非理想通道資訊及時空關連通道下之空間調變信號偵測

Abstract

空間調變(SM)是一個高效率的多天線傳輸方法。它沒有傳統多天線系統中通道間交互干擾(ICI) 的問題、不需建立多傳輸天線資料鏈的同步，傳送端只需要一個射頻鏈路(RF chain)並可將資訊置於傳送天線序號中。 為偵測、解調SM訊號，接收端需有通道資訊(CSI)。大多數的研究都假設通道是非時變的(time-invariant)並有完整正確的CSI，這樣的假設並不實際。此外一般的CSI通道估計演算通常忽略或是不考慮通道的空間與時間的關連性(spatial and temporal correlations)，其推論或估測之正確性有頗多改善的餘地。 本論文探討了決策引導(decision-directed)與基於模型(model-based) 兩種CSI估測法。我們同時考慮通道估計誤差(CSI estimation error)及其時空關連性並分別推導相關的最佳解調器。 此外，我們進一步提出了兩種可大幅降低解調器運算複雜度的解調器結構。針對所提出的各種解調器結構，我們透過了電腦模擬來檢驗其效能並與現有的方法比較。這些數據顯示我們所提出的空間調變解調器改善了現有解調器的效能，且低運算複雜度的解調器與最佳解調器的效能差別幾乎是可忽略的。

Spatial modulation (SM) is a promising multiple antennas based transmission scheme that induces no inter-(spatial) channel interference (ICI) and does not require timing synchronization amongst multiple spatial data streams. It only needs a single radio frequency (RF) chain and can exploit transmit antenna index for carrying extra information to enhance spectral e?ciency. To detect the SM signals, channel state information (CSI) is need at the receive side. Most SM detection studies assume that the channel is time-invariant and CSI is perfectly known. In reality, CSI is never perfect and varies in time. Moreover, spatial correlations among multiple spatial channels exist but are ignored in channel estimation. In this thesis we release the above assumptions and develop a class of decision-directed and a model-based channel estimator. Based on the proposed channel estimators we derive optimal detectors that take into account both the CSI error and time-spatial correlation. To simplify the optimal detector, we then proposed two low-complexity suboptimal structures for each optimal detector. Computer simulations are carried out to estimate the corresponding performance and numerical results show that the optimal detectors o?er performance gain against conventional SM detector and the suboptimal detectors incur negligible performance loss.