使用卡曼濾波器及延伸卡曼濾波器實現磁儲存系統之通道估測與鎖相回路

Abstract

本論文提出使用卡曼濾波器及延伸卡曼濾波器來實現一 磁儲存系統的通道參數估測及鎖相回路。在本論文中，我們首先以有限脈衝響應的架構來建立磁接收通道的模型。因此其模型參數可以利用卡曼濾波器來估測。然後再由通道特性及所估測之參數來建立鎖相回路的時間函數。由於此時間函數與相位之間的關係可能是非線性，因此最後即可以使用延伸卡曼濾波器來實現鎖相回路。 本論文並針對磁儲存系統，在通道特性資訊有所不足或完全缺乏的各種假設狀況下，藉由近似延伸卡曼濾波器演算法或模型的方法來解決這些問題，且可經由模擬來驗證其可行性。此方法也同樣適用於各種通訊接收系統或儲存系統。

The thesis uses the Kalman filter (KF) and the extended Kalman filter (EKF) for realizing the channel-parameter estimation and a phase-locked loop (PLL) of the magnetic storage system. The magnetic storage system, in this thesis, is modeled by a finite impulse response (FIR) structure. Thus, the model parameters can be estimated by using the KF. These channel parameters are then used to obtain a timing function for the PLL. Since the timing function is commonly a nonlinear function of the phase, the EKF is thus used to realize the phase estimation of a PLL system. The thesis studies the performance of the PLL system for the magnetic storage channel under several conditions of only partial knowledge of the channel information. Since the modeling approach and EKF can obtain approximate parameters even the knowledge regarding the channel characteristic is partially known or fully unknown, simulation results verify the usefulness of the proposed PLL. This approach also applies to most receivers of communication and storage systems.