標題: 應用於高車速正交分頻多工系統新型通道估計方法之設計與模擬
Design and Simulation of A New Channel Estimation Method for High Mobility OFDM Systems
作者: 葉哲芳
Yeh Che-Fang
黃家齊
Huang Chia-Chi
電信工程研究所
關鍵字: 通道估計;正交分頻多工系統;多路徑干擾消除;適應性牛頓法;channel estimation;OFDM;MPIC;adaptive Newton
公開日期: 2005
摘要: 在多路徑衰減通道高移動速度的正交分頻多工系統,通道估計是非常重要的課題。為因應高資料速率傳輸的需求,不損失太多頻帶使用效率的通道估計方法是緊要的,特別是在快速變動的通道下。我們提出半盲式的通道估計方法,其中包含基於多路徑干擾消除的初始通道估計子,及適應性牛頓法追踨緊接著的通道變化。值得一提的是,在通道追踨過程中,只需要稀疏的領航訊號。此外,沒有限制初始領航訊號的結構及追蹤用稀疏領航訊號的位置,所以此通道估計方法亦可推廣至一般的正交分頻多工系統。此論文中,初始通道估計子是利用IEEE 802.16 正交分頻多工模式下定義的初始領航符元模擬,而牛頓法通道追蹤則模擬在 120 km/hr及240 km/hr 的高車速下。從模擬結果知道,此半盲式通道估計子在錯誤率上有很好的表現。總之,此通道估計方法不僅達到高頻帶使用效率且有不錯的估計準確度。
Channel estimation is an important issue for high mobility OFDM systems in multipath fading channels. In demand of high data transmission rate, a channel estimation method without loss of too much bandwidth efficiency is critical, especially in fast fading channels. We propose a semi-blind channel estimation method which consists of an MPIC-based initial channel estimator, followed by an adaptive Newton tracker to track channel variations in the subsequent OFDM symbols. It is worth to mention that only sparse pilots are needed to track channel impulse responses. Moreover, there are no restrictions on the preamble structure in the initial channel estimator and no restrictions on the locations of the sparse pilots in the tracking stages as well. Thus, the semi-blind channel estimation method can be generally applied to OFDM systems. In this thesis, the initial channel estimator is simulated with the preamble defined in IEEE 802.16 OFDM mode, and the adaptive Newton tracker is verified in high mobility channels (120 km/hr and 240 km/hr).From the simulation results, we can find that the semi-blind channel estimation method performs very well in terms of bit error rate. In conclusion, the semi-blind channel estimation method can achieve both bandwidth efficiency and estimation accuracy.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT009313527
http://hdl.handle.net/11536/78342
Appears in Collections:Thesis


Files in This Item:

  1. 352701.pdf
  2. 352702.pdf
  3. 352703.pdf
  4. 352704.pdf