Performance Analysis of Transceiver Impairments in OFDM Systems
Date Issued
2014
Date
2014
Author(s)
Kuo, Chi
Abstract
The thesis presents an extended best linear unbiased estimator (EBLUE) based on a periodic training sequence, to estimate a frequency offset in orthogonal frequency division multiplexing (OFDM) systems. The structure of the EBLUE is general and flexible so that it adapts to different complexity constraints. The features of generality and flexibility are attractive for practical implementation. Performance analysis and design strategy of the EBLUE are provided to realize the best tradeoff between performance and complexity. Moreover, closed-form results of both weight and performance make the EBLUE even more attractive for practical implementation.
Time-invariant maximum likelihood channel estimator (TI-MLCE) originally designed for time-invariant channels is extended to OFDM systems in time-varying channels. The analysis of the extended TI-MLCE is performed under a general setup, and results degenerate to special cases. Simulation is then used to verify the accuracy of analysis, and the performance of the extended TI-MLCE is compared with that of other proposals. It is demonstrated that the extended TI-MLCE has better performance, lower complexity and fewer design parameters. Thus, the extended TI-MLCE is practical for implementation.
The effect of transceiver impairments (including frequency offset, phase noise and doubly-selective channel) is a key factor in determining the performance of an OFDM system. These impairments are well known and have been investigated separately in the past. However, these impairments usually arise concurrently and should be jointly considered from the perspectives of both receiver design and system evaluation. Impacts of these impairments on the OFDM system are jointly analyzed in the research, and the analysis result degenerates to the special cases where only a subset of the impairments is present. A mitigation method aided by segment-by-segment time-domain interpolation (STI) is then proposed following the analysis. The STI is general, and weights for realizing the STI can be specified according to the interpolation method and system requirements.
Signal-to-interference-plus-noise ratio (SINR) is treated as a performance measure for evaluating the receiver using the TI-MLCE with the STI, and is derived under a general setting. Consistency between our analysis and simulations verifies the validity of our derivation. Significant improvement over the receiver using only the TI-MLCE is demonstrated in the simulation results.
Subjects
正交分頻多工
最大似然估計子
通道估測
頻率偏移
相位雜訊
時變通道
傳收機減損
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-103-D94942012-1.pdf
Size
23.32 KB
Format
Adobe PDF
Checksum
(MD5):2334461a67a19467466fedeef81c5f46