【摘要】：物理層網(wǎng)絡(luò)編碼(PNC)作為一種新穎的中繼傳輸技術(shù),在頻譜效率和吞吐量方面具有巨大優(yōu)勢而備受人們關(guān)注。目前,關(guān)于PNC的研究大多假設(shè)兩個用戶節(jié)點發(fā)送的符號在中繼處完全對齊,即沒有符號同步誤差。然而,實際情況往往并不滿足這一條件,造成系統(tǒng)性能下降。因此,PNC的符號時鐘估計至關(guān)重要。本文研究雙向中繼信道PNC的符號時鐘估計問題,基于CAZAC訓練序列先后提出三種全新的適用于不同采樣率下的時鐘估計算法,分別是:過采樣率(每個碼元的采樣點數(shù))Q≥4的基于DFT插值的時鐘估計算法、過采樣率Q=2的低過采樣率時鐘估計算法,和波特率采樣的時鐘估計算法。論文首先根據(jù)PNC信號模型以及最大似然估計準則,利用訓練序列的零自相關(guān)特性,得到兩個端節(jié)點各自的時鐘誤差估計似然函數(shù),進而提出運用基于DFT的插值算法來估計時鐘誤差。然后針對上述算法采樣率過高的問題,提出一種Q=2下的低過采樣率時鐘估計算法。該算法根據(jù)似然函數(shù)的兩個離散值之差與時鐘誤差是一一對應(yīng)的關(guān)系,利用已知的發(fā)送脈沖形狀和最小二乘法求一個擬合多項式來近似這種函數(shù)關(guān)系,從而實現(xiàn)時鐘誤差估計。最后為了進一步降低采樣率,通過分析波特率采樣下的接收信號,定義一種指示時鐘誤差大小的定時函數(shù),并估計出該定時函數(shù),實現(xiàn)波特率采樣下的時鐘估計。三種算法均在MATLAB平臺上以均方誤差(MSE)為性能指標進行仿真,同時探討各種參數(shù)對算法性能的影響。仿真表明,三種算法都能取得較好的時鐘估計結(jié)果,在32位的訓練序列長度及18 d B的信噪比(SNR)條件下,MSE可達1×10~(-4)。
[Abstract]:As a novel relay transmission technology, physical layer network coding (PNC) has great advantages in spectrum efficiency and throughput. At present, most researches on PNC assume that the symbols sent by the two user nodes are completely aligned at the relay, that is, there is no symbol synchronization error. However, the actual situation often does not meet this condition, resulting in a deterioration of system performance. Therefore, the symbol clock estimation of PNC is very important. In this paper, the symbol clock estimation problem of bidirectional relay channel PNC is studied. Based on the CAZAC training sequence, three new clock estimation algorithms are proposed, which are suitable for different sampling rates. They are: the clock estimation algorithm based on DFT interpolation for oversampling rate (sampling points per symbol) Q 鈮，

本文編號：2289344