基于隐藏马尔可夫模型的信道质量预测
发布时间：2018-12-13 18:21
【摘要】：频谱预测是认知无线电中的一项关键技术,根据统计信道的历史信息,分析频谱的使用规律,从而找出不同用户的信道占用特点和频谱空洞出现的规律。根据频谱预测的结果,认知用户可以选择最优的信道,或是提前撤出主用户可能会占用的信道。所以,准确的频谱预测能够减小频谱空洞的错误感知概率,主动地减少干扰和延迟,提升频谱利用率,提高网络的吞吐量。针对一阶隐藏马尔可夫模型不能充分利用历史序列的有效信息因而准确度不足的问题,本文改进了传统的一阶隐藏马尔可夫模型来计算信道感知不完美条件下高阶的信道状态转移概率和散射概率。同时,针对现有文献中应用的高阶隐藏马尔可夫模型只适用于信道占用、空闲时长的分布为指数分布的这种假设与实际情况并不相符的问题,本文提出一种适用于所有信道状态分布的高阶频谱预测算法。在保证了准确率的同时,并不需要信道状态持续时间为指数分布。此外,根据高阶隐藏马尔可夫模型的频谱预测结果,本文提出了一种结合感知准确度和信道空闲概率的信道质量评价标准,根据该标准,认知用户可以选择接入质量更好的信道,使得频谱得到更加高效的利用。经过仿真,验证了高阶隐藏马尔可夫模型下的频谱预测算法可以有效地解决不完美感知导致的预测准确度不足的问题；同时验证了本文所提出的信道质量预测标准在不同信道分布的场景下的适用性。
[Abstract]:Spectrum prediction is a key technology in cognitive radio. According to the historical information of statistical channels, the use of spectrum is analyzed to find out the characteristics of different users' channel occupation and the regularity of spectrum holes. According to the result of spectrum prediction, cognitive users can choose the optimal channel or withdraw the channel that the primary user may occupy. Therefore, accurate spectrum prediction can reduce the error perception probability of spectrum holes, actively reduce interference and delay, improve spectrum efficiency and improve the throughput of the network. In order to solve the problem that the first order hidden Markov model can not make full use of the effective information of the historical sequence, the accuracy of the model is insufficient. In this paper, the traditional first order hidden Markov model is improved to calculate the high order channel state transition probability and scattering probability under the condition of imperfect channel perception. At the same time, the assumption that the high order hidden Markov model used in the existing literature is only applicable to channel occupancy and the distribution of idle time is exponential is not in accordance with the actual situation. In this paper, a high order spectrum prediction algorithm for all channel state distributions is proposed. At the same time, the channel state duration is not required to be exponential distribution. In addition, according to the spectrum prediction results of high order hidden Markov model, a channel quality evaluation standard combining perceptual accuracy and channel idle probability is proposed. Cognitive users can select channels with better access quality, which makes the spectrum more efficient. The simulation results show that the spectrum prediction algorithm based on high order hidden Markov model can effectively solve the problem of poor prediction accuracy caused by imperfect perception. At the same time, the applicability of the proposed channel quality prediction standard in different channel distribution scenarios is verified.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN925
[Abstract]:Spectrum prediction is a key technology in cognitive radio. According to the historical information of statistical channels, the use of spectrum is analyzed to find out the characteristics of different users' channel occupation and the regularity of spectrum holes. According to the result of spectrum prediction, cognitive users can choose the optimal channel or withdraw the channel that the primary user may occupy. Therefore, accurate spectrum prediction can reduce the error perception probability of spectrum holes, actively reduce interference and delay, improve spectrum efficiency and improve the throughput of the network. In order to solve the problem that the first order hidden Markov model can not make full use of the effective information of the historical sequence, the accuracy of the model is insufficient. In this paper, the traditional first order hidden Markov model is improved to calculate the high order channel state transition probability and scattering probability under the condition of imperfect channel perception. At the same time, the assumption that the high order hidden Markov model used in the existing literature is only applicable to channel occupancy and the distribution of idle time is exponential is not in accordance with the actual situation. In this paper, a high order spectrum prediction algorithm for all channel state distributions is proposed. At the same time, the channel state duration is not required to be exponential distribution. In addition, according to the spectrum prediction results of high order hidden Markov model, a channel quality evaluation standard combining perceptual accuracy and channel idle probability is proposed. Cognitive users can select channels with better access quality, which makes the spectrum more efficient. The simulation results show that the spectrum prediction algorithm based on high order hidden Markov model can effectively solve the problem of poor prediction accuracy caused by imperfect perception. At the same time, the applicability of the proposed channel quality prediction standard in different channel distribution scenarios is verified.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN925
【相似文献】
相关期刊论文 前10条
1 戴迎s，
本文编号：2377008
本文链接：https://www.wllwen.com/kejilunwen/wltx/2377008.html
