The Effect of Pilot Allocation for Comb-type Pliot-aided Channel Estimation in OFDM Systems / 導頻配置在OFDM系統中對梳狀導頻輔助通道估測的影響

博士 / 國立中正大學 / 電機工程所 / 96 / The purpose of this dissertation is to investigate the effect of pilot allocation on some interpolation methods for comb-type pilot-aided channel estimation in orthogonal frequency division multiplexing (OFDM) systems. The present results showed that the proposed analytic interpolator, which was derived from the relation between the unknown channel impulse response and known pilot signals under the assumption that the number of pilots is not less than the number of channel paths, outperforms the cubic-spline interpolation and linear interpolation methods under both equidistant and non-equidistant pilot allocations for a given pilot density. Moreover, equidistant pilot patterns for a given pilot density give better performance than those with non-equidistance. However, if the number of subcarriers divided by the number of pilots is not an integer, then an equidistant pilot pattern is impossible. In order to optimize the performance of the proposed analytic interpolator, an optimal pilot pattern generation algorithm is developed from the minimum mean squared error (MMSE) criterion. This algorithm generates all the pilot patterns for a given pilot density with MMSE for the proposed analytic interpolator. In addition, two PSO-based analytic interpolation schemes are presented and examined for different pilot patterns. One is the unknown data signals search method, which estimates channel impulse response with both searched data signals and pilot signals. The other is a channel impulse response search method, which directly searches for the optimal channel impulse response through the criterion of minimizing the mean squared error between the channel frequency response derived from the searched for channel impulse response and that from the least squared method on pilot subcarriers. Simulation results showed that the latter is superior to the former while the iteration number of particle swarm optimization (PSO) in both schemes is larger.

Identiferoai:union.ndltd.org:TW/096CCU05442180
Date January 2008
CreatorsSShu-Hong Lee, 李樹鴻
Contributorsnone, 溫志宏
Source SetsNational Digital Library of Theses and Dissertations in Taiwan
Languageen_US
Detected LanguageEnglish
Type學位論文 ; thesis
Format69

Page generated in 0.013 seconds