认知无线电网络中授权用户参数估计

Estimation of primary user parameters in cognitive radio networks

针对认知用户在频谱切换过程中无法实时地获取授权用户到达率与服务率的问题,提出了基于隐式马尔可夫模型的参数估计算法。首先利用排队论对授权用户队列进行建模与分析,推导出授权用户队列状态转移概率;其次利用能量感知算法检测授权用户队列真实状态,获得可观测序列值;然后利用隐式马尔可夫模型描述两种随机过程,即授权用户队列状态变化随机过程和可观测序列随机过程;最后利用forward-backward算法估计隐式马尔可夫模型,从而获得授权用户到达率与服务率。仿真结果表明,该方法能够实现实时的、较为精确的估计,从而实时地为认知用户选择频谱切换策略提供依据。

To solve the problem that secondary users cannot real-timely obtain arrival rates and service rates of primary users during the handoff period, a parameter estimation algorithm based on the hidden Markov model is proposed. Firstly, primary user queues are modeled and analyzed by the queue theory, and the state transition probability of primary user queues is derived. Real states of primary user queues are detected by the energy sensing algorithm, so that observable sequences are obtained. Then, two kind of stochastic processes are de- scribed by the hidden Markov process, namely, the stochastic process of state transition of primary user queues and the stochastic process of observable sequences. Finally, the hidden Markov process is estimated by the for- ward-backward algorithm, so that the arrival rate and the service rate of primary users are obtained. Simulation results verify that real-time and relatively accurate estimation can be realized, and a foundation for secondary us- ers is provided to choose a spectrum handoff strategy in real time.