Adaptive self-linearization
View Patent ↗A signal processing method includes receiving an unknown signal that includes a distorted component and an undistorted component, and performing self-linearization based at least in part on the unknown signal to obtain an output signal that is substantially undistorted, wherein performing self-linearization includes adaptively generating a replica distortion signal that is substantially similar to the distorted component, and subtracting the replica distortion signal from the unknown signal to obtain the output signal.
1. A method of signal processing, comprising:
receiving, at an input interface, an unknown signal that includes a distorted component and an undistorted component; and
performing self-linearization based at least in part on the unknown signal and without requiring a training signal with known characteristics to obtain an output signal that is substantially undistorted, wherein performing self-linearization includes:
adaptively generating a replica distortion signal based at least in part on the unknown signal, the replica distortion signal being substantially similar to the distorted component; and
subtracting the replica distortion signal from the unknown signal to obtain the output signal.
2. A method as recited in claim 1 , wherein the distorted component and the undistorted component have a nonlinear relationship.
3. A method as recited in claim 1 , wherein the self-linearization is independent of the modulation or encoding scheme of the unknown signal.
4. A method as recited in claim 1 , wherein performing self-linearization further includes separating from the distorted signal a reference component and a target component.
5. A method as recited in claim 4 , wherein the reference component includes the undistorted component and a harmonic of the undistorted component.
6. A method as recited in claim 4 , wherein the reference component occupies a first frequency band, and the target component occupies a second frequency band separate from the first frequency band.
7. A method as recited in claim 4 , wherein the target component occupies a first frequency channel, reference component includes a radio frequency (RF) signal in a second frequency channel, and the first frequency channel is separate from the second frequency channel.
8. A method as recited in claim 4 , wherein the undistorted component occupies a first plurality of frequency bands, the reference component occupies a second plurality of frequency bands, and the first plurality of frequency bands are separate from the second plurality of frequency bands.
9. A method as recited in claim 1 , wherein performing self-linearization includes applying the unknown signal to a persistence filter.
10. A method as recited in claim 1 , wherein performing self-linearization includes applying the unknown signal to a persistence filter having N-number of filter taps, and adjusting the value of N.
11. A method as recited in claim 1 , wherein performing self-linearization includes applying the unknown signal to a persistence filter having an adaptation step size μ, and adjusting the value of μ.
12. A method as recited in claim 1 , wherein performing self-linearization based on the unknown signal further includes:
generating a linearization output that approximates the distorted component, the linearization output has a delay of k samples;
delaying the unknown signal by k samples; and
combining the linearization output and the delayed unknown signal.
13. A method as recited in claim 1 , wherein performing self-linearization includes:
scaling a plurality of samples of the unknown signal using a plurality of coefficients;
aggregating the scaled results to produce an aggregate;
generating an error that is the difference between the aggregate and a sample of the unknown signal; and
feeding back the error to adapt the plurality of coefficients.
14. A method as recited in claim 1 , wherein performing self-linearization further includes adapting a digital signal processor (DSP), the adaptation being based at least in part on the reference component.
15. A method as recited in claim 1 , wherein the self-linearization is performed in real time.
16. A method as recited in claim 1 , wherein performing self-linearization includes adapting a DSP to obtain a filter transfer function that approximates a system distortion transfer function.
17. A method as recited in claim 16 , wherein the DSP is a nonlinear DSP.
18. A method as recited in claim 16 , wherein the DSP is a first DSP; and the
method further comprising configuring a duplicate DSP to have a duplicate transfer function that is substantially similar to the filter transfer function obtained by the first DSP.
19. A method as recited in claim 18 , wherein the duplicate DSP is configured using coefficients from the first DSP.
20. A system comprising:
an input terminal configured to receive an unknown signal that includes an undistorted component and a distorted component; and
an adaptive self-linearization module coupled to the input terminal, configured to perform self-linearization based on the unknown signal and without requiring a training signal with known characteristics to obtain an output signal that is substantially undistorted, wherein:
the adaptive self-linearization module is configured to adaptively generate a replica distortion signal based at least in part on the unknown signal, the replica distortion signal being substantially similar to the distorted component; and
the adaptive self-linearization module includes a combiner configured to subtract the replica distortion signal from the unknown signal to obtain the output signal.
21. A system as recited in claim 20 , wherein the adaptive self-linearization module includes an adaptive linearization module and a delay element.
22. A system as recited in claim 20 , further comprising a separation block configured to separate from the distorted signal a reference component and a target component.
23. A system as recited in claim 22 , wherein the separation block includes a persistence filter configured to enhance the undistorted component in the unknown signal.
24. A system as recited in claim 22 , wherein the adaptive self-linearization module further includes an adaptive DSP coupled to the separation block.
25. A system as recited in claim 24 , further comprising a duplicate DSP coupled to the adaptive DSP, configured to receive configuration parameters from the adaptive DSP.
26. A system as recited in claim 22 , wherein the separation block includes a reference component band-specific filter configured to select the reference component from the unknown signal.
27. A system as recited in claim 22 , wherein the separation block includes a target component band-specific filter configured to select the target component from the unknown signal.
28. A computer program product for signal processing, the computer program product being embodied in a computer readable medium and comprising computer instructions for:
receiving an unknown signal that includes an undistorted component and a distorted component; and
performing self-linearization based at least in part on the unknown signal and without requiring a training signal with known characteristics to obtain an output signal that is substantially undistorted, wherein performing self-linearization includes:
adaptively generating a replica distortion signal based at least in part on the unknown signal, the replica distortion signal being substantially similar to the distorted component; and
subtracting the replica distortion signal from the unknown signal to obtain the output signal.