US-12621192-B2 - Soft-demodulation based waveform repair for signals undergoing spectral suppression

US12621192B2US 12621192 B2US12621192 B2US 12621192B2US-12621192-B2

Abstract

Systems and methods for waveform repair. The methods comprise: receiving a signal of interest having a spectral distortion, the signal of interest having a sequence of symbols; generating an estimate for each symbol of a sequence of symbols in the signal of interest; and applying reconstruction filtering to the estimates to correct the spectral distortion.

Inventors

Christian Schlegel
L. Andrew Gibson, Jr.
Evan D. Poff
Edwin R. Twitchell
David R. Keller
Lance Lindsay

Assignees

L3HARRIS TECHNOLOGIES, INC.

Dates

Publication Date: 20260505
Application Date: 20240613

Claims (15)

1 . A method for waveform repair, comprising: receiving a signal of interest having a spectral distortion, the signal of interest having a sequence of symbols; performing band-rejection filtering to remove interference from the signal of interest; generating an estimate for each symbol of a sequence of symbols in the signal of interest; and applying reconstruction filtering to the estimates to correct the spectral distortion, wherein the reconstruction filtering is applied to the estimates to demodulate a portion of the signal of interest which was removed by the band-rejection filtering.
2 . The method according to claim 1 , wherein the estimate comprises a maximum a posteriori probability estimate, or an appropriate approximation thereof, for each symbol of a sequence of symbols that is generated using an output signal of the band-rejection filtering.
3 . The method according to claim 1 , further comprising: combining the band-rejection filtered portion of the signal of interest with an output signal of the reconstruction filtering; and performing upconversion and bandpass filtering using a resulting signal from said combining to reconstitute the signal of interest without interference.
4 . The method according to claim 1 , wherein the band-rejection filtering comprises notch filtering.
5 . The method according to claim 4 , wherein the notch filtering is shaped to minimize the energy of intersymbol interference.
6 . The method according to claim 1 , wherein the estimate for each symbol is generated using at least one past symbol that comes before the symbol in the sequence of symbols or at least one future symbol that comes after the symbol in the sequence of symbols.
7 . The method according to claim 1 , wherein the estimate for each symbol is generated using a BCJR algorithm.
8 . The method according to claim 1 , wherein the reconstruction filtering comprises bandpass filtering.
9 . A circuit, comprising: a soft symbol demodulator configured to generate an estimate for each symbol of a sequence of symbols in a received signal of interest having a spectral distortion; a frequency excision module configured to perform band-rejection filtering to remove interference from the signal of interest; and a reconstruction filter configured to apply reconstruction filtering to the estimates to correct the spectral distortion; wherein the reconstruction filtering is applied to the estimates to demodulate a portion of the signal of interest which was removed by the band-rejection filtering.
10 . The circuit according to claim 9 , wherein the estimate comprises a maximum a posteriori probability estimate, or an appropriate approximation thereof, for each symbol of a sequence of symbols that is generated using an output signal of the band-rejection filtering.
11 . The circuit according to claim 9 , further comprising: a combiner configured to combine the demodulated portion of the signal of interest with the output signal of the band-rejection filtering; and a filter configured to perform upconversion and bandpass filtering using an output signal of said combiner to reconstitute the signal of interest without interference.
12 . The circuit according to claim 9 , wherein the band-rejection filtering comprises notch filtering.
13 . The circuit according to claim 12 , wherein the notch filtering is shaped to minimize the energy of inter-symbol interference.
14 . A non-transitory computer-readable medium that stores instructions that, when executed by at least one computing device, will cause the at least one computing device to perform operations comprising: receiving a signal of interest having a spectral distortion, the signal of interest having a sequence of symbols; performing band-rejection filtering to remove interference from the signal of interest; generating an estimate for each symbol of a sequence of symbols in the signal of interest; and applying reconstruction filtering to the estimates to correct the spectral distortion, wherein the reconstruction filtering is applied to the estimates to demodulate a portion of the signal of interest which was removed by the band-rejection filtering.
15 . The non-transitory computer-readable medium according to claim 14 , wherein the estimate comprises a maximum a posteriori probability estimate, or an appropriate approximation thereof, for each symbol of a sequence of symbols that is generated using an output signal of the band-rejection filtering.

Description

BACKGROUND Description of the Related Art A communication signal can be negatively impacted by interference. A solution to mitigate and overcome this type of interference exist with notching the signal with products such as a frequency notching filter (FNF). For wide-band interference signals, however, current methods for such notching may remove too much of a communications signal itself and render it undecodable. SUMMARY The present disclosure concerns implementing systems and methods for waveform repair. The methods comprise: receiving a signal of interest having a spectral distortion, the signal of interest having a sequence of symbols; generating an estimate for each symbol of a sequence of symbols in the signal of interest; and applying reconstruction filtering to the estimates to correct the spectral distortion. The present disclosure also concerns a circuit, comprising: an approximate soft symbol demodulator configured to generate an estimate for each symbol of a sequence of symbols in a received signal of interest having a spectral distortion; and a reconstruction filter configured to apply reconstruction filtering to the estimates to correct the spectral distortion. The present disclosure further concerns a non-transitory computer-readable medium that stores instructions that, when executed by at least one computing device, will cause the at least one computing device to perform operations comprising: receiving a signal of interest having a spectral distortion, the signal of interest having a sequence of symbols; generating an estimate for each symbol of a sequence of symbols in the signal of interest; and applying reconstruction filtering to the estimates to correct the spectral distortion. BRIEF DESCRIPTION OF THE DRAWINGS The present solution will be described with reference to the following drawing figures, in which like numerals represent like items throughout the figures. FIG. 1 provides graphs that are useful for understanding notch filtering and signal reconstruction. FIG. 2 provides an illustration that is useful for understanding notch filtering. FIG. 3 provides an illustration of a system implementing the present solution. FIG. 4 provides a block diagram of a communication device. FIG. 5 provides a block diagram of a receiver. FIG. 6 provides a block diagram of a signal processing circuit implementing the present solution. FIG. 7 provides a graph showing absolute values of taps for a rolloff value β of 0.2. FIG. 8 provides a first graph showing signal constellation at the SOI output and SOI spectrum if no action is taken, along with a second graph showing the SOI output using symbol estimates from a BCJR algorithm to rebuild the SOI spectrum. FIG. 9 provides an illustration of a discrete time model of the inter-symbol interference (ISI) effects resulting from frequency notch filtering (FNF), covering five relevant taps, along with an example trellis section that would act as a computational blueprint for the BCJR algorithm. FIG. 10 provides a graph showing that soft-demodulation-based waveform repair, such as implemented with a BCJR-based repair filter, improves the bit error performance in the presence of interference. FIG. 11 provide a flow diagram of an illustrative method for processing a received signal. FIG. 12 provides an illustration of a computer system. DETAILED DESCRIPTION As noted above, a communication signal can be negatively impacted by interference. A solution to mitigate and overcome this type of interference exists with notching the signal with products such as an frequency notching filter (FNF) module. For wider interference signals, current methods that notch the signal remove too much of the communication signal to then be demodulated error free. The effectiveness of notching can be substantially mitigated by deriving soft symbols of the signal of interest via a soft symbol estimator, such as a decoder implementing the Bahl, Cocke, Jelinek, and Raviv (BCJR) algorithm to optimally estimate symbols in signals distorted by such notch filtering. After interference (e.g., signal 104 in graph 100 of FIG. 1) is removed, the SOI (e.g., signal 102 in graph 100 of FIG. 1) has too much spectral signal content removed by the notch filtering for the signal to be received error-free. The BCJR decoder provides a means to obtain high-accuracy estimated probabilities for each symbol of the SOI despite signal degradation due to the notched-out portion thereof and allows for the rebuilding of the spectral component that were removed by notch filtering. Consequently, the SOI may be received error free (e.g., as shown in graph 150 of FIG. 1) without further special equipment. An FNF module 200 of FIG. 2 is configured to remove interference from a signal via notching a portion of the signal that contains interference at higher power than the SOI. The FNF module 200 performs operations to notch interference out of the spectrum. If the notch is small enough, the demodulator can lock onto t