Title: High-performance channel estimation in an adaptive OFDM-system using a Kalman filter/predictor

Speaker: Gunnar Hagsköld

Time and Place:

Wednesday February 12th, at 13:15
Library, floor 2 at Magistern, Dag Hammarskjölds väg 31, Uppsala

Abstract:

In order to use adaptive modulation over a radio channel, there is a need for efficient and accurate channel estimation. This investigation evaluates a Kalman filter/predictor estimation tool, using an AR(4) model of the fading dynamics, working over an adaptive OFDM-system with 500 subcarriers on a bandwidth of 5MHz.

Training symbols and detected 4QAM control symbols, located on a subset of the subcarriers, are used as regressors in the estimation. The output from the Kalman estimator is interpolated in time and frequency, to obtain estimates of the channel between the training and control symbol locations.

The topics to be evaluated are:

• Deriving an AR(4) model for the fading dynamics.
• Adjustment of the Kalman parameters based on recieved data.
• The interpolation accuracy.
• Correction of erroneous detected control symbols using estimated channel phase differences.