基于DAQL算法的动态频谱接入方案

Dynamic spectrum access based on DAQL

针对传统的动态频谱接入方案一般没有考虑自主性,不具备普适性这一缺点,提出了一种基于双动作Q学习算法DAQL(double action Q-learning)的频谱接入方案,该方案将DAQL引入到多授权用户存在的环境下频谱接入问题中,用以降低接入未知频谱环境时的冲突概率。仿真结果表明,提出的方案与随机接入方案相比,不但有更小的冲突概率,而且能动态适应环境的变化,适合认知无线电的需要。

As for the traditional dynamic spectrum access schemes that are not fit to all kinds of instances and do not consider the ability of self-determination, a scheme based on Double action Q-learning （DAQL） was proposed to solve the problem of dynamic spectrum access. This scheme could reduce the collision probability when accessing unknown spectrum by introducing DAQL into the scenario where there were many licensed users. Simulation results show that this scheme has lower collision probability than random access scheme, and can be adaptive to the change of environment so that it can satisfy the need of cognitive radio.