异构认知网络自适应频谱决策方法

Adaptive Spectrum Decision Method for Heterogeneous Cognitive Radio Networks

To improve spectrum utilization and mini-mize interference to Primary User (PU), an adaptive spectrum decision method is proposed for Seconda-ry User (SU), while taking traffic load balancing and spectrum heterogeneity into consideration. Long-term statistics and current sensing results are integrated into the proposed decision method of spectrum access. Two decision methods, namely probability based and sensing based, are presented, compared and followed by performance analysis in terms of delay. For probability based spectrum deci-sion, Short-Time-Job-First (STJF) priority queuing discipline is employed to minimize average residual time and theoretical conclusion is derived in a novel way. For sensing based decision we treat the inter-rupted service of SU as newly incoming and re-de-cision process is initialized to find available spec-trum in a First-Available-First-Access (FAFA) fash-ion. Effect of sensing error in PHY layer is also an-alyzed in terms of extended average residual time. Simulation results show that, for relatively low arri-ving rate of SU traffic, the proposed spectrum deci-sion method yields at least a delay reduction of 39. 5% compared with non-adaptive method. The pro-posed spectrum decision can significantly improve delay performance even facing sensing errors, which cause performance degeneration to both PU and SU.