[Paper Review] Optimal Primary-Secondary user Cooperation Policies in Cognitive Radio Networks

Nestor D. Chatzidiamantis, E. Matskani+3 more|arXiv (Cornell University)|Jul 22, 2013

Cognitive Radio Networks and Spectrum Sensing4 citations

TL;DR

This paper proposes optimal primary-secondary user cooperation policies in cognitive radio networks that maximize secondary user (SU) throughput while ensuring primary user (PU) queue stability, even when PU traffic exceeds the threshold for non-cooperation. The solution involves a constrained Markov decision process with randomized, sensing-based transmission policies, proven optimal under infinite SU queue assumptions and extended to imperfect sensing and distributed implementation.

ABSTRACT

In cognitive radio networks, secondary users (SUs) may cooperate with the primary user (PU), so that the success probability of PU transmissions are improved, while SUs obtain more transmission opportunities. Thus, SUs have to take intelligent decisions on whether to cooperate or not and with what power level, in order to maximize their throughput subject to average power constraints. Cooperation policies in this framework require the solution of a constrained Markov decision problem with infinite state space. In our work, we restrict attention to the class of stationary policies that take randomized decisions in every time slot based only on spectrum sensing. The proposed class of policies is shown to achieve the same set of SU rates as the more general policies, and enlarge the stability region of PU queue. Moreover, algorithms for the distributed calculation of the set of probabilities used by the proposed class of policies are presented.

Motivation & Objective

• Address the limitation of prior work that restricts PU traffic to a low threshold for cooperation to be effective.
• Develop optimal cooperation policies for SUs that maximize their throughput while guaranteeing PU queue stability and unobstructed PU transmission.
• Extend the framework to handle imperfect spectrum sensing and distributed implementation for practical deployment.
• Characterize the optimal policy structure under both infinite and finite SU queue assumptions.
• Provide a lightweight distributed protocol enabling real-time, scalable implementation of the proposed cooperation policies.

Proposed method

• Formulate the problem as a constrained Markov decision process (CMDP) with an infinite state space, where the state is the PU queue size.
• Propose a class of stationary, randomized policies that decide SU activation and transmit power based solely on spectrum sensing outcomes.
• Use a parameterization technique to transform the original CMDP into an equivalent convex optimization problem (OPT1), enabling analytical tractability.
• Introduce a new parameter set q(b,s,i) and q(e,s,i) to represent SU behavior during PU busy and idle periods, respectively, ensuring feasibility and optimality.
• Derive necessary and sufficient conditions for the existence of optimal policies using the feasibility of the transformed optimization problem.
• Design a lightweight distributed protocol that enables autonomous, localized decision-making by SUs based on local sensing and feedback, suitable for real-world deployment.

Experimental results

Research questions

• RQ1 How can secondary users be optimally scheduled to cooperate with primary users in cognitive radio networks to maximize their own throughput while ensuring PU queue stability?
• RQ2 What is the structure of the optimal cooperation policy when PU traffic exceeds the threshold for non-cooperation, and how can it be computed efficiently?
• RQ3 How does imperfect spectrum sensing affect the design and performance of optimal cooperation policies?
• RQ4 Can the optimal policy be implemented in a distributed manner without centralized coordination?
• RQ5 Does the optimality of the policy structure persist when the assumption of infinitely backlogged SU queues is relaxed?

Key findings

• The proposed class of randomized, sensing-based stationary policies achieves the maximum possible throughput for secondary users while significantly expanding the stability region of the primary user queue.
• The optimal policy structure remains invariant even when the assumption of infinitely backlogged SU queues is relaxed, ensuring robustness to queue state variations.
• For imperfect sensing, the framework is extended to account for false alarm and miss detection probabilities, maintaining optimality under realistic channel conditions.
• A lightweight distributed protocol is proposed that enables autonomous SU decision-making based on local spectrum sensing, with no need for global coordination or centralized control.
• The existence of an optimal policy is proven to be equivalent to the feasibility of a convex optimization problem (OPT1), which is characterized by explicit bounds on the SU cooperation probability.
• The optimal policy achieves a throughput region that strictly dominates previous approaches, especially in high PU traffic regimes where prior methods fail.