Game Theoretical Approaches for Transport-Aware Channel Selection in Cognitive Radio Networks

Abstract

Effectively sharing channels among secondary users (SUs) is one of the greatest challenges in cognitive radio network (CRN). In the past, many studies have proposed channel selection schemes at the physical or the MAC layer that allow SUs swiftly respond to the spectrum states. However, they may not lead to enhance performance due to slow response of the transport layer flow control mechanism. This paper presents a cross-layer design framework called Transport Aware Channel Selection (TACS) scheme to optimize the transport throughput based on states, such as RTT and congestion window size, of TCP flow control mechanism. We formulate the TACS problem as two different game theoretic approaches: Selfish Spectrum Sharing Game (SSSG) and Cooperative Spectrum Sharing Game (CSSG) and present novel distributed heuristic algorithms to optimize TCP throughput. Computer simulations show that SSSG and CSSG could double the SUs throughput of current MAC-based scheme when primary users (PUs) use their channel infrequently, and with up to 12% to 100% throughput increase when PUs are more active. The simulation results also illustrated that CSSG performs up to 20% better than SSSG in terms of the throughput.

Publisher note

To access the full article, please see PDF.