Abstract: A mixed spectrum handoff algorithm which combines reactive spectrum handoff and proactive spectrum handoff is proposed in Cognitive Radio Networks（CRNs）. Based on the continuous-time markov chain of Primary Users’（PUs’） channels, the algorithm predicts the future channel state and performs proactive spectrum handoff for serving Secondary Users（SUs）. Meanwhile, the algorithm performs reactive spectrum handoff for collision SUs. Simulation results show that compared with the reactive spectrum handoff algorithm, the proposed algorithm can dramatically reduce the number of collisions between SUs and PUs on the premise of unchanged SU’s blocking probability and outage probability. Moreover, the spectrum efficiency of the CRN can be increased in the proposed algorithm.

Key words: cognitive radio, spectrum handoff, continuous-time markov chain, number of collisions

摘要： 针对认知无线电网络，提出了一种将被动式频谱切换与主动式频谱切换相结合的混合频谱切换算法。该算法基于主用户信道的连续时间马尔可夫链模型，预测出信道的未来状态信息，根据该预测结果周期性地对正在通信的认知用户执行主动式频谱切换。该算法对于由于碰撞而退出信道的认知用户执行被动式频谱切换。仿真结果表明，相对于被动频谱切换算法，混合频谱切换算法在保持认知用户阻塞概率和中断概率不变的前提下可显著减少认知用户和主用户间的碰撞次数，能够提高认知无线电网络的频谱利用率。

关键词: 认知无线电, 频谱切换, 连续时间马尔可夫链, 碰撞次数