Optimal Myopic Sensing and Dynamic Spectrum Access in Cognitive Radio Networks with Low-Complexity Implementations

Abstract

Cognitive radio techniques allow secondary users (SU's) to opportunistically access underutilized primary channels that are licensed to primary users. We consider a group of SU's with limited spectrum sensing capabilities working cooperatively to find primary channel spectrum holes. The objective is to design the optimal sensing and access policies that maximize the total secondary throughput on primary channels accrued over time. Although the problem can be formulated as a Partially Observable Markov Decision Process (POMDP), the optimal solutions are intractable. Instead, we find the optimal sensing policy within the class of myopic policies. Compared to other existing approaches, our policy is more realistic because it explicitly assigns SU's to sense specific primary channels by taking into account spatial and temporal variations of primary channels. Contributions: (1) formulation of a centralized spectrum sensing/access architecture that allows exploitation of all available primary spectrum holes; and (2) proposing sub-optimal myopic sensing policies with low-complexity implementations and performance close to the myopic policy. We show that our proposed sensing/access policy is close to the optimal POMDP solution and outperforms other proposed strategies. We also propose a Hidden Markov Model based algorithm to estimate the parameters of primary channel Markov models with a linear complexity.