在無線通訊中衰減是一個重要的問題,因其引入幾乎無法修正的錯誤層級,並造成通訊品質的嚴重降低,尤其在高速行動通信中,衰減補償更是首先要解決的問題。 近年來,使用指標符號的方式常用於無線高速傳輸系統,以估測指標符號位置的通道頻率響應,再藉由DFT (Discrete Fourier transform)內插運算有效的估測其餘通道的通道頻率響應,藉由這些通道的評估值,對訊號做適當的補償,以找回經由通道衰減的信號。 由於通道衰減本質上屬於一有限頻帶運作,在實務上以DFT內插法應用在頻域上補零 (Zero-padding)為之頗為吻合,然而傳統內插方式使用傅氏轉換的時候,因為訊號兩端點的不連續,造成邊界效應影響了內插的精確度,因而降低通到補償的效果。在論文中,闡述兩種使兩端點能連續或近似成連續訊號的方法,運用時域邊界匹配法之觀念,經由邊界匹配的調整,可改善因傅氏轉換內插法而產生的誤差,進而提升估測通道衰減的精確度。
There are many problems occurred in wireless communications destructing the signals transmitted. Fading is one of the main problems to affect the transmitting signals, hence the fading compensation is the most important problem to be soled especially in high-speed mobile communication. Pilot symbol-assisted modulation is an effective scheme used to solve fading problems providing good effect to recover the destructed signals to the original ones. This modulation uses the pilots to estimate the fading effect and remove the estimated fading effect and then the signals will be recovered. The boundary effect by discontinuously of the two ends of the points due to the period extend characteristic of the DFT could reduce the accuracy of the evaluation in pervious research. In this thesis, we discuss the Windowed method and Boundary Matched method to improve the traditional Discrete Fourier transform interpolation in channel estimation. By the simulating results, we can find these two methods can reduce the error of DFT interpolation, and improve the accuracy of channel estimation.