Methods and apparatus for weighted equalization
View Patent ↗Systems and techniques for equalization of a response of a receiver. In one embodiment, a response of a device to a known signal is measured to generate a measured signal. A processor is operated to apply an initial weighting to the measured signal and the known signal. The processor is operated to iteratively perform the operations of constructing an equalizer based on the weighted signals, testing the performance of the equalizer on the measured signal, and adjusting weighting applied to the measured signal and the known signal, until the performance of the equalizer on the measured signal meets a predetermined criterion.
1. A method comprising:
retrieving from a computer readable memory a test signal;
measuring a response of a device to the test signal to generate a measured signal;
operating a processor to apply an initial weighting to at least a given frequency component of the measured signal and at least said given frequency component of the test signal; and
operating a processor to iteratively perform the operations of constructing an equalizer based on the weighted signals, testing the performance of the equalizer on the measured signal, and adjusting weighting applied to the measured signal and the test signal,
wherein:
the test signal comprises a plurality of frequency sub-bands within a frequency band,
the measured signal comprises values representing a measurement on each of the plurality of frequency sub-bands,
operating the processor to apply the initial weighting to the measured signal and the test signal comprises constructing a weighting vector comprising a plurality of weighting values to be applied to each of the plurality of sub-bands of the test signal and the measured signal; and
operating the processor to iteratively perform the operations of constructing an equalizer based on the weighted signals comprises, in each successive iteration after the initial weighting is applied, adjusting the weighting comprises increasing a per-sub-band weighting value applied to each sub-band that is subject to the next successive iteration of the operations of constructing the equalizer.
2. The method of claim 1 , wherein constructing the equalizer based on the weighted signals comprises constructing the equalizer to map the weighted measured signal to the weighted test signal.
3. The method of claim 1 , further comprising constructing a constraint waveform and adding the constraint waveform to the weighted measured signal.
4. The method of claim 1 , wherein the test signal is a digital representation of a radio frequency signal and wherein the measured signal is a digitized output of the response of the device.
5. The method of claim 1 , wherein the test signal is a cyclic signal exhibiting equal energy at all in-band frequencies and no energy at out-of-band frequencies.
6. The method according to claim 1 , wherein for each sub-band subject to a given successive iteration after the initial weighting, the per-sub-band weighting value is proportional to a ratio of an error value to a maximum error value, where the error value is used in the testing of the performance of the equalizer that was performed in the immediately preceding iteration.
7. An apparatus comprising:
a receiver configured to produce a response to a test signal;
a measurement device configured to measure the receiver response to produce a measured signal;
data storage configured to store the test signal and the measured signal; and
a processor configured to:
apply an initial weighting to at least a given frequency component of the measured signal and at least said given frequency component of the test signal, wherein operating the processor to apply the initial weighting to the measured signal and the test signal comprises constructing a weighting vector comprising a plurality of weighting values to be applied to each of a plurality of sub-bands of the test signal and the measured signal; and
iteratively perform the operations of constructing an equalizer based on the weighted signals, testing the performance of the equalizer on the measured signal, and adjusting weighting applied to the measured signal and the test signal,
wherein operating the processor to iteratively perform the operations of constructing the equalizer based on the weighted signals further comprises, in each successive iteration after the initial weighting is applied, adjusting the weighting comprises increasing a per-sub-band weighting value applied to each sub-band that is subject to the next successive iteration of the operations of constructing the equalizer.
8. The apparatus of claim 7 , wherein the receiver comprises a filter and wherein the response of the receiver includes artifacts produced by the filter.
9. The apparatus of claim 8 , wherein the filter is an analog baseband filter.
10. The apparatus of claim 8 , wherein the artifacts include a ripple in a wanted frequency band of the measured signal.
11. The apparatus of claim 7 , wherein the test signal includes adjacent channel interference.
12. The apparatus of claim 7 , wherein the processor is configured to test the performance of the equalizer by at least performing a comparison between each sub-band of an equalized measured signal and a corresponding sub-band of a test signal, the equalized measured signal being generated by equalizing the measured signal, the processor being further configured to determine, for each sub-band of the equalized measured signal and the corresponding sub-band of the test signal, an error value based on the comparison, the processor being further configured, for each error value, to determine if the error value meets the predetermined criterion.
13. The apparatus of claim 8 , wherein the receiver comprises an analog to digital converter converting an output of the filter to digital form, wherein the response of the receiver comprises a digital representation of a waveform produced by the analog to digital converter and wherein the analog to digital converter introduces aliasing into the response of the receiver.
14. The apparatus according to claim 7 , wherein for each sub-band subject to a given successive iteration after the initial weighting, the per-sub-band weighting value is proportional to a ratio of an error value to a maximum error value, where the error value is used in the testing of the performance of the equalizer that was performed in the immediately preceding iteration.