Learning-Based Opportunistic Spectrum Access with Adaptive Hopping Transmission Strategy

This paper presents an adaptive hopping transmission strategy for secondary users (SUs) to access temporarily idle frequency-slots of a licensed frequency band in consideration of the random return of primary users (PUs), aiming to maximize the overall SU throughput. A SU dynamically hops over multiple idle frequency-slots, each with an adaptive activity factor to avoid high-risk data loss due to possible PU return. SU activity factor optimization problems are formulated to develop the optimal opportunistic spectrum access (OSA) algorithms for SUs based on the Lagrange dual decomposition method. Subsequently, a fully distributed learning-based OSA algorithm is developed in which each SU independently adapts its activity factors to the optimal values over time by learning other SUs' behavior from locally available information. The convergence and convergence rate that characterize its asymptotic behavior and efficiency are analyzed. It is shown that the proposed learning-based OSA algorithm converges with probability of 1 to the optimal solution. Illustrative results confirm its effectiveness and performance gain as compared to existing OSA schemes.