Modeling and Prediction of Radio Channels for Orthogonal Frequency Division Multiplexing
(2005) In MSc ThesesDepartment of Automatic Control
- Abstract
- In a wireless communication system, the channel is ever changing due to multipath properties such as excess delays and constructive and destructive interference. In a step to optimize system performance, it would be beneficial to have access to channel state information of both present subchannels' future development and subchannels outside the currently used frequency interval. In this thesis, radio channels for Orthogonal Frequency Division Multiplexing (OFDM) is concerned. Channels are modelled and a prediction algorithm appli-cable in time or frequency direction of the frequency response is developed. The algorithm derives an autoregressive (AR) model of the channel and exploits the close connection between prediction of a time... (More)
- In a wireless communication system, the channel is ever changing due to multipath properties such as excess delays and constructive and destructive interference. In a step to optimize system performance, it would be beneficial to have access to channel state information of both present subchannels' future development and subchannels outside the currently used frequency interval. In this thesis, radio channels for Orthogonal Frequency Division Multiplexing (OFDM) is concerned. Channels are modelled and a prediction algorithm appli-cable in time or frequency direction of the frequency response is developed. The algorithm derives an autoregressive (AR) model of the channel and exploits the close connection between prediction of a time discrete signal and an autoregres-sive process. The great over sampling at the receiver, compared to the maximum Doppler shift or maximum excess delay of the channel, yields it feasible to derive the model with decreased sampling rate and thus only use pilot symbols. This rate reduction gives longer accurate predictions but also an interpolation issue that is solved by Wiener filters. The mean performance of the approach is evaluated by calculating the root mean square error (RMSE) of predictions as functions of various channel and algorithm parameters. Simulations show that the algorithm is able to predict several coherence times or coherence bandwidths of a channel. The algorithm is also extended to use several OFDM symbols and an averaging finite impulse response (FIR) filter, yielding a considerable increase in prediction performance. It is also appraised as a tool for adaptive symbol mapping in a frequency interval not currently used. Results show an opportunity of substantially higher data throughput than for a mapping scheme based on the mean SNR of the channel. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/8847868
- author
- Larsson, Per-Ola
- supervisor
- organization
- year
- 2005
- type
- H3 - Professional qualifications (4 Years - )
- subject
- publication/series
- MSc Theses
- report number
- TFRT-5758
- ISSN
- 0280-5316
- language
- English
- id
- 8847868
- date added to LUP
- 2016-03-18 10:38:51
- date last changed
- 2016-03-18 10:38:51
@misc{8847868,
abstract = {{In a wireless communication system, the channel is ever changing due to multipath properties such as excess delays and constructive and destructive interference. In a step to optimize system performance, it would be beneficial to have access to channel state information of both present subchannels' future development and subchannels outside the currently used frequency interval. In this thesis, radio channels for Orthogonal Frequency Division Multiplexing (OFDM) is concerned. Channels are modelled and a prediction algorithm appli-cable in time or frequency direction of the frequency response is developed. The algorithm derives an autoregressive (AR) model of the channel and exploits the close connection between prediction of a time discrete signal and an autoregres-sive process. The great over sampling at the receiver, compared to the maximum Doppler shift or maximum excess delay of the channel, yields it feasible to derive the model with decreased sampling rate and thus only use pilot symbols. This rate reduction gives longer accurate predictions but also an interpolation issue that is solved by Wiener filters. The mean performance of the approach is evaluated by calculating the root mean square error (RMSE) of predictions as functions of various channel and algorithm parameters. Simulations show that the algorithm is able to predict several coherence times or coherence bandwidths of a channel. The algorithm is also extended to use several OFDM symbols and an averaging finite impulse response (FIR) filter, yielding a considerable increase in prediction performance. It is also appraised as a tool for adaptive symbol mapping in a frequency interval not currently used. Results show an opportunity of substantially higher data throughput than for a mapping scheme based on the mean SNR of the channel.}},
author = {{Larsson, Per-Ola}},
issn = {{0280-5316}},
language = {{eng}},
note = {{Student Paper}},
series = {{MSc Theses}},
title = {{Modeling and Prediction of Radio Channels for Orthogonal Frequency Division Multiplexing}},
year = {{2005}},
}