Methods, devices, and computer programs for compensating nonlinearities of a communication channel
A method, carried out by a gateway transmitter ( 400 ), aims at compensating the nonlinearities of a communication channel ( 500 ) comprising a repeater ( 510 ). A plurality of digital signals is modulated (s 10 ) on a plurality of carriers, wherein symbols of the constellation diagram used for modulation of each carrier are distorted in accordance with a pre-distortion function. The modulated signals are then frequency division multiplexed (s 20 ), and sent (s 30 ) for transmission, through the communication channel ( 500 ), to at least one receiver ( 600 ). The pre-distortion function involves a plurality of polynomial functions, each of which taking as input the symbols from all the carriers. The polynomial functions' coefficients, called “pre-distortion coefficients”, are computed according to a direct learning approach, performed jointly for the plurality of carriers. The pre-distortion coefficients are iteratively updated based on received signals being fed back from a receiver ( 600 ). The invention also relates to transmitters and computer programs.
1. Method, carried out by a transmitter, hereinafter referred to as “gateway transmitter”, for compensating, at least partially, nonlinearities of a communication channel comprising at least one repeater, the method comprising:
modulating a plurality of digital signals on a plurality of carriers respectively, wherein symbols of a constellation diagram used for modulation of each carrier are distorted in accordance with a function, hereinafter referred to as “pre-distortion function”;
frequency division multiplexing the plurality of modulated digital signals; and
sending the frequency division multiplexed plurality of modulated digital signals for transmission, through the communication channel, to at least one receiver;
wherein
the pre-distortion function involves a plurality of polynomial functions, each of which taking as input the symbols from all the carriers, and the polynomial functions' coefficients being hereinafter referred to as “pre-distortion coefficients”, and
the pre-distortion coefficients are computed in accordance with a direct learning approach, the computation being performed jointly for the plurality of carriers, and the computation comprising iteratively updating the pre-distortion coefficients based on received signals being fed back from at least one among the at least one receiver to the gateway transmitter.
2. Method of claim 1 , wherein the at least one repeater comprises a transponder on board a satellite.
3. Method of claim 2 , wherein the transponder on board the satellite comprises at least one of an input multiplexing filter, an amplifier and an output multiplexing filter.
4. Method of claim 1 , wherein the at least one receiver is a plurality of receivers.
5. Method of claim 1 , wherein, before received fed back signals are available, the pre-distortion coefficients are set as if the communication channel had a linear gain with a gain factor being unity.
6. Method of claim 1 , wherein, before received fed back signals are available, the pre-distortion coefficients are set based on channel model parameters obtained in advance through at least one of tests, simulations and studies carried out on, or regarding, at least one of
the communication channel; and
at least one among the at least one repeater.
7. Method of claim 1 , wherein, for each carrier, only one of the plurality of polynomial functions is used for distorting symbols of the constellation diagram used for modulation of that carrier.
8. Method of claim 1 , wherein the direct learning approach employs at least one of a recursive least squares filter and a least mean squares filter.
9. Method of claim 1 , wherein
the plurality of digital signals being subject to modulation are originally using packets,
each of the packets is assigned a packet identifier,
the packets are kept stored in the gateway transmitter together with their packet identifiers, after transmission,
the received fed back signals are in the form of packets or rearranged, by the gateway transmitter, in the form of packets, and
the computation of the pre-distortion coefficients involves comparing the contents of the received fed back signals' packets with the contents of buffered packets to estimate an error rate caused by the communication channel.
10. Method of claim 1 , wherein fed back information on the received signals on all active carriers is required for triggering the process of computing the pre-distortion coefficients.
11. Transmitter, hereinafter referred to as “gateway transmitter”, for compensating, at least partially, nonlinearities of a communication channel comprising at least one repeater, the gateway transmitter comprising:
a modulator for modulating a plurality of digital signals on a plurality of carriers respectively, wherein, in operation, symbols of a constellation diagram used for modulation of each carrier are distorted in accordance with a function, hereinafter referred to as “pre-distortion function”;
a multiplexer for frequency division multiplexing the plurality of modulated digital signals;
a transmitter element for sending the frequency division multiplexed plurality of modulated digital signals for transmission, through the communication channel, to at least one receiver;
the gateway transmitter being such that
the pre-distortion function involves a plurality of polynomial functions, each of which taking as input the symbols from all the carriers, and the polynomial functions' coefficients being hereinafter referred to as “pre-distortion coefficients”, and
the pre-distortion coefficients are computed in accordance with a direct learning approach, the computation being performed jointly for the plurality of carriers, and the computation comprising iteratively updating the pre-distortion coefficients based on received signals being fed back from at least one among the at least one receiver to the gateway transmitter.
12. Gateway transmitter according to claim 11 , being such that, in operation, before received fed back signals are available, the pre-distortion coefficients are set as if the communication channel had a linear gain with a gain factor being unity.
13. Gateway transmitter according to claim 11 , being such that, in operation, before received fed back signals are available, the pre-distortion coefficients are set based on channel model parameters obtained in advance through at least one of tests, simulations and studies carried out on, or regarding, at least one of
the communication channel; and
at least one among the at least one repeater.
14. Gateway transmitter according to claim 11 , being such that, for each carrier, only one of the plurality of polynomial functions is used for distorting symbols of the constellation diagram used for modulation of that carrier.
15. Gateway transmitter according to claim 11 , wherein the direct learning approach employs at least one of a recursive least squares filter and a least mean squares filter.
16. Gateway transmitter according to claim 11 , being such that, in operation,
the plurality of digital signals being subject to modulation are originally using packets,
each of the packets is assigned a packet identifier,
the packets are kept stored in the gateway transmitter together with their packet identifiers, after transmission,
the received fed back signals are in the form of packets or rearranged, by the gateway transmitter, in the form of packets, and
the computation of the pre-distortion coefficients involves comparing the contents of the received fed back signals' packets with the contents of the buffered packets to estimate an error rate caused by the communication channel.
17. Gateway transmitter according to claim 11 , being such that fed back information on the received signals on all active carriers is required for triggering the process of computing the pre-distortion coefficients.
18. System comprising a gateway transmitter according to claim 11 , and the at least one repeater.
19. System of claim 18 , further comprising the at least one receiver.
20. Computer program comprising a non-transitory computer readable medium including computer-readable instructions configured, when executed on a computer, to cause the computer to
determine coefficients, hereinafter referred to as “pre-distortion coefficients”, involved in a function, hereinafter referred to as “pre-distortion function”, for use in distorting symbols of each of a constellation diagrams used for modulation of, respectively, a plurality of carriers carrying a plurality of digital signals,
wherein
distorting is suitable, in a transmitter, hereinafter referred to as “gateway transmitter”, to compensate at least partially, the nonlinearities of a communication channel comprising at least one repeater;
distorting is carried out prior to frequency division multiplexing the plurality of modulated digital signals, and prior to sending the frequency division multiplexed plurality of modulated digital signals for transmission, through the communication channel, to at least one receiver,
the pre-distortion function involves a plurality of polynomial functions, each of which taking as input the symbols from all the carriers, and the polynomial functions' coefficients being the above-mentioned pre-distortion coefficients, and
the pre-distortion coefficients are computed in accordance with a direct learning approach, the computation being performed jointly for the plurality of carriers, and the computation comprising iteratively updating the pre-distortion coefficients based on received signals being fed back from at least one among the at least one receiver.