利用卡曼濾波器設計之DS-CDMA接收機

Abstract

在通訊系統中，多路徑通道一直是需要被解決的問題。在接收機方面，準確的通道訊息有助於訊號的估計，在時變通道的情況尤其重要。本篇論文提出了一種使用卡曼濾波器 (Kalman filter)為基礎所設計的接收機。我們將估計通道以及訊號的兩個卡曼濾波器串接，稱之為串接式卡曼濾波器 (CKF)。但是此一接收機只能依序的估計通道及訊號，本論文又提出了一種利用擴展式卡曼濾波器 (extended Kalman filter, EKF)以同時估計兩者。配合一些假設的情況下，近似過的擴展式卡曼濾波器與串接式卡曼相等。在模擬的結果中，我們比較串接的卡曼濾波器以及配合卡曼使用的決策回授等化器(DFE)接收機。結果可以顯現串接卡曼濾波器的優越性。

The multipath channel is always a problem in communication system. The receiver needs accuracy channel knowledge to estimate symbol. The accuracy of channel is more critical under time-varying fading channel. In this thesis, we proposed a DS-CDMA receiver based on cascading two Kalman filters, estimating the channel and symbol, respectively. We call it the cascaded Kalman filter (CKF). But CKF needs to estimate channel and symbol sequentially, we propose a joint channel and symbol estimator by extended Kalman filter (EKF). With some assumptions, the approximated EKF is the same with CKF. We compare CKF and DFE receiver in simulation. The results prove that CKF perform better than DFE.