Iterative equalization with non-linear soft interference cancellation in non-linear satellite channels
A method for removing distortions in a transmitted signal transmitted by a high power amplifier in a satellite communications system. The method characterizes the high power amplifier to define a series of Volterra coefficients and uses those coefficients in an equalizer in a receiver in the communications system to remove the distortions. The equalizer is a non-linear soft interference cancellation and minimum mean square error equalizer that employs three processing operations including parallel soft interference cancellation, minimum mean square error filtering and a priori log-likelihood ratio calculations.
1. A method for removing distortions in an actual transmitted signal transmitted by a high power amplifier, said actual transmitted signal including bits, said method comprising:
providing a model of the high power amplifier that generates a reproduced transmitted signal of the actual transmitted signal;
defining a series of Volterra coefficients from the reproduced transmitted signal using a Volterra model;
receiving the actual transmitted signal in a receiver; and
processing the actual transmitted signal in an equalizer in the receiver to generate a digital signal including the bits, wherein processing the actual transmitted signal in the equalizer includes providing parallel soft interference cancellation of the actual transmitted signal that cancels both linear and non-linear interference for each symbol of interest in the actual transmitted signal, filtering the actual transmitted signal after the parallel soft interference cancellation using a minimum mean square error filter that employs the Volterra coefficients to provide equalizer coefficients, and using the equalizer coefficients to generate a new extrinsic log-likelihood ratio (LLR) of the bits in the actual transmitted signal.
2. The method according to claim 1 wherein processing the actual transmitted signal in an equalizer includes providing the new extrinsic LLR of the bits to an outer channel decoder that decodes the bits.
3. The method according to claim 2 wherein the outer channel decoder provides an a priori LLR of the bits as a feedback signal to the equalizer.
4. The method according to claim 3 wherein providing parallel soft interference cancellation of the actual transmitted signal includes using the a priori LLR of the bits and a symbol mean to subtract the interference.
5. The method according to claim 2 wherein using the equalizer coefficients to generate a new extrinsic LLR of the bits includes generating an a posteriori LLR of the bits using the equalizer coefficients and subtracting the a priori LLR of the bits from the a posteriori LLR of the bits to generate the new extrinsic LLR of the bits.
6. The method according to claim 1 wherein filtering the actual transmitted signal after signal cancellation in the equalizer includes solving an optimization problem that identifies a series of filter taps.
7. The method according to claim 1 wherein using the equalizer coefficients to generate the new extrinsic LLR includes employing a Gaussian approximation of inter-symbol interference plus noise.
8. The method according to claim 1 wherein the high power amplifier is a traveling-wave tube amplifier or a solid-state power amplifier.
9. The method according to claim 1 wherein the actual transmitted signal is a satellite signal.
10. A method for removing distortions from a transmitted signal received by a receiver using an equalizer in the receiver, said transmitted signal including bits, said method comprising:
receiving an a priori log-likelihood ratio (LLR) of the bits in the equalizer;
providing parallel soft interference cancellation of the transmitted signal that cancels both linear and non-linear interference for each symbol of interest in the transmitted signal using the a priori LLR of the bits;
filtering the transmitted signal after the soft interference cancellation using a minimum mean square error filter that employs a series of Volterra coefficients provided by a Volterra model to generate equalizer coefficients;
generating an a posteriori LLR of the bits in the transmitted signal using the equalizer coefficients; and
subtracting the a priori LLR of the bits from the a posteriori LLR of the bits to generate a new extrinsic LLR of the bits.
11. The method according to claim 10 wherein providing parallel soft interference cancellation of the transmitted signal includes using the a priori LLR of the bits and a symbol mean to subtract the interference.
12. The method according to claim 10 wherein filtering the transmitted signal after soft interference cancellation in the equalizer includes solving an optimization problem that identifies a series of filter taps.
13. The method according to claim 10 wherein using the equalizer coefficients to generate an a priori LLR includes employing a Gaussian approximation of inter-symbol interference plus noise.
14. The method according to claim 10 wherein receiving an a priori LLR of the bits in the equalizer includes receiving the a priori LLR of the bits from an outer channel decoder.
15. The method according to claim 10 wherein the transmitted signal is transmitted from a traveling-wave tube amplifier or a solid-state power amplifier.
16. The method according to claim 10 wherein the transmitted signal is a satellite signal.
17. A receiver receiving a transmitted signal from a high power amplifier, said transmitted signal including bits, said receiver comprising:
an outer channel decoder decoding the bits in the transmitted signal and providing an a priori log-likelihood ratio (LLR) of the bits; and
an equalizer receiving the transmitted signal, said equalizer including a parallel soft interference cancellation processor that cancels both linear and non-linear interference for each symbol of interest in the transmitted signal using the a priori LLR of the bits, said equalizer further including a minimum mean square error filter that receives the transmitted signal from the parallel soft interference cancellation processor and filters the transmitted signal using Volterra coefficients to provide equalizer coefficients, said equalizer further including an LLR processor that generates an a posteriori LLR of the bits in the transmitted signal using the equalizer coefficients and generates an extrinsic LLR of the bits by subtracting the a priori LLR of the bits from the a posteriori LLR of the bits.
18. The receiver according to claim 17 wherein the parallel soft interference cancellation processor uses the a priori LLR of the bits and a symbol mean to subtract the interference.
19. The receiver according to claim 17 wherein the minimum mean square error filter solves an optimization problem that identifies a series of filter taps.
20. The receiver according to claim 17 wherein the LLR processor employs a Gaussian approximation of inter-symbol interference plus noise to generate the a priori LLR.