NON-LINEAR ENCODER FOR FIBER OPTIC TRANSMISSION

Description

TECHNICAL FIELD

[0001] The present disclosure relates to data communication, and more particularly to an encoder for fiber-optic communication. Moreover, the present disclosure also relates to a fiber-optic transmission apparatus having such encoder, for modulating an optical carrier signal.BACKGROUND

[0002] Fiber-optic communication is a method that transmits an information from one place to another place, through an optical fiber, by sending an optical signal which is modulated to carry the information. Transmission of the optical signal over the optical fiber is usually impaired by a significant amount of noise, such as additive white gaussian noise (AWGN) and nonlinear noise. Such noise can distort the optical signal en route that tends to affect bit-error-rate (BER), symbol error rate (SER), or a maximum achievable transmission distance.

[0003] Optical fiber nonlinear effects such as self-phase modulation (SPM), cross-phase modulation (XPM), four-wave mixing (FWM) and scattering effects, may happen due to use of increased input power, higher bit rates, longer transmission distances, increased numbers of channels, and variation in refractive index of the optical fiber (also called Kerr effect). Further, in coherent fiber optics, there is a nonlinear Shannon limit which is caused by Kerr nonlinearity accumulated during propagation of the optical signal. The nonlinear Shannon limit has been identified as a fundamental barrier to a maximum rate of transmitted information in optical communications.

[0004] Several probabilistic shaping approaches are proposed for improving modulation of the optical signal. The probabilistic shaping approach is employed to limit occurrence of high-power symbols in order to reduce optical fiber nonlinear effects and to increase signal-to-noise ratio (SNR) for all other symbols. Existing methods are developed to realize probabilistic shaping such as enumeration scheme, trellis-coded modulation (TCM), and block shaping.

[0005] The probabilistic shaping approach may be used for capacity penalty reduction of traditional quadrature amplitude modulation (QAM) signals in an AWGN channel. TCM approach may be used for channels with memory to achieve better performance. Enumeration scheme may be optimal for data block with a finite length, but it is difficult to unite the data block with a forward error correction (FEC) module. Since it may cause an error propagation effect that should be taken into account when uniting the data block with the FEC module.

[0006] Some existing approaches may focus on increasing the input power. Increasing the input power may increase SNR but it leads to a nonlinear regime immediately which may produce additional penalty. Some existing approaches may consider different nonlinear metrics minimization, but they mostly considered it from hard restriction point of view. Some of the existing approaches propose perturbation-based model for nonlinearity compensation. Further, with respect to shaping with memory, memory-less super gaussian shaping may provide performance in links with compact disc (CD) management. Temporal probabilistic shaping may support short memory effects.

[0007] Even though there have been several probabilistic shaping methods available, proper metrics for nonlinear constellation shaping are still unknown. There is no common methodology to cover nonlinear metric minimization of arbitrary form within a block of big memory.

[0008] Therefore, there arises a need to address the aforementioned technical drawbacks in existing systems or technologies in reducing the optical fiber nonlinear effects.

[0009] Document US 10 587 358 B1 describes an optical transmitter, which is operative to generate, from a plurality of encoded client bits, a set of symbols exhibiting non-uniform visitation probabilities in at least one dimension, to encode the set of symbols across a plurality of frequency division multiplexing (FDM) subcarriers, and to transmit an optical signal comprising the plurality of FDM subcarriers across which the set of symbols is encoded. An optical receiver is operative to decode a set of symbol estimates from a plurality of FDM subcarriers, the symbol estimates comprising estimates of symbols exhibiting non-uniform visitation probabilities in at least one dimension, and to recover client bits from the set of symbol estimates.

[0010] Document JP H11 215091 A describes a sending side that applies scramble systems #1 to #N to a sending data system. Outputs of scrambling parts 10#1 to 10#N are modulated, and plural OFDM symbols are produced. A selecting part 16 selects one of plural OFDM symbols and sends it. A receiving side descrambles digital data obtained from a received signal by descrambling parts 112#1 to 112#N. One among plural descrambled outputs is a correct data system. A selecting part outputs the data system that is correctly descrambled as a receiving data system.SUMMARY

[0011] It is an object of the present disclosure to provide an improved approach for reducing nonlinearity accumulation during propagation of an optical carrier signal through an optical fiber.

[0012] This object is achieved by the features of the independent claims. Further implementation forms are apparent from the dependent claims, the description, and the figures.

[0013] The present disclosure provides an encoder for fiber-optic communication and also a fiber-optic transmission apparatus having such encoder.

[0014] According to a first aspect, there is provided an encoder for fiber-optic communication. The encoder includes an input, a digital signal processor and an output. The input is configured to receive an information bit sequence. The digital signal processor (DSP) is configured to translate the information bit sequence into a modulation symbol sequence for modulating an optical carrier signal. The output is configured to output the modulation symbol sequence. The DSP is configured to translate the information bit sequence into the modulation symbol sequence such that the modulation symbol sequence has, among all symbol sequences in a candidate set of symbol sequences, a smallest value of a metric. The symbol sequences in the candidate set are different representations of the information bit sequence in a symbol space. The metric is defined as S = ∑ m ∈ M w XX , m ⋅ S XX , m + w YY , m ⋅ S YY , m + w XY , m ⋅ S XY , m for any symbol sequence Z = (Z[0], ..., Z[N - 1]) in the candidate set, where, N is a common length of the symbol sequences in the candidate set, Z[k] is a symbol at a position k in the symbol sequence (Z[0], ... , Z[N - 1]), k ∈ {0, ..., N - 1}, the symbol Z[k] includes one or both of the following: a complex-valued first symbol X[k] related to a first polarization and a complex-valued second symbol Y[k] related to a second polarization, S is a value of the metric, m is a summation index, M is a set of one or more whole numbers in the range of -N + 2 to N - 2, ∑ m∈M means summation over all values of m in M, where m=0 to N-1, w XX,m , w XY,m , w YY,m are real-valued weights that depend on the value of m, S XX,m , S XY,m , S XY,m are non-negative numbers defined as S XX , m = ∑ k ∈ K X k X * k + m α , S YY , m = ∑ k ∈ K Y k Y * k + m α , S XY , m = ∑ k ∈ K X k Y * k + m α , | | means an absolute value of a complex number, α is a positive real number, ( ) α< means "to the power of α", k is a summation index, K is a set of one or more whole numbers, ∑ k∈K means summation over all values k in K, and ( ) *< means complex conjugation.

[0015] The encoder enables the addition of nonlinear properties in the optical carrier signal with low rate loss to decrease nonlinearity accumulation during propagation of the optical carrier signal through an optical fiber. The optical carrier signal modulated with the modulation symbol sequence may reduce optical nonlinear effects such as self-phase modulation (SPM), cross-phase modulation (XPM), and four-wave mixing (FWM) together during the propagation of the optical carrier signal. The encoder optionally employs a shaping procedure to translate the information bit sequence into the modulation symbol sequence with the smallest value of the metric. The shaping procedure is flexible, and may be applied to quadrature amplitude modulation (QAM) constellation, and phase-shift keying (PSK) constellation. The shaping procedure may be applied to any number of bit streams or even part of the bit streams, so it is possible to combine the shaping procedure with convenient enumeration probabilistic shaping. Physics relative performance of the shaping procedure may be increased with growing numbers of shaped subcarriers in dense wavelength division multiplexing (DWDM) signal. The shaping procedure may not affect forward error correction (FEC) work and vice versa.

[0016] In one possible implementation form, the set K includes at least one number that is different from zero.

[0017] In another possible implementation form, wherein S XX , m = ∑ k = 0 N − 1 − m X k X * k + m α , S YY , m = ∑ k = 0 N − 1 − m Y k Y * k + m α , S XY , m = ∑ k = 0 N − 1 − m X k Y * k + m α , form ∈ {0, ...,N - 2} and, wherein S XX , m = ∑ k = − m N − 1 X k X * k + m α , S YY , m = ∑ k = − m N − 1 Y k Y * k + m α , S XY , m = ∑ k = − m N − 1 X k Y * k + m α . form ∈ {-N + 2, ..., -1}.

[0018] In yet another possible implementation form, wherein S XX , m = ∑ k = 0 N − 1 X k X * k + m α , S YY , m = ∑ k = 0 N − 1 Y k Y * k + m α , S XY , m = ∑ k = 0 N − 1 X k Y * k + m α , for any m ∈ M, with the following cyclic definitions: X k + p ∗ N = X k , Y k + p ∗ N = Y k , p ∈ − 1 , 0 , 1 .

[0019] The set M may be a superset of at least one of the following sets: 0 1 , − 1 , 0 , − 1 , 0 , 1 , 0 1 2 , − 2 , − 1 , 0 , − 2 , − 1 , 0 , 1 , 2 , 0 1 2 3 , − 3 , − 2 , − 1 , 0 , − 3 , − 2 , − 1 , 0 , 1 , 2 , 3 , where M=3.

[0020] In yet another possible implementation form, wherein α = 1.

[0021] In yet another possible implementation form, Z[k] includes only the first symbol X [k] and no further symbol. The metric is defined as S = ∑ m w XX , m ⋅ S XX , m .

[0022] The metric S may be a measure of expected distortion, the expected distortion being distortion of a hypothetical optical carrier signal that is modulated with the symbol sequence Z and transmitted via an optical fiber. The encoder may select a symbol sequence as the modulation symbol sequence that is able to yield a modulated optical carrier signal which may be less affected by distortion when propagating through the optical fiber, when compared to other symbol sequences in the candidate set of symbol sequences.

[0023] In yet another possible implementation form, each of X[k] and Y[k] is complex-valued. Each of X[k] and Y[k] is optionally an amplitude and phase shift keying (APSK) symbol. Each of X[k] and Y[k] is optionally a quadrature-amplitude modulation (QAM) symbol. Each of X[k] and Y[k] is optionally real-valued. Each of X[k] and Y[k] is optionally a pulse-amplitude modulation (PAM) symbol.

[0024] In yet another possible implementation form, the DSP is configured to obtain the modulation symbol sequence by selecting, among the symbol sequences in the candidate set, a symbol sequence that has the smallest value of the metric. The optical carrier signal modulated with the modulation symbol sequence may be less affected by the distortion during propagation through the optical fiber.

[0025] In yet another possible implementation form, the DSP is configured to generate the candidate set of symbol sequences by generating a plurality of shaped bit sequences, each of the shaped bit sequences being a representation of the information bit sequence; and mapping each shaped bit sequence to the symbol space.

[0026] In yet another possible implementation form, each of the shaped bit sequences includes an encoded form of the information bit sequence, and an indication of a code word for decoding the encoded form.

[0027] In yet another possible implementation form, the encoded form of the information bit sequence includes one or more non-encoded bits of the information bit sequence. The encoded form of the information bit sequence optionally does not comprise any non-encoded bits of the information bit sequence.

[0028] In a possible implementation form, the DSP is configured to generate the encoded form of the information bit sequence by applying the code word to the information bit sequence.

[0029] In another possible implementation form, applying the code word to the information bit sequence includes, for each bit of the code word, a logical XOR operation on a pair of bits, the bit pair of bits including the bit of the code word and a bit of the information bit sequence. The output is optionally connected or connectable to an optical transmitter.

[0030] According to a second aspect, there is provided a fiber-optic transmission apparatus. The fiber-optic transmission apparatus includes the encoder as described above and an optical transmitter. The optical transmitter includes an input connected to the output of the encoder and an output connected or connectable to an optical fiber.

[0031] The fiber-optic transmission apparatus may add nonlinear properties in the optical carrier signal with low rate loss to decrease nonlinearity accumulation during propagation of the optical carrier signal through an optical fiber. The optical carrier signal modulated with the modulation symbol sequence may reduce optical nonlinear effects such as self-phase modulation (SPM), cross-phase modulation (XPM), and four-wave mixing (FWM) together during propagation of the optical carrier signal. The encoder optionally employs a shaping procedure to translate the information bit sequence into the modulation symbol sequence with the smallest value of the metric. The shaping procedure is flexible, and may be applied to quadrature amplitude modulation (QAM) constellation, and phase-shift keying (PSK) constellation. The shaping procedure may be applied to any number of bit streams or even part of the bit streams, so it is possible to combine the shaping procedure with convenient enumeration probabilistic shaping. Physics relative performance of the shaping procedure may be increased with growing numbers of shaped subcarriers in dense wavelength division multiplexing (DWDM) signal. The shaping procedure may not affect forward error correction (FEC) work and vice versa.

[0032] In a possible implementation form, the optical transmitter is configured to modulate an optical carrier signal based on the modulation symbol sequence, and output the modulated optical carrier signal at the output of the optical transmitter.

[0033] A technical problem in the prior art is resolved, where the technical problem concerns reducing the optical fiber nonlinear effects.

[0034] Therefore, in contradistinction to the prior art, according to the encoder for fiber-optic communication in the present disclosure, the shaping procedure is employed to translate the information bit sequence into the modulation symbol sequence for modulating the optical carrier signal. The modulation symbol sequence has the smallest value of the metric among all the symbol sequences in the candidate set of symbol sequences. The optical carrier signal modulated with the modulation symbol sequence may reduce optical nonlinear effects such as SPM, XPM, and FWM together during the propagation. Further, the shaping procedure has low overhear.

[0035] These and other aspects of the present disclosure will be apparent from and the implementation(s) described below.BRIEF DESCRIPTION OF DRAWINGS

[0036] Implementations of the present disclosure will now be described, by way of example only, with reference to the accompanying drawings, in which: FIG. 1 is a block diagram of an encoder for fiber-optic communication in accordance with an implementation of the present disclosure; FIG. 2 is an exemplary schematic illustration of a process of translating an information bit sequence into a modulation symbol sequence using the encoder of FIG.1 in accordance with an implementation of the present disclosure; FIG. 3 is an exemplary schematic illustration of a process of translating an information bit sequence into a modulation symbol sequence for a binary phase shift keying (BPSK) modulation in accordance with an implementation of the present disclosure; FIG. 4 is an exemplary schematic illustration of a quadrature phase shift keying (QPSK) shaping in accordance with an implementation of the present disclosure; FIG. 5 is an exemplary graphical representation of a normalized mean squared error (NMSE) versus an optical signal-to-noise ratio (OSNR) for a QPSK modulation in accordance with an implementation of the present disclosure; FIG. 6 is an exemplary graphical representation of a bit-error-rate (BER) versus an OSNR for a QPSK modulation in accordance with an implementation of the present disclosure; FIG. 7 is an exemplary schematic illustration of a 16 quadrature amplitude modulation (16QAM) shaping in accordance with an implementation of the present disclosure; FIG. 8 is an exemplary graphical representation of a normalized mean squared error (NMSE) versus an optical signal-to-noise ratio (OSNR) for a 16QAM in accordance with an implementation of the present disclosure; FIG. 9 is an exemplary graphical representation of a bit-error-rate (BER) versus an OSNR for a 16QAM in accordance with an implementation of the present disclosure; FIG. 10 is an exemplary graphical representation of a measure of power spectral density (PSD) for nonlinear term X*conjugate(Y) in accordance with an implementation of the present disclosure; FIG. 11 is a block diagram of a fiber-optic transmission apparatus in accordance with an implementation of the present disclosure; and FIG. 12 is a flow diagram of a method of operating an encoder for fiber-optic communication in accordance with an implementation of the present disclosure. DETAILED DESCRIPTION OF THE DRAWINGS

[0037] Implementations of the present disclosure provide an encoder for fiber-optic communication that translates an information bit sequence into a modulation symbol sequence for modulating an optical carrier signal and a fiber-optic transmission apparatus having such encoder, for modulating the optical carrier signal.

[0038] To make solutions of the present disclosure more comprehensible for a person skilled in the art, the following implementations of the present disclosure are described with reference to the accompanying drawings.

[0039] Terms such as "a first", "a second", "a third", and "a fourth" (if any) in the summary, claims, and foregoing accompanying drawings of the present disclosure are used to distinguish between similar objects and are not necessarily used to describe a specific sequence or order. It should be understood that the terms so used are interchangeable under appropriate circumstances, so that the implementations of the present disclosure described herein are, for example, capable of being implemented in sequences other than the sequences illustrated or described herein. Furthermore, the terms "include" and "have" and any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, a method, a system, a product, or a device that includes a series of steps or units, is not necessarily limited to expressly listed steps or units but may include other steps or units that are not expressly listed or that are inherent to such process, method, product, or device.

[0040] FIG. 1 is a block diagram of an encoder 100 for fiber-optic communication in accordance with an implementation of the present disclosure. The encoder 100 includes an input 102, a digital signal processor (DSP) 106, and an output 108. The input 102 is configured to receive an information bit sequence 110. The digital signal processor (DSP) 106 is configured to translate the information bit sequence 110 into a modulation symbol sequence 112 for modulating an optical carrier signal. The output 108 is configured to output the modulation symbol sequence 112. The information bit sequence 110 may be a bit sequence or a representation of digital content in an assembly of a fundamental unit of digital bits. The optical carrier signal is a waveform that is modulated (or modified) with an information signal (or message signal) for a purpose of conveying information or a message.

[0041] The DSP 106 is configured to translate the information bit sequence 110 into the modulation symbol sequence 112 such that the modulation symbol sequence 112 has, among all symbol sequences in a candidate set of symbol sequences, a smallest value of a metric. The symbol sequences in the candidate set are different representations of the information bit sequence 110 in a symbol space. The symbol space optionally means a dimension where information can be expressed or transferred.

[0042] The metric is defined as: S = ∑ m ∈ M w XX , m ⋅ S XX , m + w YY , m ⋅ S YY , m + w XY , m ⋅ S XY , m for any symbol sequence Z = (Z[0], ..., Z[N - 1]) in the candidate set, where N is a common length of the symbol sequences in the candidate set, Z[k] is a symbol at a position k in the symbol sequence (Z[0], ... , Z[N - 1]), k ∈ {0, ..., N - 1}, the symbol Z[k] includes one or both of the following: a complex-valued first symbol X[k] related to a first polarization and a complex-valued second symbol Y[k] related to a second polarization, S is a value of the metric, m is a summation index, M is a set of one or more whole numbers in the range of -N + 2 to N - 2, ∑ m∈M means summation over all values of m in M, where m=0 to N-1, w XX,m , w XY,m , w YY,m are real-valued weights that depend on the value of m, S XX,m , S XY,m , S XY,m are non-negative numbers defined as S XX , m = ∑ k ∈ K X k X * k + m α , S YY , m = ∑ k ∈ K Y k Y * k + m α , S XY , m = ∑ k ∈ K X k Y * k + m α , | | means an absolute value of a complex number, α is a positive real number, ( ) α< means "to the power of α", k is a summation index, K is a set of one or more whole numbers, ∑ k∈K means summation over all values k in K, and ( ) *< means complex conjugation.

[0043] The encoder 100 may further include a memory 104 that stores one or more code words (i.e. a code book). The memory 104 may be communicatively coupled with the DSP 106. The code book may be a binary code book. The code book may be used to shape one or more bits of the information bit sequence 110. Each code word may include N coding bits, where N is a natural number. The information bit sequence 110 may include M data bits, where M is a natural number greater than N. The encoder 100 may use the code book to generate the candidate set of symbol sequences.

[0044] Further, the output 108 may be connected or connectable to an optical transmitter. The length of the symbol sequences is optionally equal to a number of elements of the symbol sequences.

[0045] In a first implementation, the set K includes at least one number that is different from zero. In a second implementation, non-cyclic definition of [k + m], form ∈ {0, ..., N - 2}, includes S XX , m = ∑ k = 0 N − 1 − m X k X * k + m α , S YY , m = ∑ k = 0 N − 1 − m Y k Y * k + m α , S XY , m = ∑ k = 0 N − 1 − m X k Y * k + m α .

[0046] The non-cyclic definition of [k + m], for m ∈ {-N + 2, ..., -1}, may include S XX , m = ∑ k = − m N − 1 X k X * k + m α , S YY , m = ∑ k = − m N − 1 Y k Y * k + m α , S XY , m = ∑ k = − m N − 1 X k Y * k + m α .

[0047] In a third implementation, cyclic definition of [k + m], includes S XX , m = ∑ k = 0 N − 1 X k X * k + m α , S YY , m = ∑ k = 0 N − 1 Y k Y * k + m α , S XY , m = ∑ k = 0 N − 1 X k Y * k + m α , for any m∈M, with the following cyclic definitions: X k + p ∗ N = X k , Y k + p ∗ N = Y k , p ∈ − 1 , 0 , 1 .

[0048] In a fourth implementation, the set M is a superset of at least one of the following sets: 0 1 , − 1 , 0 , − 1 , 0 , 1 , 0 1 2 , − 2 , − 1 , 0 , − 2 , − 1 , 0 , 1 , 2 , 0 1 2 3 , − 3 , − 2 , − 1 , 0 , − 3 , − 2 , − 1 , 0 , 1 , 2 , 3 , where M=3. Optionally, M is natural number 1,2,3,4,5,6, etc.

[0049] In a fifth implementation, where α = 1. For example, power = 1. The encoder 100 may be used in phase-shift keying (PSK) modulation, amplitude and phase shift keying (APSK) modulation, quadrature-amplitude modulation (QAM), or pulse-amplitude modulation (PAM).

[0050] In a sixth implementation, each of X[k] and Y[k] is complex-valued. Each of X[k] and Y[k] may be real-valued. Each of X[k] and Y[k] is optionally an amplitude and phase shift keying (APSK) symbol. Each of X[k] and Y[k] may be a quadrature-amplitude modulation (QAM) symbol. Each of X[k] and Y[k] may be a pulse-amplitude modulation (PAM) symbol.

[0051] The modulation symbol sequence 112 with a minimized nonlinear metric may reduce optical nonlinear effects such as self-phase modulation (SPM), cross phase modulation (XPM), and four-wave mixing (FWM) together during propagation of the optical carrier signal.

[0052] With reference to FIG.1, FIG. 2 is an exemplary schematic illustration of a process of translating the information bit sequence 110 into the modulation symbol sequence 112 using the encoder 100 of FIG.1 in accordance with an implementation of the present disclosure. The DSP 106 may translate the information bit sequence 110 into the modulation symbol sequence 112 by encoding one or more bits of the information bit sequence 110 using one or more code words of a code book 212. The information bit sequence 110 may include a shaping block 202 to be encoded, and an information block 204 not to be encoded. The DSP 106 may be configured to generate the candidate set of symbol sequences by generating a plurality of shaped bit sequences, each of the shaped bit sequences being a representation of the information bit sequence 110 and mapping each shaped bit sequence to the symbol space. The DSP 106 may be configured to obtain the modulation symbol sequence 112 by selecting, among the symbol sequences in the candidate set, a symbol sequence that has the smallest value of the metric.

[0053] Each of the plurality of shaped bit sequences may include an encoded form of the information bit sequence 110, and an indication of a code word for decoding the encoded form. The encoded form of the information bit sequence 110 may include one or more non-encoded bits of the information bit sequence 110. The encoded form of the information bit sequence 110 may not include any non-encoded bits of the information bit sequence 110. For example, the modulation symbol sequence 112 includes encoded bits 206 of the information bit sequence 110, and an indication 208 of the code word of the code book 212 for decoding the encoded bits 206. Optionally, the modulation symbol sequence 112 includes non-encoded bits 210 of the information bit sequence 110. The modulation symbol sequence 112 may not include any non-encoded bits of the information bit sequence 110 (not shown in FIG.2).

[0054] The DSP 106 is configured to generate the encoded form of the information bit sequence 110 by applying the code word to the information bit sequence 110. Applying the code word to the information bit sequence 110 may include, for each bit of the code word, a logical XOR operation on a pair of bits, the bit pair of bits including the bit of the code word, and a bit of the information bit sequence 110. In the code book 212, a length of the code word may be represented as 214 and a size of the code book 212 may be represented as 216. The length of the code word 214 may be equal to a length of the shaping block 202 of the information bit sequence 110. A number of the code word may be a power of 2.

[0055] The DSP 106 may apply the code word to the information bit sequence 110 by performing the logical XOR operation on the pair of bits. The pair of bits include the bit of the code word and the bit of the information bit sequence 110. The DSP 106 may generate the encoded bits 206 of the modulation symbol sequence 112 by applying the code word of the code book 212 to the information bit sequence 110. The DSP 106 may convert an index of the code word to a binary image and add the binary image as the indication of the code word in each of the shaped bit sequence.

[0056] After applying the code word, the DSP 106 may calculate a value of the metric S for each of the shaped bit sequences using the formula (1) as illustrated above, resulting value of the metric S may be presented as linear combination of them, weights w XX,m , w XY,m , w YY,m may be optimized for application. The shaping block 202 with an applied code word corresponding to a minimal metric value has added as the encoded bits 206 in the modulation symbol sequence 112. The information block 204 may be added in the modulation symbol sequence 112 without changes.

[0057] The metric S may be a measure of expected distortion. The expected distortion is a distortion of a hypothetical optical carrier signal that is modulated with the symbol sequence Z and transmitted via the optical fiber. A measure of the metric S may be depended on the weights w XX,m , w XY,m , w YY,m . The weights may be set at a factory. The weights can be hard-coded or soft-coded. The weights may be chosen (e.g. based on experiments conducted prior to production of the encoder 100 ) such that the encoder 100 may select a symbol sequence that may yield a modulated optical carrier signal which will be less affected by the distortion when propagating through the optical fiber, compared to other symbol sequences in the candidate set.

[0058] With reference to FIG.1, FIG. 3 is an exemplary schematic illustration of a process of translating an information bit sequence 302 into a modulation symbol sequence 310 for a binary phase shift keying (BPSK) modulation in accordance with an implementation of the present disclosure. The BPSK modulation is a two-phase modulation scheme, where the 0's and 1's in a binary message are represented by two different phase states in an optical carrier signal: θ = 0 0< for binary 1 and θ = 180 0< for binary 0. The input 102 of the encoder 100 may receive the information bit sequence 302. The DSP 106 may translate the information bit sequence 302 into the modulation symbol sequence 310 for modulating the optical carrier signal. The output 108 may output the modulation symbol sequence 310 to the optical transmitter for modulating the optical carrier signal with the modulation symbol sequence 310. In the BPSK modulation, 0 and 1 bits may correspond to -1 and 1 values. One polarization and one criteria Sxx,1 are considered. For example, Z[k] includes only a first symbol X[k] and no further symbol, and the metric may be defined as S = ∑ m w XX , m ⋅ S XX , m .

[0059] The term w YY,m · S YY,m + w XY,m · S XY,m in the above formula for the value of the metric S may be zero.

[0060] The DSP 106 may apply a first code word 304A and a second code word 304B to the information bit sequence 302. The DSP 106 may perform the logical XOR operation between the information bit sequence 302 and the first code word 304A and the second code word 304B to generate a first shaped bit sequence 306A and a second shaped bit sequence 306B. The first shaped bit sequence 306A may be a first representation of the information bit sequence 302. The second shaped bit sequence 306B may be a second representation of the information bit sequence 302. Further, the DSP 106 places a binary number of the first code word 304A and the second code word 304B as a first bit of the first shaped bit sequence 306A and the second shaped bit sequence 306B, after the XOR operation. The DSP 106 may map the first shaped bit sequence 306A and the second shaped bit sequence 306B to the symbol space to generate a first symbol sequence 308A and a second symbol sequence 308B respectively. The binary number of the first code word 304A and the second code word 304B may represent as an indication of the first code word 304A and the second code word 304B for decoding an encoded form of the first shaped bit sequence 306A and the second shaped bit sequence 306B. The DSP 106 derives a first metric value as 1 for the first symbol sequence 308A and a second metric value as 3 for the second symbol sequence 308B. The DSP 106 may select the second symbol sequence 306B, among the first symbol sequence 306A and the second symbol sequence 306B, as the modulation symbol sequence 310 that has a smallest value of the metric. The modulation symbol sequence 310 may be used to modulate the optical carrier signal.

[0061] FIG. 4 is an exemplary schematic illustration of a quadrature phase shift keying (QPSK) shaping in accordance with an implementation of the present disclosure. The QPSK is a form of phase shift keying in which two bits are modulated at once, selecting one of four possible carrier phase shifts (0, 90, 180, or 270 degrees). A QPSK modulation may be used in submarine application. Each QPSK symbol is described with 2 bits (I and Q) with Gray mapping. For X and Y polarization, 4 bits per symbol Xi, Xq, Yi, and Yq are considered. In FIG. 4, a code book with 128 code words may be used at each 64 symbols. Shaping cost of an information bit sequence is 7 / 256, i.e. 0.11 bit / symbol (~3%) rate loss. The Xq and Yq bits are encoded with the code book. First 4 (or 3) bits of Xq and 3 (or 4) bits of Yq are placed with an indication of the code word. Nonlinear metrics of m=0, 1, 2 orders are used. A modulated symbol sequence 400 is represented in FIG. 4 after quadrature phase shift keying (QPSK) shaping. The modulated symbol sequence 400 includes one or more non-encoded bits 402A-N, an indication of the code word 404A-N, and one or more encoded bits 406A-N. The one or more non-encoded bits 402A-N, and the one or more encoded bits 406A-N may be an encoded form of the information bit sequence.

[0062] With reference to FIG. 4, FIG. 5 is an exemplary graphical representation of a normalized mean squared error (NMSE) versus an optical signal-to-noise ratio (OSNR) for the QPSK modulation in accordance with an implementation of the present disclosure. The following simulation conditions are considered for the QPSK modulation: (1) QPSK modulation with and without nonlinear shaping of an information bit sequence, (2) 100 gigahertz (GHz) frequency band for each optical channel, (3) 11 optical channels, (4) 120 GHz optical channel spacing, (5) roll-off factor (RRC) as 0.1, (6) 5 decibel-milliwatts (dBm) of input power per optical channel, (7) 16.89 picoseconds / kilometer (ps / km) dispersion, (8) 1 / Watt / kilometer (1 / W / km) Kerr nonlinearity (including factor 8 / 9), (9) 0.18 decibel / kilometer (dB / km) attenuation, (10) 80 kilometer (km) span length, and (11) 140 spans (11200 km). In the exemplary graphical representation, the OSNR is plotted in an X-axis and the NMSE is plotted in a Y-axis. The NMSE is mean square error (MSE) normalized by variance of desired values. In order to evaluate a performance of an optical carrier signal after nonlinear shaping, the NMSE is plotted against the OSNR. In the exemplary graphical representation, a NMSE / OSNR curve 502 represents the QPSK modulation without encoding (i.e. without nonlinear shaping) the information bit sequence into the modulation symbol sequence 400 for modulating the optical carrier signal. A NMSE / OSNR curve 504 represents the QPSK modulation with encoding (i.e. with nonlinear shaping) the information bit sequence into the modulation symbol sequence 400 for modulating the optical carrier signal. The exemplary graphical representation as shown in FIG. 5 illustrates that the performance of the optical carrier signal, which is modulated with the modulation symbol sequence 400, is increased. An increase in the OSNR may increase an optical network throughput and energy efficiency and it enables to minimize a number of retransmissions of information.

[0063] With reference to FIG. 4 and FIG. 5, FIG. 6 is an exemplary graphical representation of a bit-error-rate (BER) versus an OSNR for the QPSK modulation in accordance with an implementation of the present disclosure. In the exemplary graphical representation, the OSNR is plotted in an X-axis, and the BER is plotted in a Y-axis. The BER is defined as the percentage of bits that have errors relative to a total number of bits received in a transmission. The OSNR may be used to quantify a degree of optical noise interference in an optical carrier signal. In the exemplary graphical representation, a BER / OSNR curve 602 represents the QPSK modulation without encoding (i.e. without nonlinear shaping) the information bit sequence into the modulation symbol sequence 400 for modulating the optical carrier signal. A BER / OSNR curve 604 represents the QPSK modulation with encoding (i.e. with nonlinear shaping) the information bit sequence into the modulation symbol sequence 400 for modulating the optical carrier signal. The exemplary graphical representation, as shown in FIG. 6, illustrates that 1dB OSNR improvement is achieved at BER 4.5E-2 level.

[0064] FIG. 7 is an exemplary schematic illustration of a 16 quadrature amplitude modulation (16QAM) shaping in accordance with an implementation of the present disclosure. The 16QAM is a higher order modulation technique, which allows one single radio wave to represent four bits of data by manipulating an amplitude and a phase of the radio wave into one of 16 different discrete and measurable states. The 16QAM may be used for long haul applications. The following simulation conditions are considered for 16QAM: (1) Each 16QAM symbol is described with 4 bits, (2) use of Gray mapping (e.g. 2 amplitude bits and 2 sign bits), (3) for X and Y polarization, 8 bits per symbol Xa1, Xa2, Xs1, Xs2, Ya1, Ya2, Ys1, Ys2 are considered, (4) a code book with 16 code words is used at each 64 symbols, (5) shaping cost of an information bit sequence is 4 / 512, i.e. 0.06 bit / symbol (~0.8%), (6) Xs2 and Ys2 bits are shaped using the code book, (7) first 2 bits of Xs2 and Ys2 are placed with an indication of the code word, (8) nonlinear metrics of m=0, 1 and 2 orders are used. A modulated symbol sequence 700 is represented in FIG. 7 after 16QAM shaping. The modulated symbol sequence 700 includes one or more non-encoded bits 702A-N, an indication of the code word 704A-N, and one or more encoded bits 706A-N. The one or more non-encoded bits 702A-N, and the one or more encoded bits 706A-N may be an encoded form of the information bit sequence.

[0065] With reference to FIG. 7, FIG. 8 is an exemplary graphical representation of a normalized mean squared error (NMSE) versus an optical signal-to-noise ratio (OSNR) for the 16QAM in accordance with an implementation of the present disclosure. In the exemplary graphical representation, the OSNR is plotted in an X-axis and the NMSE is plotted in a Y-axis. In the exemplary graphical representation, a NMSE / OSNR curve 802 represents the 16QAM without encoding (i.e. without nonlinear shaping) an information bit sequence into the modulation symbol sequence 700 for modulating an optical carrier signal. A NMSE / OSNR curve 804 represents the 16QAM with encoding (i.e. with nonlinear shaping) the information bit sequence into the modulation symbol sequence 700 for modulating the optical carrier signal. The exemplary graphical representation as shown in FIG. 8 illustrates that the performance of the optical carrier signal, which is modulated with the modulation symbol sequence 700, is increased. An increase in the OSNR may increase an optical network throughput and energy efficiency and it enables to minimize a number of retransmissions of information.

[0066] With reference to FIG. 7 and FIG. 8, FIG. 9 is an exemplary graphical representation of a bit-error-rate (BER) versus an OSNR for the 16QAM in accordance with an implementation of the present disclosure. In the exemplary graphical representation, the OSNR is plotted in an X-axis and BER is plotted in a Y-axis. A BER / OSNR curve 902 represents the 16QAM without encoding (i.e. without nonlinear shaping) an information bit sequence into the modulation symbol sequence 700 for modulating an optical carrier signal. A BER / OSNR curve 904 represents the 16QAM with encoding (i.e. with nonlinear shaping) the information bit sequence into the modulation symbol sequence 700 for modulating the optical carrier signal. The exemplary graphical representation, as shown in FIG. 9, illustrates that 0.7 decibel (dB) OSNR improvement is achieved at BER 4.5E-2 level.

[0067] FIG. 10 is an exemplary graphical representation of power spectral density (PSD) for nonlinear term X*conjugate(Y) in accordance with an implementation of the present disclosure. In the exemplary graphical representation, a relative frequency is plotted in an X-axis and a power / frequency is plotted in a Y-axis. Minimization of single or dual polarization nonlinear metrics is presented as low frequency PSD suppression of corresponding nonlinear terms. The nonlinear terms for an optical carrier signal for 16 quadrature amplitude modulation (QAM) or phase shift keying (PSK) modulator is calculated by: X ⋅ delay X * , k , X ⋅ delay Y * , k , Y ⋅ delay Y * , k ,

[0068] Where delay (X,k) means delay X at k samples X ⋅ delay X * , 0 = X 2 , Y ⋅ delay Y * , 0 = Y 2 ,

[0069] Any PSD of presented nonlinear term, excluding (2) which demonstrates low frequency suppression. PSD estimated at big enough block size shows suppression of nonlinear metric as shown in FIG. 10. Terms |X| 2< , |Y| 2< may correspond to power shaping and their PSD low frequency suppression.

[0070] FIG. 11 is a block diagram of a fiber-optic transmission apparatus 1100 in accordance with an implementation of the present disclosure. The fiber-optic transmission apparatus 1100 includes the encoder 100 (as shown in FIG. 1) and an optical transmitter 1102. The encoder 100 includes the input 102 that is configured to receive the information bit sequence 110, the memory 104, the digital signal processor 106, and the output 108. The optical transmitter 1102 includes an input 1104 that is connected to the output 108 of the encoder 100 and an output 1106 that is connected or connectable to an optical fiber. The optical transmitter 1102 is configured to modulate an optical carrier signal based on a modulation symbol sequence, and output the modulated optical carrier signal at the output 1106 of the optical transmitter 702. The optical transmitter 1102 may be a pulse amplitude modulation (PAM) transmitter, a quadrature amplitude modulation (QAM) transmitter, a binary phase shift keying (BPSK) modulation transmitter, and a quadrature phase shift keying (QPSK) modulation transmitter.

[0071] With reference to FIG. 1, FIG. 12 is a flow diagram of a method of operating the encoder 100 for fiber-optic communication in accordance with an implementation of the present disclosure. At a step 1202 , the information bit sequence 110 is received at the input 102 of the encoder 100. At a step 1204 , the information bit sequence 110 is translated into the modulation symbol sequence 112 such that the modulation symbol sequence 112 has, among all symbol sequences in a candidate set of symbol sequences, a smallest value of a metric. At a step 1206 , the modulation symbol sequence 112 is outputted at the output 108. The modulation symbol sequence 112 may be provided as an input to the optical transmitter for modulating the optical carrier signal.

[0072] Although the present disclosure and its advantages have been described in detail, it should be understood that various changes, substitutions, and alterations can be made herein without departing from the scope of the present disclosure as defined by the appended claims.

Claims

1. An encoder (100) for fiber-optic communication, comprising: an input (102) for receiving an information bit sequence (110, 302); a digital signal processor, DSP, (106) configured to translate the information bit sequence (110, 302) into a modulation symbol sequence (112, 310, 400, 700) for modulating an optical carrier signal; and an output (108) configured to output the modulation symbol sequence (112, 310, 400, 700), wherein the DSP (106) is configured to translate the information bit sequence (110, 302) into the modulation symbol sequence (112, 310, 400, 700) such that the modulation symbol sequence (112, 310, 400, 700) has, among all symbol sequences in a candidate set of symbol sequences, a smallest value of a metric, wherein the symbol sequences in the candidate set are different representations of the information bit sequence (110, 302) in a symbol space, wherein the metric is defined as S = ∑ m ∈ M w XX , m ⋅ S XX , m + w YY , m ⋅ S YY , m + w XY , m ⋅ S XY , m for any symbol sequence Z = (Z[0], ..., Z[N - 1]) in the candidate set, wherein N is a common length of the symbol sequences in the candidate set, Z[k] is a symbol at a position k in the symbol sequence (Z[0], ..., Z[N - 1]), k ∈ {0, ..., N - 1}, the symbol Z[k] comprises one or both of the following: a complex-valued first symbol X[k] related to a first polarization and a complex-valued second symbol Y[k] related to a second polarization, S is a value of the metric, m is a summation index, M is a set of one or more whole numbers in the range of -N + 2 to N - 2, ∑m∈M means summation over all values of m in M, where m=0 to N-1, wXX,m, wXY,m, wYY,m are real-valued weights that depend on the value of m, SXX,m, SXY,m, SXY,m are non-negative numbers defined as S XX , m = ∑ k ∈ K X k X * k + m α , S YY , m = ∑ k ∈ K Y k Y * k + m α , S XY , m = ∑ k ∈ K X k Y * k + m α , | | means an absolute value of a complex number, α is a positive real number, ( )α means "to the power of α", k is a summation index, K is a set of one or more whole numbers, ∑k∈K means summation over all values k in K, and ()* means complex conjugation.

2. The encoder (100) of claim 1, wherein the set K includes at least one number that is different from zero.

3. The encoder (100) of claim 1 or 2, wherein S XX , m = ∑ k = 0 N − 1 − m X k X * k + m α , S YY , m = ∑ k = 0 N − 1 − m Y k Y * k + m α , S XY , m = ∑ k = 0 N − 1 − m X k Y * k + m α , form ∈ {0, ...,N - 2} and wherein S XX , m = ∑ k = − m N − 1 X k X * k + m α , S YY , m = ∑ k = − m N − 1 Y k Y * k + m α , S XY , m = ∑ k = − m N − 1 X k Y * k + m α . for m ∈ {-N + 2, ..., -1}.

4. The encoder (100) of claim 1 or 2, wherein S XX , m = ∑ k = 0 N − 1 X k X * k + m α , S YY , m = ∑ k = 0 N − 1 Y k Y * k + m α , S XY , m = ∑ k = 0 N − 1 X k Y * k + m α , for any m ∈ M, with the following cyclic definitions: X k + p ∗ N = X k , Y k + p ∗ N = Y k , p ∈ − 1 , 0 , 1 .

5. The encoder (100) of any one of claims 1 to 4, wherein the set M is a superset of at least one of the following sets: 0 1 , − 1 , 0 , − 1 , 0 , 1 , 0 1 2 , − 2 , − 1 , 0 , − 2 , − 1 , 0 , 1 , 2 , 0 1 2 3 , − 3 , − 2 , − 1 , 0 , − 3 , − 2 , − 1 , 0 , 1 , 2 , 3 , where M=3.

6. The encoder (100) of any one of claims to 5, wherein α = 1.

7. The encoder (100) of any one of claims to 6, wherein Z[k] comprises only the first symbol X[k] and no further symbol, and the metric is defined as S = ∑ m w XX , m ⋅ S XX , m .

8. The encoder (100) of any one of claims to 7, wherein the metric S is a measure of expected distortion, the expected distortion being distortion of a hypothetical optical carrier signal that is modulated with the symbol sequence Z and transmitted via an optical fiber.

9. The encoder (100) of any one of claims to 8, wherein each of X[k] and Y[k] is complex-valued.

10. The encoder (100) of claim 9, wherein each of X[k] and Y[k] is an amplitude and phase shift keying, APSK, symbol.

11. The encoder (100) of claim 9, wherein each of X[k] and Y[k] is a quadrature-amplitude modulation, QAM, symbol.

12. The encoder (100) of any one of claims to 8, wherein each of X[k] and Y[k] is real-valued.

13. The encoder (100) of claim 12, wherein each of X[k] and Y[k] is a pulse-amplitude modulation, PAM, symbol.

14. A fiber-optic transmission apparatus (700), comprising: the encoder (100) of any one of claims to 13; and an optical transmitter (702), the optical transmitter (702) having an input (704) connected to the output (108) of the encoder (100) and an output (706) connected or connectable to an optical fiber.

15. The fiber-optic transmission apparatus (700) of claim 14, wherein the optical transmitter (702) is configured to: modulate an optical carrier signal based on the modulation symbol sequence (112, 310, 400, 700), and output the modulated optical carrier signal at the output (706) of the optical transmitter (702).