Systems and methods for efficiently-transformed digital self-interference cancellation
An efficiently-transformed digital self-interference canceller, preferably including an FD transformer, a TD transformer, a channel estimator, a composer, and a controller. The canceller can optionally include a channel memory, a predictor, and/or an extender. A method for digital self-interference cancellation, preferably including receiving inputs, transforming the inputs, generating outputs based on the transformed inputs, transforming the outputs, and/or generating a cancellation signal based on the outputs.
1. A system for digital self-interference cancellation, the system comprising:
a first transformer that transforms a digital transmit signal and a digital residue signal of a communication system from a time domain to a frequency domain;
a channel estimator, communicatively coupled to the first transformer, that generates a self-interference channel estimate in the frequency domain based on the digital transmit signal and the digital residue signal;
a second transformer that transforms the self-interference channel estimate from the frequency domain to the time domain, the second transformer comprising:
a decimating filter that, in the frequency domain, downsamples the self-interference channel estimate to generate a low-dimension channel estimate;
an inverse fast Fourier transform (IFFT) processor that transforms the low-dimension channel estimate from the frequency domain to the time domain; and
a post-compensator that, in the time domain, generates a compensated channel estimate based on the low-dimension channel estimate; and
a composer, communicatively coupled to the post-compensator, that, in the time domain, generates a digital self-interference cancellation signal based on the digital transmit signal and the compensated channel estimate;
wherein the digital self-interference cancellation signal is combined with a receive signal of the communication system to form the digital residue signal.
2. The system of claim 1 , wherein:
the decimating filter defines a window function that is substantially non-flat in the time domain, wherein the decimating filter distorts the low-dimension channel estimate based on the window function; and
the post-compensator generates the compensated channel estimate based further on the window function.
3. The system of claim 2 , wherein the window function is associated with a discrete prolate spheroidal function.
4. The system of claim 2 , wherein generating the compensated channel estimate comprises dividing the low-dimension channel estimate by the window function.
5. The system of claim 1 , wherein the decimating filter is a finite impulse response filter.
6. The system of claim 1 , wherein:
the IFFT processor implements an IFFT butterfly structure defining a plurality of stages, wherein a stage of the plurality defines a radix; and
the decimating filter downsamples the self-interference channel estimate by a downsampling factor greater than or equal to the radix.
7. The system of claim 6 , wherein the downsampling factor is greater than or equal to a square of the radix.
8. The system of claim 7 , wherein the radix is greater than 3.
9. The system of claim 1 , wherein the decimating filter downsamples the self-interference channel estimate by a downsampling factor greater than or equal to 10.
10. The system of claim 1 , further comprising:
a third transformer, coupled to the second transformer, comprising:
a fast Fourier transform (FFT) processor that transforms the compensated channel estimate into a low-dimension frequency-domain self-interference channel estimate; and
an upsampling filter that, in the frequency domain, upsamples the low-dimension frequency-domain self-interference channel estimate to generate an upsampled channel estimate; and
a channel memory, coupled to the upsampling filter, that stores the upsampled channel estimate;
wherein the system modifies, in the frequency domain, the frequency-domain self-interference channel estimate based on the upsampled channel estimate; wherein the second transformer transforms the frequency-domain self-interference channel estimate after modification based on the upsampled channel estimate.
11. The system of claim 10 , wherein:
the upsampling filter defines a first window function that is substantially non-flat in the time domain, wherein the upsampling filter distorts the upsampled channel estimate based on the first window function;
the third transformer further comprises a pre-compensator that further compensates the compensated channel estimate based on the first window function; and
the FFT processor transforms the compensated channel estimate into the low-dimension frequency-domain self-interference channel estimate after compensation by the pre-compensator.
12. The system of claim 11 , wherein:
the decimating filter defines a second window function that is substantially non-flat in the time domain, wherein the decimating filter distorts the low-dimension channel estimate based on the second window function; and
the post-compensator generates the compensated channel estimate based further on the second window function.
13. The system of claim 12 , wherein the first window function is substantially equal to the second window function.
14. The system of claim 11 , wherein the first window function is associated with a discrete prolate spheroidal function.
15. The system of claim 11 , wherein further compensating the compensated channel estimate comprises dividing the compensated channel estimate by the first window function.
16. The system of claim 11 , wherein the upsampling filter is a finite impulse response filter.
17. The system of claim 10 , wherein:
the FFT processor implements an FFT butterfly structure defining a plurality of stages, wherein a stage of the plurality defines a radix; and
the upsampling filter upsamples the low-dimension frequency-domain self-interference channel estimate by an upsampling factor greater than or equal to the radix.
18. The system of claim 17 , wherein:
the IFFT processor implements an IFFT butterfly structure defining a second plurality of stages, wherein a stage of the second plurality defines a second radix equal to the radix; and
the decimating filter downsamples the self-interference channel estimate by an upsampling factor greater than or equal to the radix.
19. The system of claim 1 , wherein the first transformer comprises:
a pre-compensator that, in the time domain, generates a compensated transmit signal based on the digital transmit signal and generates a compensated residue signal based on the digital residue signal;
a fast Fourier transform processor that transforms the compensated transmit signal into a low-dimension frequency-domain transmit signal, and that transforms the compensated residue signal into a low-dimension frequency-domain residue signal; and
an upsampling filter that, in the frequency domain, upsamples the low-dimension frequency-domain transmit signal to generate an upsampled transmit signal, and that upsamples the low-dimension frequency-domain residue signal to generate an upsampled residue signal;
wherein the channel estimator generates the self-interference channel estimate based on the upsampled transmit signal and the upsampled residue signal.