Feedforward Adaptive Sampling Time Synchronization Technique

• Main Authors: Ming-Fong Tseng, 曾明鋒
• Other Authors: Chi-Kuang Hwang
• Format: Others
• Language: en_US
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/44464784177068845038

碩士 === 中華大學 === 電機工程學系碩士班 === 95 === This thesis concentrates on the sampling time offset (STO) estimation scheme that employs the correlation of the short symbols specified in the preamble field of 802.11a WLAN frame structure. A simple detection strategy that uses four values and two values of the correlation is introduced. We also present a new algorithm by which the STO can be adaptively adjusted symbol by symbol. The adaptive algorithm inherits robustness against the variation of roll off rate and can improve the magnitude of the desired four points of correlated sequence. To take advantage of the adaptive algorithm, a linear curve fitting that focuses on the vicinity of zero STO is proposed, and then a state space based on the linear curve fitting can be constructed for a smoothing Kalman filter. Additionally, a threshold that can improve the STO detection is proposed to cooperate with the filter. The mean of estimation error resulted from this new scheme has been simulated for AWGN channels and multipath environments. Performance characteristics of the STO estimation using the non-adaptive method and the adaptive mode are evaluated via numerical examples. Our results show that the proposed method can effectively reduce STO.