- 學位論文
Blind Maximan-Likelihood Detection for Decode-and-forward Randomized Distributed OSTBC
解碼轉送隨機分散式正交空時區塊編碼之盲蔽最大似然偵測
摘要
In this thesis, we consider the randomized distributed orthogonal space-time block coding (DOSTBC) system proposed by Sirkeci-Mergen and Scaglione [3], and consider a situation where some of the cooperative relays can unexpectedly disconnect from the network during the data frame transmission. Such an event translates into abrupt changes of the virtual channel matrix at the receiver; and a coherent maximum-likelihood (ML) detector, which assumes that the virtual channel matrix is perfectly known and static, can be severely degraded in performance. In fact, the simulation results reveal that the coherent ML detector with mismatched channel state information (CSI) exhibits an error floor for high signal-to-noise ratio (SNR). We propose a blind maximum-likelihood (ML) detector and a non-intersecting subspace (NIS) code scheme for the randomized DOSTBC system. With a mild assumption on the transmission protocol, we show that the proposed blind ML detector is robust against the unexpected relay disconnection problem. Moreover, we show that the randomized NIS code scheme can achieve the maximum transmit diversity with the blind ML receiver. Some simulation results are presented to demonstrate the efficacy of the proposed method. Furthermore, we propose a full blind ML detector which does not require any pilot bits for solving the inherent sign ambiguity problem, which thereby is more spectrum efficient. Some simulation results are presented to demonstrate the efficacy of the proposed method.
並列摘要
在訊號從傳送端透過一群中繼器（relay）替代傳統多輸入多輸出系統的天線，傳送資料到接收端的合作式通訊之中，本論文應用Mergen與Scaglione[3]所提出的隨機分散式正交空時區塊編碼（randomized distributed orthogonal space-time block coding）系統，考慮有一些中繼器在資料傳輸途中突然從中繼網路斷訊的問題；因而造成中繼器與接收端之間的虛擬通道矩陣（virtual channel matrix）轉變，對於一個假設已知虛擬通道資訊的同調最大似然檢測器（coherent maximum-likelihood (ML) detector）而言，其效能會嚴重變差。在我們所做的模擬結果，顯示了同調最大似然檢測器在資料傳輸的過程當中，遇到中繼器無預警斷訊的問題時，在訊雜比（signal-to-noise）很高時，會出現錯誤地板（error floor）現象。 吾等提出盲蔽最大似然檢測器（blind ML detector），其使用子空間無交集碼（non-intersecting subspace codes）方法的隨機分散式空時區塊編碼系統，只需一個前導位元（pilot bit），就能保證資料鑑別的唯一性（unique data identifiability）。加上一些通訊協定上的簡單假設，分析證明我們所提出來的盲蔽最大似然檢測器可以有效抵抗中繼器無預警斷訊的問題。此外，我們也證明了盲蔽最大似然接收端使用隨機子空間無交集碼的方法，可以達到最大之傳送分集（maximum transmit diversity）。另外，我們還提出了全盲蔽最大似然檢測器（full blind maximum-likelihood detector），其不需要前導位元就能解決內在符號歧異性（inherent sign ambiguity）的問題，因此傳輸效率更高。最後的模擬結果亦印證了我們所提出的方法之效能。