Energy-Efficient Delay-Constrained Transmission and Sensing for Cognitive Radio Systems

TL;DR

This paper develops a low-complexity policy for energy-efficient, delay-constrained transmission in cognitive radio systems, balancing spectrum sensing overheads and data transmission to optimize energy and delay performance.

Contribution

It formulates the problem as a Markov Decision Process and proposes a Certainty Equivalent Control-based policy that effectively manages energy-overheads and delay constraints.

Findings

Energy-overheads increase SU transmission rate but with diminishing marginal impact.

Impact of energy-overheads is more significant for delay-insensitive traffic.

Reducing spectrum sensing errors improves SU transmission strategies.

Abstract

In this work we study energy-efficient transmission for Cognitive Radio (CR) which opportunistically operates on Primary User's (PU's) channel through spectrum sensing. Spectrum sensing and compulsory idling (for incumbent protection) introduce energy-overheads for Secondary User's (SU's) operations, and thus an appropriate balance between energy consumption in data transmission and energy-overheads is required. We formulate this problem as a discrete-time Markov Decision Process (MDP) in which the SU aims at minimizing its average cost (including both energy consumption and delay cost) to finish a target traffic payload through an appropriate rate allocation. Based on Certainty Equivalent Control, we propose a low-complexity rate-adaptation policy that achieves comparable performance as the optimal policy. With the low-complexity policy, we quantify the impact of energy-overheads (including the power consumption for spectrum sensing and compulsory idling) on the SU transmission strategy. Specifically, the SU rate increases with the increase of energy-overheads, whose marginal impact, however, diminishes. Moreover, the marginal impact of energy-overheads is more significant for delay-insensitive traffic compared to that for delay-sensitive traffic. To mitigate the loss due to imperfect spectrum sensing, we quantify that the SU decreases (increases) its rate with a larger mis-detection probability (false alarm probability).