全雙工異質性小細胞網路:具較真實干擾模型下之傳輸速率分析

Abstract

在此篇論文裡，我們採用了較貼近真實情形下的方針去建立一個全雙工的異質性小細胞網路，由能夠進行全雙工和單工的小基地台以及因為自干擾太強而只能進行傳統單工的大基地台所組成。我們將兩種現實中會遭遇的兩種現象 – 閒置的基地台與不完美的自干擾消除導入此論文的討論內，並讓使用者利用兩種不同的連接基地台方式 – 選擇最接近的與選擇平均最強的方式，來觀察分析全雙工和自干擾是如何影響這個異質性網路。首先，我們研讀使用者與這些基地台的拉普拉斯訊雜比倒數，並將他們整理條列出來；藉由這些式子，我們可以進一步去探討全雙工網路的傳輸速率，並進一步呈現全雙工網路不是在任何情況下都優於傳統的單工傳輸。最後我們展示如何去合適地設置這些基地台的密度，以便讓全雙工和單工的傳輸速率都達到此網路的最佳效率。

In this thesis, we adopt a realistic approach to modeling all interferences in a two-tier full-duplex heterogeneous cellular network (HetNet) consisting of a tier of small cell base stations (BSs) that can perform both half-duplex (HD) and full-duplex (FD) transmission and a tier of macro BSs that can only perform HD transmission due to their strong self-interference. We bring two realistic issues into this thesis - void cell phenomenon and imperfect self-interference cancellation with two kinds of cell association scheme, nearest base station association (NBA) and mean maximum received power association (MRPA), to analyze how full-duplex interference and self-interference affect the downlink throughput of the HetNet. First, we study the Laplace transforms of the reciprocals of the signal-to-interference power ratios of users and small cell BSs and derive the explicit expressions of the ergodic rate with general fading channel models. With these formulas, we then show that the throughput performance of network-wise FD is not always superior to network-wise HD transmission. Also, we demonstrate how to make hybrid FD and HD transmission benefit the ergodic rate by deploying macro and small cell BSs with appropriate intensities.