【摘要】：頻譜預(yù)測(cè)是認(rèn)知無(wú)線電中的一項(xiàng)關(guān)鍵技術(shù),根據(jù)統(tǒng)計(jì)信道的歷史信息,分析頻譜的使用規(guī)律,從而找出不同用戶的信道占用特點(diǎn)和頻譜空洞出現(xiàn)的規(guī)律。根據(jù)頻譜預(yù)測(cè)的結(jié)果,認(rèn)知用戶可以選擇最優(yōu)的信道,或是提前撤出主用戶可能會(huì)占用的信道。所以,準(zhǔn)確的頻譜預(yù)測(cè)能夠減小頻譜空洞的錯(cuò)誤感知概率,主動(dòng)地減少干擾和延遲,提升頻譜利用率,提高網(wǎng)絡(luò)的吞吐量。針對(duì)一階隱藏馬爾可夫模型不能充分利用歷史序列的有效信息因而準(zhǔn)確度不足的問(wèn)題,本文改進(jìn)了傳統(tǒng)的一階隱藏馬爾可夫模型來(lái)計(jì)算信道感知不完美條件下高階的信道狀態(tài)轉(zhuǎn)移概率和散射概率。同時(shí),針對(duì)現(xiàn)有文獻(xiàn)中應(yīng)用的高階隱藏馬爾可夫模型只適用于信道占用、空閑時(shí)長(zhǎng)的分布為指數(shù)分布的這種假設(shè)與實(shí)際情況并不相符的問(wèn)題,本文提出一種適用于所有信道狀態(tài)分布的高階頻譜預(yù)測(cè)算法。在保證了準(zhǔn)確率的同時(shí),并不需要信道狀態(tài)持續(xù)時(shí)間為指數(shù)分布。此外,根據(jù)高階隱藏馬爾可夫模型的頻譜預(yù)測(cè)結(jié)果,本文提出了一種結(jié)合感知準(zhǔn)確度和信道空閑概率的信道質(zhì)量評(píng)價(jià)標(biāo)準(zhǔn),根據(jù)該標(biāo)準(zhǔn),認(rèn)知用戶可以選擇接入質(zhì)量更好的信道,使得頻譜得到更加高效的利用。經(jīng)過(guò)仿真,驗(yàn)證了高階隱藏馬爾可夫模型下的頻譜預(yù)測(cè)算法可以有效地解決不完美感知導(dǎo)致的預(yù)測(cè)準(zhǔn)確度不足的問(wèn)題；同時(shí)驗(yàn)證了本文所提出的信道質(zhì)量預(yù)測(cè)標(biāo)準(zhǔn)在不同信道分布的場(chǎng)景下的適用性。
[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.
【學(xué)位授予單位】：西安電子科技大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：TN925

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 戴迎s，

本文編號(hào)：2377008