Edge MMSE filters
View Patent ↗A method for performing channel estimation of an OFDM channel includes, in part, interpolating pilots for sub-channels positioned within a first range of an OFDM symbol, and estimating frequency response of sub-channels positioned within a second range of the OFDM symbol. The first range is defined by subchannels positioned substantially away from channel edges and the second range is defined by subchannels positioned substantially near channel edges. The method optionally includes transforming the pilots from frequency domain into the time-domain, time-domain windowing to obtain a channel impulse response having a multitude of discrete values, estimating the discrete values within the channel impulse response; and transforming the channel impulse response to the frequency domain.
1. A channel estimator disposed in an Orthogonal Frequency Division Multiplexing (OFDM) receiver, the channel estimator comprising:
an interpolator adapted to interpolate pilots for sub-channels positioned within a first range of an OFDM symbol; and
a filter adapted to estimate frequency response of sub-channels positioned within a second range of the OFDM symbol, said first range defined by subchannels positioned substantially away from channel edges and said second range defined by subchannels falling substantially near channel edges.
2. The channel estimator of claim 1 wherein the interpolator comprises:
an IFFT block for transforming pilots from frequency domain into the time-domain;
a time-domain windowing filter for obtaining a channel impulse response having a plurality of discrete values;
a threshold estimator having a plurality of threshold levels for estimating the discrete values within the channel impulse response; and
an FFT block for transforming the channel impulse response to the frequency domain.
3. The channel estimator of claim 2 wherein the time-domain windowing filter has a rectangular shape with a time base corresponding to an estimated delay spread.
4. The channel estimator of claim 2 wherein the threshold estimator estimates the discrete values within the channel impulse response in accordance with predefined threshold levels.
5. The channel estimator of claim 1 wherein the filter is an MMSE filter adapted to estimate the frequency response of sub-channels positioned within the second range in accordance with an auto-covariance matrix of pilots positioned within the second range and a cross-covariance pilot vector.
6. The channel estimator of claim 5 wherein the MMSE filter coefficients are defined by a product of the auto-covariance matrix of the pilots positioned within the second range and the cross-covariance of the pilots.
7. The channel estimator of claim 1 wherein the frequency response of each sub-channel positioned within the second range of the OFDM symbol is estimated in accordance with a product of the MMSE filter coefficients and the pilots.
8. The channel estimator of claim 2 wherein the threshold estimator removes the estimated discrete values within the channel impulse response when the estimated discrete values are below an optimum threshold level.
9. The channel estimator of claim 8 wherein the optimum threshold level is adaptively adjusted based on a standard deviation of the discrete values.
10. A method for performing channel estimation of an Orthogonal Frequency Division Multiplexing (OFDM) channel, the method comprising:
interpolating pilots for sub-channels positioned within a first range of an OFDM symbol; and
estimating frequency response of sub-channels positioned within a second range of the OFDM symbol using a filter, said first range defined by subchannels positioned substantially away from channel edges and said second range defined by subchannels positioned substantially near channel edges.
11. The method of claim 10 further comprising:
transforming pilots from frequency domain into the time-domain;
time-domain windowing to obtain a channel impulse response having a plurality of discrete values;
estimating the discrete values within the channel impulse response; and
transforming the channel impulse response to the frequency domain.
12. The method of claim 11 wherein the time-domain windowing has a rectangular shape with a time base corresponding to an estimated delay spread.
13. The method of claim 11 further comprising estimating the discrete values within the channel impulse response in accordance with predefined threshold levels.
14. The method of claim 11 wherein the frequency response of sub-channels positioned within the second range is estimated in accordance with an auto-covariance matrix of pilots positioned within the second range and a cross-covariance pilot vector.
15. The method of claim 14 wherein the frequency response of sub-channels positioned within the second range is estimated in accordance with a product of the auto-covariance matrix of the pilots positioned within the second range and the cross-covariance pilot vector.
16. The method of claim 15 wherein the frequency response of sub-channels positioned within the second range is estimated in accordance with a product of filter coefficients and the pilots.
17. The method of claim 11 further comprising removing the discrete values within the channel impulse response when the estimated discrete values are below an optimum threshold level.
18. The method of claim 17 wherein the optimum threshold level is adaptively adjusted based on a standard deviation of the discrete values.