摘要
频谱数据库是一种直接获得频谱信息的方式,针对在复杂多变的网络环境中用户之间缺少信息交互,研究了数据库协助和没有数据库情况下的动态频谱接入算法。一方面,次用户通过历史感知数据估计信道的可用性而进行信道选择,另一方面,次用户通过数据库获得更加可靠的频谱信息从而做出策略。证明了用户之间的博弈是一个超模博弈,通过提出的分布式学习算法能收敛到一个纯策略纳什均衡点。仿真结果表明,提出的联合数据库感知算法和数据库协助算法比依靠感知的结果收敛更快,获得的系统吞吐量接近最大,减少了用户决策的时延,提高了频谱利用率。
Spectrum database could directly acquire work environment Which is lacking information exc spectrum information, and in view of the complex net- hange of between the users, dynamic spectrum access algorithm with and without the aid of database is discussed in this paper. On the one hand, secondary us- ers estimate the availability probability and select channels with historically sensed data, on the other, they acquire farly reliable spectrum information from spectrum database and make proper decision. The game of between the users proves to be a supermodular game, and two distributed leaning algorithms are proposed to achieve the pure strategic NE (Nash Equilibrium). Simulation results show that the proposed database- jointed and database-aided perception algorithms could enjoy fairly fast convergence, nearly maximum throughput, and thus decrease the decision time-delay of decision and improve the spectrum utilization.
出处
《通信技术》
2016年第4期426-430,共5页
Communications Technology
基金
国家自然科学基金(No.61301161
No.61471395)
江苏省自然科学基金(No.BK20141070)~~