Noise Filtering Inverse Transformation

• US 20100322364A1
• Filed: 03/12/2008
• Published: 12/23/2010
• Est. Priority Date: 03/12/2007
• Status: Active Grant

• Current Assignee: John W. Bogdan
• Original Assignee: John W. Bogdan
• Inventors: Bogdan, John W.
• Granted Patent: US 8,284,877 B2
• Time in Patent Office: Days
• US Class Current: 375/350
• CPC Class Codes: H04L 27/2649 Demodulators

Abstract

A method, system and apparatus for noise filtering inverse transformation (NFIT), recovering phases and amplitudes of singular cycles of data carrying tones or sub-bands from a composite signal such as OFDM, is presented herein. Such NFIT comprises adaptive inverse transformation of non-linear channel transform function and instant accommodation of time variant quickly changing characteristics of transmission channel caused by interferences including line loads, cross-talk or predictable noise.

Claims

1. A method for noise filtering inverse transformation (NFIT) comprising a recovery of an original signal, received from a transmission channel which transforms the original signal into a received signal space affected by a transmission channel distortion or interference such as that of a transmission link noise or adjacent channel or frequency band, by reversing this transformation of original signal in order to recover the original signal free of such channel distortion or such interference;

the NFIT method comprising the steps of;

capturing an over-sampled waveform of said received signal;

pre-filtering or analyzing such captured waveform in order to detect which predefined set of waveform shapes corresponds to this pre-filtered or analyzed waveform, wherein such predefined set of shapes is characterized by its frame of reference;

the recovery of said original signal by applying an inverse transformation of said original signal transformation to this frame of reference;

wherein such original signal transformation and its inverse transformation can be derived using theoretical models or results of training session or results of adaptive optimization process.

2. A method for noise filtering inverse transformation (NFIT) comprising a recovery of an original signal, received from a transmission channel which transforms the original signal into a received signal space affected by a transmission channel distortion or interference such as that of a transmission link noise or adjacent channel or frequency band, by reversing this transformation of original signal in order to recover the original signal free of such channel distortion or such interference, the NFIT method comprising the steps of:

capturing an interval of an over-sampled waveform of said received signal;

comparing this captured interval with a mask used as frame of reference characterizing a set of predefined shapes, in order to verify which set of predefined shapes such captured interval corresponds to,

wherein such comparison includes

producing a proximity estimate such as correlation integral or deviation integral between samples belonging to the captured interval and their counterparts belonging to the mask,

and using such proximity estimate for verifying if the set of shapes characterized by the mask used corresponds to a shape of the captured interval;

applying the inverse transformation of the known transmission channel transformation to the mask characterizing such corresponding set of shapes, in order to recover the original signal.

3. (canceled)

4. A method for data recovery from a composite frame (DRCR) comprising discrete multiple tones (DMT) or multiple sub-bands (MSB) wherein such data recovery method comprises combination of frequency domain and time domain signal processing methods, the DRCR method comprising the steps of:

frequency domain filtering of a composite signal carrying such composite frame wherein frequency filters produce discrete tones or sub-bands maintaining known phase relation to said DMT or MSB frame i.e. said frequency filters keep their outputs in phase with the composite frame;

said time domain time signal processing of such discrete tones or sub-bands performed in phase with the composite frame, measures amplitudes and/or phases of singular half-cycles or cycles of discrete tones or sub-bands;

such measured amplitudes and/or phases are inversely transformed by reversing their distortions introduced by a transmission channel of said composite signal, including distortions caused by previous processing stages;

such inversely transformed amplitudes and phases are used to recover data symbols originally encoded in the transmitted signal.

5. A DRCR as claimed in claim 4, wherein the DRCR method comprises the step of:

selecting the most probable symbol by using statistical methods for processing a set of data symbols recovered from amplitudes and/or phases of a particular tone or sub-band.

6. A method for data recovery from a composite signal (DRCS) comprising discrete multiple tones (DMT) or multiple sub-bands (MSB) wherein a combination of frequency domain and time domain signal processing methods is utilized, wherein amplitudes and/or phases of signals surrounding a particular data carrying tone or sub-band are sensed and distortions of said particular tone or sub-band are reversed;

the DRCS method comprising the steps of;

frequency domain filtering of said composite signal producing discrete tones or sub-bands maintaining known phase relation to said DMT or MSB frame;

said time domain signal processing of such discrete tones or sub-bands measures amplitudes and/or phases of singular half-cycles or cycles of discrete tones or sub-bands;

detection of amplitudes and/or phases of noise and/or other signals surrounding such particular tone or sub-band, performed in frequency domain and/or in time domain;

using such detected amplitudes and/or phases for deriving estimates of distortions introduced by the surrounding noise and/or other signals to such particular tone or sub-band;

using such distortions estimates for performing reverse transformation of said detected amplitudes and/or phases of such singular half-cycles or cycles of the particular tone or sub-band into the amplitudes and/or phases corresponding to data transmitted originally.

such reversely transformed amplitudes and phases are used to recover data symbols originally encoded in the transmitted signal.

7. (canceled)

8. (canceled)

9. (canceled)

10. (canceled)

11. (canceled)

12. A method of inverse transformation of a band filtered signal (ITBFS) for a recovery of an original signal by applying an inverse transformation to a pre-filtered signal, wherein such inverse transformation includes compensation of transmission channel change caused by an instant interference;

wherein the ITBFS method comprises the steps of;

using a band-pass filter for producing a pre-filtered signal from a captured waveform of a received signal;

producing real time evaluation of such instant interference which remains in phase with the pre-filtered signal by maintaining known phase alignment to such signal;

using such real time evaluation for selection of frames of reference accommodating such instant interference,

using such selected frames of reference for detecting a set of shapes corresponding to the pre-filtered signal affected by the instant interference,

wherein such detection includes comparison of such frames of reference, characterizing sets of predefined signal shapes, with a shape of said affected signal;

said recovery of the original signal by performing the inverse transformation of the frame of reference which detected such set of shapes corresponding to the signal affected by interference.

13. An ITBFS method as claimed in claim 12, wherein said band filtered signal is a tone or sub-band and said received signal is a composite OFDM signal transmitted over a wireline or wireless link and said original signal is the original tone or sub-band incorporated into the composite OFDM signal.

14. A method for reversal of non-linearity (RNL) of amplifier gain or signal attenuation by applying a polynomial transformation to a non-linear signal, wherein the RNL method comprises the steps of:

identification of dependency between amplitude of an original signal and said non-linear signal;

using such dependency for defining polynomial approximation thresholds and their slope coefficients and their exponents;

calculating an exponential component for every said threshold exceeded by a sample of said non-linear signal,

wherein such exponential component for the maximum exceeded threshold, is calculated by rising a difference, between the non-linear signal sample and the maximum threshold, to a power defined by said exponent for the maximum threshold;

wherein such exponential component for every other exceeded threshold, is calculated by rising a difference, between the next exceeded threshold and such other exceeded threshold, to a power defined by said exponent for the other exceeded threshold;

calculating an approximation component for every such approximation threshold exceeded by said non-linear signal sample, by multiplying such exponential component by its slope coefficient;

addition of such approximation components, calculated for approximation thresholds exceeded by or equal to the non-linear signal sample;

wherein by such addition of the approximation components, calculated for the approximation thresholds exceeded by or equal to the non-linear signal sample, said non-linearity is reversed.

15. A method for inverse normalization (IN) of OFDM tone or sub-band, comprising reversal of frequency dependent distortion of said tone or sub-band, by applying normalizing coefficients adjusted to a frequency of this tone or sub-band in order to equalize amplitude and phase distortions introduced to the tone or sub-band by an OFDM transmission channel and a signal processing applied;

such IN method comprises the steps of;

identification of the frequency related amplitude or phase distortion of such tone or sub-band by sampling and analyzing of a waveform performed by a programmable control unit (PCU);

calculating such normalizing coefficients for the tone or sub-band, performed by said PCU;

using such normalizing coefficients for equalizing such frequency related distortions, performed by a real-time processing unit (RTP).

16. (canceled)

17. (canceled)