrum with a single PRx. For it, we present an SEP-optimal joint antenna selection and power adaptation rule that applies to the class of fading models with continuous cumulative distribution function. We show that the optimal rule achieves a one to two orders of magnitude reduction in the average SEP compared to the existing rules in the literature. We then propose a simpler, yet near-optimal, variant called the linear rule. We derive tight bounds for its average SEP and interference-outage probability.

Lastly, we study a practically-motivated system model in which the STx only has the statistical information of the links from itself to the PRx. We derive the joint optimal antenna and power adaptation rule for it. We show that it has an intuitive and separable form for the general class of stochastic interference constraints. We evaluate the impact of the power adaptation, interference constraint, and multiple antennas on the performance of the secondary and primary systems.