RU2713211C1 - Optical signal receiving device and method - Google Patents

Abstract

FIELD: optical communication technique.SUBSTANCE: invention relates to optical communication and can be used to receive optical signals. According to one embodiment, the optical coherent receiver has a heterodyne configuration, wherein in this configuration, spectral-efficient multi-level DP M-QAM modulation formats, such as DP-16QAM, DP-64QAM and DP-256QAM, in which the optical signal can be presented in form of two polarized modes, each of which is modulated with M-QAM formats, i.e. such as 16QAM, 64QAM and 256QAM, respectively, and digital signal processing methods based on coherent optical communication, which have significant performance advantages compared to direct detection, which enable to recover information on the phase of the optical signal.EFFECT: high efficiency of detecting a received signal and high spectral efficiency of the communication system.18 cl, 29 dwg

Description

The invention relates to optical communication technology and can be used for coherent signal reconstruction in systems in which a signal is represented as several components, each of which is modulated with an M-QAM modulation format such as QPSK, 16QAM or higher-level formats such as 64QAM or 256QAM for example, in fiber-optic systems using DP M-QAM dual polarization modulation formats such as DP-QPSK, DP-16QAM or higher-level formats such as DP-64QAM or DP-256QAM, in which the transmitted optical signal is consists of two components, which are orthogonally polarized modes, each of which is modulated by the M-QAM modulation format, i.e. QPSK, 16QAM or higher-level formats such as 64QAM or 256QAM.

BACKGROUND

Optical signals are actively used in digital communication systems. In fiber optic communication lines (FOCL), the signal generated by the laser is transmitted through the optical fiber. The transmitted information is superimposed on the beam by modulating the laser current or by using external beam modulation by applying voltage to a modulator coupled to the laser source.

The main parameters of the fiber schematically depicted in FIG. 1, its parameters include core diameter 1, sheath diameter 2, optical loss and length, numerical aperture proportional to the sine of the maximum angle between the axis of the fiber and the beam introduced into the fiber, and the root of the difference in the square of the refractive indices of the core and fiber sheath. The core 1 and the sheath 2 of the fiber provide the distribution of radiation, and the outer coating 3 protects the fiber from external influences. The waveguide properties of an optical fiber depend not only on its parameters, but also on the wavelength of the propagating radiation. To take this factor into account, a normalized frequency is introduced, proportional to the diameter of the fiber core, the numerical aperture, and inversely proportional to the radiation wavelength, as described in the article by E.M. Dianov, A.S. Kurkov, “Fiber Optics,” Journal of Physics, No. 23, 2006, which is incorporated herein by reference. The value of the normalized frequency determines, in particular, the mode composition of the radiation in the fiber. The mode corresponds to a stable state of the electromagnetic field of radiation propagating along a certain trajectory inside the fiber. When the normalized frequency is less than 2.4, only one mode propagates in the fiber, and such an optical fiber is called single-mode. With a normalized frequency value greater than 2.4, higher-order modes appear, and such a fiber is called multimode.

With the advent of optical amplifiers, the method of increasing the fiber-optic transmission bandwidth due to the simultaneous transmission of multiple information channels on spectrally separated optical signals through the fiber, called Wave Division Multiplexing, has become widely used. Depending on the frequency interval between the channels and the number of WDM channels, the devices are divided into systems with sparse spectral multiplexing (coarse WDM (CWDM)), dense spectral multiplexing (dense WDM (DWDM)) and wide spectral multiplexing (wide WDM (WWDM)), defined in ITU G.671 and ITU G.694.1 standards of the International Telecommunication Union (ITU-International Telecommunication Unit). In the WDM literature, devices with channel spacing of 50 GHz or less are sometimes referred to as High Dense WDM (HDWDM) devices.

The methods used in fiber optic communication to reconstruct the transmitted signal can be divided into categories of direct and coherent detection. In direct detection, the photodetector receives an optical signal using, for example, a modulation format such as On-Off keying (OOK), and converts it into an electrical signal representing the power of the optical signal, without the possibility of determining the phase or frequency of the carrier. In this case, a further increase in the data transfer rate is achieved through the use of DPSK binary differential phase modulation, which encodes information on the phase difference between two adjacent symbols, as well as the use of higher order modulation formats.

Due to the appearance of erbium fiber amplifiers in the early 90s of the 20th century and the high price of equipment, coherent detection systems did not immediately find their application in fiber-optic communication lines. The methods of coherent detection of fiber-optic systems are based on the use of high-order amplitude and phase modulations, polarization multiplexing, coherent detection by a local heterodyne laser, as well as high-speed CMOS analog-to-digital converters (ADCs). The data digitization rate of such ADCs is comparable to the data transfer rate in fiber-optic communication lines, which allows the use of digital signal processing methods to restore the transmitted signal. Coherent signal reception allows the detection of signals of very low intensity, as well as restoring information about the phase of the optical signal, which requires stable synchronization in phase and / or frequency between the received optical signal and the optical local oscillator (local oscillator, LO, local oscillator, LO), used in a coherent receiver, as described for the special case of a coherent receiver for a signal with one polarization in patent RU 2394377 C1 "CO-OPERATED OPTICAL RECEIVER WITH FEEDBACK CONTROL AND WITH Electrons compensation / correction ", which is incorporated herein by reference. In coherent detection, the received optical signal is mixed with the optical signal of the optical local oscillator, due to which the frequency of the signal is reduced from the frequency of the optical carrier (of the order of 100 THz) to a frequency of usually several GHz. The mixed signals are then detected by a photodetector so that the photocurrent contains a component at a frequency f _IF = f _C -f _{LO of the} difference between the center frequency of the received signal f _C and the frequency f _{LO of the} local oscillator. This difference is known as an intermediate frequency (IF) and contains all the information (amplitude and phase) transmitted by the optical signal. Coherent receivers only see signals that are close in wavelength to the local oscillator, and therefore, when the local oscillator wavelength changes, the coherent receiver works similarly to a tunable filter.

If the signal frequency and the local oscillator frequency are equal, then the detection method is called "homodyne", otherwise the detection method is called "heterodyne". For heterodyne systems, the intermediate frequency must be at least twice the data rate of the optical signal. Homodyne reception requires that the LO LO signal is phase locked to the input optical sine, while LO detection requires synchronization with the received frequency signal. Although homodyne systems can provide higher sensitivity than heterodyne systems, homodyne detection poses more requirements for its implementation, mainly due to the strict requirement for phase synchronization. In heterodyne detection, since the IF is much less than f _C , all information can be restored using standard methods for demodulating a radio signal. At the same time, digital signal processing (DSP), based on coherent optical communication, provides significant performance advantages over direct detection. The need for new modulation formats is due to the characteristics of the channel of fiber-optic communication lines. At relatively low data rates of 10 Gb / s, the OOK modulation format was generally acceptable for DWDM systems with a channel spacing of 50 GHz. However, at a data transfer rate of 40 Gb / s, the spectral width of the signal is already 4 times larger for the OOK modulation format, which leads to losses in the channel spacing of 50 GHz.

Due to the rapid growth of fiber-optic communication lines and the need for more bandwidth, considerable efforts are directed to research with the aim of finding effective multi-level modulation formats. Any digital modulation scheme uses a finite number of different signals to represent digital data. So phase modulation (Phase-shift-keying, PSK) uses a finite number of phase values, each of which is one-to-one corresponds to a given sequence of bits. Usually each phase value encodes the same number of bits, and each such sequence of bits forms the character specified by this phase value. A demodulator designed to recover the sequences of characters used by a modulator determines the phase of the received signal and maps it to its corresponding symbol, thereby restoring the transmitted signal. The receiver compares the phase of the received signal with the reference signal. This technique uses coherent detection and is called coherent phase shift keying (CPSK).

A laser beam propagating in a fiber can be represented as two polarized modes. The modulation format DP-QPSK (quadrature phase shift keying with double polarization, dual-polarization quarature phase-shift keying) uses two orthogonal polarizations of the laser beam, denoted below by X and Y, each of which is modulated by the QPSK format, and thus encodes four bits per symbol. This reduces the spectral width of the signal, making the spectrally efficient DP-QPSK format suitable for DWDM systems with channel spacing of 50 GHz. This format was recommended by the Optical Internetworking Forum (OIF) as a data modulation format in 100Gb / s fiber optic systems.

In recent years, much attention has been paid to high-level modulation formats such as 16-QAM and 64-QAM, capable of achieving high data rates in an optical signal with high spectral efficiency. A combination of advanced DP M-QAM modulation formats such as DP-QPSK, DP-16QAM, DP-64QAM and DP-256QAM, in which the optical signal can be represented as two polarized modes, each of which is modulated with M-QAM formats, t .e. such as 16QAM, 64QAM and 256QAM respectively, with spectral shaping and forward error correction (FEC) coding methods is a necessary requirement to achieve high spectral efficiency of optical communication systems.

Scheme of a coherent optical signal receiver described in Seb J. Savory, "Digital Coherent Optical Receivers: Algorithms and Subsystems.", IEEE Journal of Selected Topics in Quantum Electronics, vol. 16, no. 5, September / October 2010, which is incorporated herein by reference, schematically depicted in FIG. 2, includes four key subsystems:

1) The optical unit 4 for processing the optical input optical signal 5, called a coherent optical receiver (coherent optical receiver).

2) Analog-to-digital converter (ADC, Analog to Digital Converter, ADC) 6.

3) Digital demodulator (digital demodulator) 7, which converts the digitized samples of signals from the ADC into a set of signals with a symbol repetition rate.

4) External receiver (outer receiver) 8, including the block error correction. The optical unit 4 for processing the input optical signal 5 is intended for linear mapping of the optical field into a set of electrical signals. This block is often implemented according to the circuit shown in FIG. 3, including a polarization beam splitter (PBS) 9, dividing the optical signal 5 into two components 10 and 11, a local oscillator 12, a power divider 13, dividing the local oscillator 12 into two components 14 and 15, two 90 ° hybrid (90 ° hybrid, 90 degree hybrid) 16 and 17, each of which mixes component 10 and 11 respectively from the output of the polarization divider 9 and component 14 and 15, respectively, from the output of the power divider 13, due to which the signal frequency decreases from the frequency of the optical carrier to IF, balanced photodetectors 18-21, to which signals are output s 90 ° hybrids 16 and 17. An example of such an optical assembly may serve as a coherent receiver (coherent receiver), described in US Patent 8,295,713 B2 "DUAL STAGE CARRIER PHASE ESTIMATION IN A COHERENT OPTICAL SIGNAL RECEIVER", which is incorporated herein by reference.

There are various options for the implementation of the 90 ° hybrid, for example, described in the publication M. Seimetz; CM. Weinert "Options, Feasibility, and Availability of 2 × 4 90 ° Hybrids for Coherent Optical Systems" JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 24, NO. 3, MARCH 2006, which is incorporated herein by reference. One possible implementation of the 90 ° hybrid is to use a 3-dB coupler system 22 with an additional phase delay of 23 ninety degrees in one of the branches, as shown in FIG. 4. The coupler 22 consists of two parallel waveguides 24, 25 located quite close to each other, so that the energy exchange between them is possible. For the mathematical description of such devices, a transfer matrix is used, as described, for example, in Massood Tabib-Azar's book "Integrated Optics, Microstructures, and Sensors", Kluwer Academic Publishings, Boston, MA (1995), which is incorporated herein by reference links.

Electrical signals from balanced photodetectors 18-21 are fed to the inputs of analog-to-digital converters 26-29 of block 6 shown in FIG. 5. Analog-to-digital converters 26-29 of block 6 convert electrical signals with a sampling frequency into a set of time-quantized quantized signals: the ADC 26 converts the electrical signal from the output of the photodetector 18, the ADC 27 converts the electrical signal from the output of the photodetector 19, the ADC 28 converts the electrical signal from the output of the photodetector 20, the ADC 29 converts the electrical signal from the output of the photodetector 21. High-speed ADCs allow you to digitize the signal at a speed sufficient to completely restore the transmitted data . When transmitting data at a speed of S characters per second, the minimum digitization speed is S hertz. Asynchronous digitization, which allows you to restore the frequency and phase of the sequence of characters, requires a 2S digitization speed. The frequency of the transmitter symbols as a rule differs from the digitization frequency of the ADC receiver, but can be restored by interpolation and rescritization of the digitized signal, performed in block 30 of the digital demodulator 7, following the ADC block. In FIG. Figure 5 shows the location of the subsystems of the demodulator unit 7 for one of the possible implementations, as described in Seb J. Savory, "Digital Coherent Optical Receivers: Algorithms and Subsystems.", IEEE Journal of Selected Topics in Quantum Electronics, vol. 16, no. September 5, 2010.

Optical signals received by the receiver through a standard fiber-optic communication line are distorted by the effects of chromatic dispersion (CD), polarization mode dispersion (PMD), rotation of polarization angles and polarization losses (PDL) . The polarization effects in the fiber lead to the rotation of the beam polarizations so that they are no longer orthogonal at the receiver and do not coincide with the directions of the polarization beam divider 9. As a result of polarization at the output of the polarization beam divider 9, they contain the energies of both polarizations of the transmitted signal, including distortions such as CD and PMD. Due to the fact that each of the polarizations of the transmitted signal contains the corresponding transmitted data, it is necessary not only to compensate for distortions caused by effects such as chromatic dispersion and polarization mode dispersion, but also to separate these data signals from each other. The goal of digital coherent technology is to simultaneously obtain both the amplitude and phase of the modulated signal for each of the polarized modes of the transmitted signal so that a linear digital compensator with a finite impulse response (FIR) can be used to perform chromatic dispersion compensation, restoration of polarization and compensation of polarization mode dispersion in the electric domain.

The digital demodulator unit shown in FIG. 5, in addition to the interpolation and oversampling unit 30, may include such units as a unit 31 for eliminating delays between the signals of different channels (deskew) and orthogonalization (orthogonalization); block 32 static compensation of signal distortions such as chromatic dispersion (CD, chromatic dispersion, CD); unit 33 for dynamic compensation of time-varying effects such as polarization rotation and polarization mode dispersion (PMD), the output signal of which is separated components, polarized modes of the transmitted signal with uncompensated deviations of the frequency and phase of the carrier; a carrier frequency recovery unit 34; a carrier phase reconstruction unit 35, the output signal of which is constellation signals for each of the polarized modes of the transmitted signal with the restored carrier frequency and phase.

As can be seen from FIG. 5, there are many different feedbacks between blocks of a coherent receiver. Some of them, for example, the relationship between phase and carrier frequency 36 are natural, others depend on the algorithms used. For example, feedbacks 37 for character estimates and decoded data are required for algorithms using data aided data (DA), but not for blind algorithms. Similarly, for synchronous digitization of data at a symbol rate, feedback 38 is required from the interpolation and oversampling unit 30 to the analog-to-digital converter unit 6, which is not required for asynchronous signal sampling. Other possible feedbacks include feedback 39 from the dynamic equalizer unit 33 to the static equalizer unit 32, feedback 40 from the carrier frequency recovery unit 34 to the local oscillator unit 12, and also feedback 41 from the carrier phase recovery unit 35 to the dynamic equalizer unit 33 .

The function of the external receiver unit is to optimally decode the demodulated signals to obtain the best estimate of the sequence of bits encoded at the transmitter. This can be done in the form of forward error correction with soft decision making (soft-decision forward error correction) or character evaluation followed by advanced error correction with hard decision making. Such a block, shown schematically in FIG. 6 may include hard decision (HD) or soft decision (SD) calculation blocks 42 and 43, decoders 44 and 45, for example, decoders based on LDPC codes (Low density parity check code, low density code parity checks) calculating the sequence of bits 46 and 47 transmitted over the communication channel, which can then be used in block 48 to decode the transmitted data packet, for example, in the OTU-4 format (Optical Transport Unit, Optical Transport Unit), as described in article C-S. Choi, N. Lee, N. Kaneda, Y.-K. Chen, "Concatenated Non-Binary LDPC and HD-FEC Codes for 100Gb / s Optical Transport Systems", 2012 IEEE International Symposium on Circuits and Systems, which is incorporated herein by reference, as well as 37 character ratings for demodulation block 7.

Orthogonalization performed in block 31 is intended to compensate for imbalances in the amplitudes and phases of demodulated signals arising both from imperfections of 90 ° of hybrids 16 and 17, and other factors such as inaccurate tuning of the polarization beam splitter 9 or deviations in the sensitivity of photodiodes. Such orthogonalization can be performed, for example, using the Gram-Schmidt process described in I. Fatadin, Seb J. Savory and D. Ives "Compensation of Quadrature Imbalance in an Optical QPSK Coherent Receiver", IEEE PHOTONICS TECHNOLOGY LETTERS, VOL. 20, NO. 20, OCTOBER 15, 2008, which is incorporated herein by reference, for valid signals. This algorithm can be generalized to the case of complex signals. Let r _X and r _Y be the digitized signals of two polarizations X and Y, respectively, received at the input of block 31. The orthonormal signal X for the signal r _{X is} written in the form

обозначает комплексно-сопряженную величину.where E {} means the average value, and

denotes a complex conjugate.

Then the orthonormal signal Y for r _{Y is} written as

Where

so that

It is easy to see that for signals X and Y the condition

orthogonality of signals. Thus, orthogonalization in block 31 can be performed as follows:

Chromatic dispersion of the signal E (z, t) = [E _X (z, t), E _Y [z, t)] ^T , where t is time, az is the coordinate along the fiber, in the absence of nonlinear effects can be described by the differential equation

Where

поэтому переданный сигнал может быть восстановлен из сигнала, искаженного хроматической дисперсией. Как показано в публикации S. J. Savory, "Digital filters for coherent optical receivers", Opt. Exp., vol. 16, no. 2, pp. 804-817, 2008, которая включена здесь в качестве ссылки, для принятого сигнала, оцифрованного с интервалом T_ADC секунд, переданный сигнал может быть восстановлен при помощи фильтра с конечной импульсной характеристикой (КИХ, finite impulse response, FIR), чьи коэффициенты h_cd[k] заданы величинамиand β ₂ is the group delay dispersion (GDD), approximately equal to 21nc ² / km for a standard single-mode fiber, J is the imaginary unit. The solution of equation (7) for the fiber length L _total gives that

therefore, the transmitted signal can be reconstructed from a signal distorted by chromatic dispersion. As shown in SJ Savory, "Digital filters for coherent optical receivers", Opt. Exp., Vol. 16, no. 2, pp. 804-817, 2008, which is incorporated herein by reference, for a received signal digitized at an interval of T _ADC seconds, the transmitted signal can be reconstructed using a filter with finite impulse response (FIR), whose coefficients are h _cd [k] are given by

, а ρ=2πβ₂L_total/T² _ADC.where N is the number of samples given by

, and ρ = 2πβ ₂ L _total / T ² _ADC .

Although a filter for dispersion compensation can also be implemented using a shorter filter with an infinite impulse response (IIR), as shown in G. Goldfarb and G. Li, "Chromatic dispersion compensation using digital IIR filtering with coherent detection, "IEEE Photon. Technol. Lett., Vol. 19, no. 13, pp. 969-971, Jul. 1, 2007, which is incorporated herein by reference, a FIR implementation is often preferable because it can be efficiently implemented in the frequency domain, for example, using the overlap-add method, as described, for example, in publication M Kuschnerov, F. Hauske, K. Piyawanno, B. Spinnler, M. Alfiad, A. Napoli, and B. Lankl, "DSP for coherent single-carrier receivers," J. Lightw. Technol., Vol. 27, no. 16, pp. 3614-3622, Aug. 15, 2009, which is used here as a reference. Chromatic dispersion compensation schemes in the time domain 49 and frequency domain 50 are shown in FIG. 7. For block 32, implemented in the form of a circuit 49, the compensation of chromatic dispersion in the signals for X and Y polarizations is performed using filters 51 and 52, respectively. For block 32, implemented in the form of a circuit 50, signals for X and Y polarizations are transferred from the time domain to the frequency domain using FFT (Fast Fourier Transform, Fast Fourier Transform, FFT) filters 53 and 54, respectively, and the chromatic dispersion of these signals is compensated by filters 55 and 56, respectively, after which the signals are again transferred from the frequency domain to the time domain using IFFT (Inverse Fast Fourier Tranform, Inverse Fast Fourier Transform, IFFT) filters 57 and 58, respectively.

The dynamic equalizer unit 33 may be implemented using a set of four filters shown in FIG. 8, as described, for example, in WO 2009/070881 A1 "SIGNAL EQUALIZER IN A COHERENT OPTICAL RECEIVER", which is incorporated herein by reference. Dynamic equalizer 33 performs signal conversion using the inverse-Jones matrix of the channel, defined by the relations

и

- скользящие блоки N отсчетов поляризованных мод сигналаwhere h _XX [k], h _XY [k], h _YX [k] are the coefficients of the FIR filters 59-62 of length N, 63 and 64 are the adders of the signals at the outputs of the filters 59, 60 and 61, 62, respectively,

and

- sliding blocks of N samples of polarized signal modes

- значения сигналов поляризаций на выходе эквалайзера 33. Коэффициенты фильтров 59-62 вычисляются обычно при помощи адаптивных алгоритмов, показанных на фиг. 8 в виде блока 65, на вход которого поступают сигналы

с выходов сумматоров 63, 64, оценки символов по обратной связи 37 и символы 41 после компенсации частоты и фазы несущей с выхода блока 35. Коэффициенты фильтров 59-62 могут вычисляться согласно алгоритму стохастического усредненного градиента (stochastic gradient algorithm)a

- the values of the polarization signals at the output of the equalizer 33. The filter coefficients 59-62 are usually calculated using the adaptive algorithms shown in FIG. 8 in the form of a block 65, the input of which receives signals

from the outputs of adders 63, 64, symbol feedback estimates 37 and symbols 41 after compensation of the frequency and phase of the carrier from the output of block 35. Filter coefficients 59-62 can be calculated according to the stochastic gradient algorithm

обозначает операцию комплесного сопряжения и транспонирования вектора

, как показано, например, в статье T.F. Portela et. al., "Analysis of Signal Processing Techniques for Optical 112 Gb/s DP-QPSK Receivers with Experimental Data", Journal of Microwaves, Optoelectronics and Electromagnetic Applications, Vol. 10, No. 1, June 2011, которая включена здесь в качестве ссылки. Так, для QPSK формата модуляции для вычисления величин ε_X и ε_Y широко используется алгоритм слепого выравнивания (constant modulus algorithm, СМА) для слепого восстановления QPSK сигнала, предложенный Годардом, изложенный в статье D.N. Godard, "Self-recovering equalization and carrier tracking in two-dimensional data communication systems", IEEE Trans. Communications, Vol. Com-28, Nov. 11, 1980, pp. 1867-1875, которая включена здесь в качестве ссылки. В этом случае для сигнала для поляризаций X и Y минимизируются ошибкиwhere μ is the coefficient determining the convergence rate of the algorithm, the quantities ε _X [k] and ε _Y [k] are error signals that control the filter adaptation algorithm, and

denotes the operation of complex conjugation and transposition of the vector

as shown, for example, in TF Portela et. al., "Analysis of Signal Processing Techniques for Optical 112 Gb / s DP-QPSK Receivers with Experimental Data", Journal of Microwaves, Optoelectronics and Electromagnetic Applications, Vol. 10, No. 1, June 2011, which is incorporated herein by reference. So, for the QPSK modulation format for calculating ε _X and ε _Y values, the constant modulus algorithm (CMA) is widely used for blind QPSK signal recovery proposed by Godard, described in DN Godard's article "Self-recovering equalization and carrier tracking in two-dimensional data communication systems ", IEEE Trans. Communications, Vol. Com-28, Nov. 11, 1980, pp. 1867-1875, which is incorporated herein by reference. In this case, for the signal for polarizations X and Y, errors are minimized

- абсолютные значения компонент сигнала

после компенсации искажений и восстановления поляризаций переданного сигнала, а сигнал каждой из поляризаций нормирован и имеет единичную энергию. На фиг. 9 показана диаграмма созвездия (constellation diagram) QPSK после выполнения алгоритма слепого выравнивания. В управляемом решением компенсаторе (Decision Directed Equalizer, DD-EQ) сигналы x_r[k] и y_r[k] 41 с выхода блока 35, представляющие собой сигналы x_out[k] и y_out[k] со скомпенсированными частотой и фазой несущей, и оценки

переданных символов 37 поступают на контур принятия решений 65 такой, что минимизируются ошибкиWhere

- absolute values of the signal components

after distortion compensation and restoration of polarizations of the transmitted signal, and the signal of each of the polarizations is normalized and has a unit energy. In FIG. Figure 9 shows the QPSK constellation diagram after running the blind alignment algorithm. In a solution-controlled equalizer (Decision Directed Equalizer, DD-EQ), the signals x _r [k] and y _r [k] 41 from the output of block 35, which are signals x _out [k] and y _out [k] with compensated frequency and phase carrier, and estimates

the transmitted characters 37 arrive at the decision loop 65 such that errors are minimized

и

могут быть получены как при помощи декодирования мягких решений 42, 43 декодерами 44 и 45, так и вычислением жестких решений 42 и 43, например, при помощи стандартной прямоугольной решетки областей решений созвездия (standard rectilinear grid of decision regions in the constellation), дающей в случае созвездия QPSK оценки символов 37:Grades

and

can be obtained both by decoding soft decisions 42, 43 by decoders 44 and 45, and by calculating hard decisions 42 and 43, for example, using the standard rectangular lattice of decision regions in the constellation, giving the case of the QPSK constellation character rating 37:

where the function csgn (x) is given by the formula

The convergence of this solution to the correct solution is guaranteed if the blind alignment algorithm is previously used to approximate the filter sample values of the dynamic distortion compensation unit 33 of the signal to the required minimum.

In contrast to the above blind algorithms, in the equalizer based on the training sequence (training based equalizer), the equalizer block 65 is tuned by the known training sequence 66 such that the equalizer has complete information about the transmitted data. It is also assumed that the laser frequency band is sufficiently small, so that the phase of the carrier can be considered constant throughout the training sequence. In this case, if the sequences of symbols s _X and s _Y, respectively, were transmitted for the polarizations X and Y, then the compensator minimizes the functions

группа 68 включает точки

а группа 69 - точки

Таким образом, созвездие 16QAM содержит три группы, причем группы 67 и 69 содержат по четыре точки, образующие QPSK созвездия с разными амплитудами. В этом случае предлагается сначала использовать обычный алгоритм СМА для предобработки сигнала для обеспечения правильных начальных условий для сходимости алгоритма адаптивной коррекции искажений сигнала, а затем использовать для каждой компоненты поляризации управляемый радиусом компенсатор искажений (radially directed equalizer), величина ошибки которого вычисляется какFor a high-level modulation format such as 16QAM, the QPSK format blind alignment algorithm can be generalized if you notice that the points of the 16QAM signal constellation can be divided into three groups, the points of each of which lie on a circle of the same radius centered at the center of the constellation, as shown in FIG. 10: group 67 includes dots

group 68 includes dots

and group 69 - points

Thus, the 16QAM constellation contains three groups, and groups 67 and 69 contain four points each, forming a QPSK constellation with different amplitudes. In this case, it is proposed to first use the usual CMA algorithm for signal preprocessing to provide the correct initial conditions for the convergence of the adaptive signal distortion correction algorithm, and then use for each polarization component a radius-controlled distortion compensator (radially directed equalizer), the error value of which is calculated as

where r _X = | x _out | for the polarization X, and r _Y = | y _out | for polarization Y, the signal of each of the polarizations is normalized and has unit energy, and the radius R ₀ is given by the following conditions:

могут быть получены как при помощи декодирования мягких решений 42, 43 декодерами 44 и 45, так и вычислением жестких решений 42 и 43, например, при помощи функции csgn_16QAM(x), определяемой стандартной прямоугольной решеткой областей решений созвездия (standard rectilinear grid of decision regions in the constellation),The diagram of the 16QAM constellation after executing the radius-controlled distortion compensator algorithm is shown in FIG. 11. As noted by Seb J. Savory, “Digital Coherent Optical Receivers: Algorithms and Subsystems.”, IEEE Journal of Selected Topics in Quantum Electronics, vol. 16, no. 5, September / October 2010, the compensator (14) controlled by the solution can also be used for the 16QAM format, provided that the signal is first processed either by the blind alignment algorithm (13) or by the radius-controlled compensator algorithm (18). As with the QPSK constellation, estimates

can be obtained both by decoding soft decisions 42, 43 by decoders 44 and 45, and by calculating hard decisions 42 and 43, for example, using the csgn _16QAM (x) function defined by the standard rectangular grid of decision regions of the constellation regions in the constellation),

as described in I. Fatadin, D. Ives, and Seb J. Savory, "Blind Equalization and Carrier Phase Recovery in a 16-QAM Optical Coherent System", JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 27, NO. 15, AUGUST 1, 2009, which is incorporated herein by reference.

A training sequence based compensator (17) can also be used for the 16QAM format.

на выходе блока 33 имеет формуFor higher-level modulation formats such as 64QAM and 256QAM, the development of its own method of adaptive correction of signal distortions is required. Therefore, one of the important problems in the development of a coherent optical signal receiver is the search for a suitable method for adaptive distortion correction for high-level modulation formats. In Seb J. Savory, "Digital Coherent Optical Receivers: Algorithms and Subsystems.", IEEE Journal of Selected Topics in Quantum Electronics, vol. 16, no. 5, September / October 2010, a separate implementation of frequency recovery units 34 and carrier phase 35 is proposed, since this not only reduces the phase necessary for tracking by the carrier recovery system, but also improves the efficiency of carrier recovery, since many phase recovery schemes give an unbiased estimate only at zero carrier frequency error. If the signal

at the output of block 33 has the form

- сигнал на входе блока восстановления несущей,

- сигнал переданных символов созвездия, а φ[k] и Δƒ - фаза и частота несущей, то задача блока 34 состоит в оценке величины Δƒ, а блока 35 - в оценке величины φ[k]. При этом полученные значения фазы могут использоваться по обратной связи 36, показанной на фиг. 5, для определения частоты несущей. В этой же работе предлагается ряд методов восстановления частоты и фазы несущей для формата модуляции QPSK, основанных на возведении сигнала в четвертую степень, устраняющем модуляцию сигнала. Так частотаwhere T _sym is the character spacing,

- signal at the input of the carrier recovery unit,

is the signal of the transmitted constellation symbols, and φ [k] and Δƒ is the phase and frequency of the carrier, then the task of block 34 is to estimate Δƒ, and block 35 to assess the value of φ [k]. In this case, the obtained phase values can be used by feedback 36 shown in FIG. 5 to determine the carrier frequency. In the same work, a number of methods are proposed for reconstructing the frequency and phase of the carrier for the QPSK modulation format, based on raising the signal to the fourth power, which eliminates signal modulation. So frequency

carrier can be estimated by the formula

or a reverse order formula

возводится в четвертую степень 70, сигнал

на выходе задержки 71 возводится в четвертую степень в блоке 72 и после операции комплексного сопряжения в блоке 73 перемножается с сигналом, полученным в блоке 70, после чего выполняется суммирование полученных значений в блоке 75 и вычисление аргумента результата в блоке 76. В случае алгоритма (23) аргумент от сигнала, полученного в блоке 74, умноженный на коэффициент сходимости в блоке 77, складывается в блоке 78 с предыдущей оценкой частоты несущей Δƒ[k-1], умноженной в блоке 79 на коэффициент (1-μ).where μ is the convergence factor. In FIG. 12 shows carrier frequency recovery schemes corresponding to algorithms (22) and (23). In the case of algorithm (22), the signal

raised to the fourth power of 70, the signal

at the output of the delay 71 is raised to the fourth power in block 72 and after the complex pairing operation in block 73 is multiplied with the signal received in block 70, after which the obtained values are summed in block 75 and the result argument is calculated in block 76. In the case of algorithm (23 ) the argument from the signal received in block 74, multiplied by the convergence coefficient in block 77, is added in block 78 with the previous estimate of the carrier frequency Δƒ [k-1] multiplied in block 79 by a factor (1-μ).

The phase can be estimated by the formula

- весовые коэффициенты. Другой предлагаемый алгоритм оценки фазы имеет видWhere

- weighting factors. Another proposed phase estimation algorithm has the form

для уменьшения влияния шума, далее вычисляется фаза сигнала в блоке 82, после чего для полученной фазы (24) выполняется развертка в блоке 83. Полученная в блоке 84 поправка используется для компенсации фазы несущей символа

в блоке 85 с задержкой сигнала, полученной в блоке 86, так, что на выходе блока 85 получается сигнал

который может быть далее послан на блок внешнего приемника 8. Алгоритму (25) соответствует схема восстановления фазы несущей, в которой блок 82 находится перед блоком 81.In FIG. 13 shows a carrier phase recovery scheme corresponding to algorithm (24) based on the scheme presented by TF Portela et. al., "Analysis of Signal Processing Techniques for Optical 112 Gb / s DP-QPSK Receivers with Experimental Data", Journal of Microwaves, Optoelectronics and Electromagnetic Applications, Vol. 10, No. 1, June 2011, which is incorporated herein by reference. The signal is raised to the fourth power in block 80 (M = 4) to eliminate signal modulation; it is passed through a low-pass filter 81, presented in formulas (24), (25) by weight coefficients

to reduce the effect of noise, the signal phase in block 82 is then calculated, after which the obtained phase (24) is scanned in block 83. The correction obtained in block 84 is used to compensate for the phase of the carrier symbol

in block 85 with a delay of the signal received in block 86, so that at the output of block 85 a signal is obtained

which can then be sent to the external receiver unit 8. Algorithm (25) corresponds to the carrier phase recovery circuit, in which block 82 is located in front of block 81.

Для высокоуровневых форматов модуляции таких, как 64QAM и 256QAM, так же требуется разработка своих методов восстановления частоты несущей. Кроме того для высокоуровневых форматов модуляции требуются свои методы восстановления фазы несущей.By Irshaad Fatadin and Seb J. Savory, "Compensation of Frequency Offset for 16-QAM Optical Coherent Systems Using QPSK Partitioning", IEEE Photonics Technology Letters, 23 (17) 1246-1248, Sep. 2011, which is incorporated herein by reference, is proposed to use groups 67 and 69 of the 16QAM constellation to estimate the carrier frequency of this modulation format. To do this, it is proposed to take N pairs of consecutive symbols such that in each pair each point belongs to either group 67 or 69. Then the carrier frequency is calculated for these pairs in the same way as QPSK modulation according to formula (22), where the kth pair of symbols contains the symbols

For high-level modulation formats such as 64QAM and 256QAM, the development of their own carrier frequency recovery methods is also required. In addition, high-level modulation formats require their own carrier phase recovery methods.

SUMMARY OF THE INVENTION

In an embodiment according to the invention, there is provided a device comprising: an optical unit for processing an input optical signal, comprising a polarization divider dividing the input optical signal into two components, a local oscillator, a power divider dividing the local oscillator signal into two components, two 90 ° hybrids, each which mixes the corresponding component from the output of the polarizing divider of the input optical signal and the corresponding component from the output of the power divider of the local oscillator signal, photodetectors which receive signals from the outputs of 90 ° hybrids; a unit comprising analog-to-digital converters, the inputs of which receive electrical signals from photodetectors, converting electrical signals with a sampling frequency into a set of time-quantized quantized signals; a digital demodulator that receives signals from the outputs of analog-to-digital converters, including blocks such as an orthogonalization block, a block for static compensation of signal distortions such as chromatic dispersion, a block for dynamic compensation of time-varying effects such as polarization rotation and polarization mode dispersion, both for the DP-QPSK modulation format and for high-level DP M-QAM modulation formats such as DP-16QAM, DP-64QAM and DP-256QAM, a carrier frequency recovery unit as for the DP-QPSK modulation format, and for high-level DP M-QAM modulation formats such as DP-16QAM, DP-64QAM and DP-256QAM, a carrier phase recovery unit, the output signal of which is constellation signals for each of the polarized modes of the transmitted signal with the restored carrier frequency and phase both for the DP-QPSK modulation format and for high-level DP M-QAM modulation formats such as DP-16QAM, DP-64QAM and DP-256QAM; an external receiver including an error correction unit.

The use of heterodyne detection gives the advantage that the signal frequency is reduced from the frequency of the optical carrier to the IF, due to which all information can be restored using standard methods of demodulation of the radio signal.

The use of a digital demodulator provides the advantage that digital signal processing based on coherent optical communication provides significant performance advantages over direct detection.

Another advantage of this embodiment according to the invention is that a combination of multi-level modulation formats DP M-QAM such as DP-16QAM, DP-64QAM and DP-256QAM, in which the optical signal can be represented as two polarized modes, each which are modulated by M-QAM formats, i.e. such as 16QAM, 64QAM, and 256QAM, respectively, with spectral filtering and noiseless coding methods FEC is a necessary requirement to achieve high spectral efficiency of optical communication systems.

In another embodiment, the invention provides a method comprising: dividing the input optical signal into two components, dividing the local oscillator signal into two components, mixing the corresponding components of the input optical signal and the local oscillator signal, converting the electrical signals into a set of time-quantized quantized signals, orthogonalization, compensation of signal distortions such as chromatic dispersion, compensation of time-varying effects such as polarization expansion and polarization mode dispersion, both for the DP-QPSK modulation format and for high-level DP M-QAM modulation formats such as DP-16QAM, DP-64QAM and DP-256QAM, carrier frequency recovery as for the DP-QPSK modulation format and for high-level DP M-QAM modulation formats such as DP-16QAM, DP-64QAM and DP-256QAM, carrier phase recovery, both for DP-QPSK modulation format and for high-level DP M-QAM modulation formats, like DP-16QAM, DP-64QAM and DP-256QAM, error correction of the demodulated signal.

These options for implementation and other implied benefits, provided patterns and functions will become apparent when studying the detailed description below.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying figures are included to facilitate understanding of the information presented. They illustrate embodiments and, in combination with the recited material, may help an understanding of the principles of the invention. Other embodiments and other implied advantages, provided patterns and functions will become apparent upon examination of the detailed description that follows. Elements of drawings are not necessarily drawn in real terms in relation to each other. General entities are indicated by the same numbers in the figures.

FIG. 1 shows a schematic representation of an optical fiber,

FIG. 2 shows a diagram of a coherent optical signal receiver.

FIG. 3 shows an optical processing unit of an input optical signal for linearly mapping an optical field into a set of electrical signals.

FIG. 4 shows the structure of a 90 ° hybrid using a 3-dB coupler system with an additional phase delay of 90 degrees 20 in one of the branches, as well as a 3-dB coupler structure.

FIG. 5 shows the elements of the optical unit, the analog-to-digital converter unit, and the components of the digital demodulator unit.

FIG. 6 shows a schematic representation of an external receiver unit.

FIG. 7 shows compensation schemes for chromatic dispersion in the time and frequency domains.

FIG. 8 shows a block diagram of a dynamic equalizer.

FIG. 9 shows a QPSK constellation diagram after executing a blind alignment algorithm, FIG. 10 shows a 16QAM constellation with constellation points divided into groups of points lying on circles of different radii.

FIG. 11 shows a diagram of a 16QAM constellation after executing a radius-controlled equalizer algorithm.

FIG. 12 shows carrier frequency recovery steps for QPSK constellation signals.

FIG. 13 shows carrier phase recovery steps for QPSK constellation signals.

FIG. 14 shows an optical input signal processing unit for linearly mapping an optical field into a set of electrical signals, and an analog-to-digital converter unit, in accordance with one embodiment of the invention.

FIG. 15 shows the location of the subsystems of the digital demodulator unit for one of the possible implementations, in accordance with one embodiment of the invention, FIG. 16 shows a block diagram of a dynamic equalizer in accordance with one embodiment of the invention.

FIG. 17 shows a diagram of a 64QAM constellation at the output of a dynamic equalizer block for polarized mode X for a modulation format DP-64QAM, in accordance with one embodiment of the invention.

FIG. 18 shows a diagram of a 64QAM constellation at the output of a dynamic equalizer block for a polarized mode Y for a modulation format DP-64QAM, in accordance with one embodiment of the invention.

FIG. 19 shows a 64QAM constellation with constellation points divided into groups of points lying on circles of different radii, in accordance with one embodiment of the invention.

FIG. 20 shows a 256QAM constellation with constellation points divided into groups of points lying on circles of different radii, in accordance with one embodiment of the invention.

FIG. 21 shows a 256QAM constellation with constellation points divided into groups of points lying on circles of different radii, in accordance with one embodiment of the invention.

FIG. 22 shows a carrier frequency recovery unit, in accordance with one embodiment of the invention.

FIG. 23 shows a diagram of a 64QAM constellation at the output of a dynamic equalizer block for polarized mode X for a modulation format DP-64QAM after carrier frequency compensation, in accordance with one embodiment of the invention.

FIG. 24 shows a diagram of a 64QAM constellation at the output of a dynamic equalizer block for a polarized Y mode for the DP-64QAM modulation format after carrier frequency compensation, in accordance with one embodiment of the invention.

FIG. 25 shows steps for recovering a carrier phase, in accordance with one embodiment of the invention.

FIG. 26 shows a carrier frequency recovery unit, in accordance with one embodiment of the invention.

FIG. 27 shows carrier phase recovery steps, in accordance with one embodiment of the invention.

FIG. 28 shows a diagram of a 64QAM constellation for polarized mode X for a DP-64QAM modulation format after recovering a carrier phase and eliminating slippage, in accordance with one embodiment of the invention.

FIG. 29 shows a diagram of a 64QAM constellation for polarized mode Y for the DP-64QAM modulation format after recovering the carrier phase and eliminating slippage, in accordance with one embodiment of the invention.

DETAILED DESCRIPTION OF EMBODIMENTS

The figures show a device including: an optical input optical signal processing unit, including a polarizing divider dividing the input optical signal into two components, a local oscillator, a power divider dividing the local oscillator signal into two components, two 90 ° hybrids, each of which mixes the corresponding component from the output of the polarizing divider of the input optical signal and the corresponding component from the output of the polarizing divider of the local oscillator signal, photodetectors, to which the signals from moves 90 ° hybrids; a unit comprising analog-to-digital converters, the inputs of which receive electrical signals from photodetectors, converting electrical signals with a sampling frequency into a set of time-quantized quantized signals; a digital demodulator that receives signals from the outputs of analog-to-digital converters, including blocks such as an orthogonalization block, a block for static compensation of signal distortions such as chromatic dispersion, a block for dynamic compensation of time-varying effects such as polarization rotation and polarization mode dispersion, both for the DP-QPSK modulation format and for high-level DP M-QAM modulation formats such as DP-16QAM, DP-64QAM and DP-256QAM, a carrier frequency recovery unit as for the DP-QPSK modulation format, and for high-level DP M-QAM modulation formats such as DP-16QAM, DP-64QAM and DP-256QAM, a carrier phase recovery unit, the output signal of which is constellation signals for each of the polarized modes of the transmitted signal with the restored carrier frequency and phase both for the DP-QPSK modulation format and for high-level DP M-QAM modulation formats such as DP-16QAM, DP-64QAM and DP-256QAM; an external receiver including an error correction unit. FIG. 2 shows a block diagram of a coherent optical signal receiver consisting of an optical unit 4 for processing an optical input optical signal 5, an analog-to-digital converter 6, a digital demodulator 7, and an external receiver 8 including an error correction unit.

The shown optical unit 4 in some embodiments according to the invention consists of a polarization beam splitter 9, dividing the optical signal 5 into two components 10 and 11, a local oscillator 12, a power divider 13, dividing the local oscillator 12 into two components 14 and 15, two 90 ° hybrids 16 and 17, receiving signals 10 and 11, respectively, from the outputs of the polarization beam splitter 9 and power divider 13, photodetectors 18-21, converting the optical signals at the 90 ° output of hybrids 16 and 17 into electrical signals, as shown in FIG. 14.

The shown block of the analog-to-digital converter 6, in some embodiments according to the invention, consists of analog-to-digital converters 26-29, converting, with a sampling frequency, the electrical signals from the outputs of the photodetectors 18-21, respectively, into a set of time-quantized quantized signals, as shown in FIG. 14. When transmitting data at a speed of S characters per second, the minimum digitization speed is S hertz. Asynchronous digitization, which allows you to restore the frequency and phase of the sequence of characters, requires a 2S digitization speed. The transmitter symbol repetition rate usually differs from the digitizing frequency of the ADC receiver, but can be restored by interpolating and rescriting the digitized signal.

In FIG. 15 shows the location of the subsystems of the digital demodulator unit 7, in accordance with one embodiment of the invention using a set of machine instructions that can be sewn into a digital signal processor (Digital Signal Processor, DSP), or stored in a separate memory that is read out by the main memory a processor or a set of processors from a computer-readable storage medium such as read-only memory (ROM) or another type of hard magnetic disk, optical media, tape or flash drive Yati (flash memory). In the case of a program stored on a separate medium, executing a sequence of instructions in a module causes the processor to perform the steps described above. Embodiments of the present invention are not limited to a specific combination of hardware and software, and the computer program code necessary to implement the invention set forth below can be developed by a person skilled in the art.

The term “machine-readable medium” as used above refers to any machine-encoded medium that provides or participates in providing instructions to a processor. Such media includes, but is not limited to, non-volatile devices, volatile devices, and transmitting devices. For example, non-volatile devices may include optical or magnetic memory disks. Volatile devices include dynamic random access memory (dynamic random access memory, dynamic random access memory, DRAM), which typically includes main memory. Traditional computer readable media are well known and do not need to be described in detail.

The signals from the outputs of the analog-to-digital converters 26-29 of the ADC unit 6 shown in FIG. 14, arrive at the orthogonalization block 31. Orthogonalized signals from the output of block 31 are fed to the input of the chromatic dispersion compensation block 32, performed according to scheme 49 for the time domain or according to scheme 50 for the frequency domain. Signals with residual chromatic dispersion and other distortions, such as polarization mode dispersion, are input to the symbol frequency recovery unit 30. The signal from the symbol frequency recovery unit 30 is supplied to the dynamic equalizer unit 33, which can be implemented using a set of four filters, shown in FIG. 16. In some embodiments of the invention, filter coefficients 59-62 are calculated in block 65 using an adaptive algorithm (12). For this, a signal compensator based on the training sequence 66 can be used for signal preprocessing, in which functions (17) are minimized for the polarizations X and Y, and the CMA algorithm (13) can also be used. For the QPSK format, a solution-controlled compensator defined by expressions (14) can then be used, provided that the blind alignment algorithm is previously used to approximate the filter sample values 59-62 to the required minimum.

For 16QAM and higher-level formats such as 64QAM and 256QAM, a compensator (17) based on training sequence 66 or a compensator based on a blind alignment algorithm CMA (13), as well as for QPSK, can be used for signal pre-processing.

For 16QAM and higher-level formats such as 64QAM and 256QAM, after the filter coefficients 59-62 are close enough to the required minimum, then a radius-controlled distortion equalizer (radially directed equalizer) is used for each polarization component. Errors are minimized for the constellation of 16QAM and higher-level formats such as 64QAM and 256QAM

- весовые коэффициенты, r_X=|x_out|, r_Y=|y_out|, а сигнал каждой из поляризаций нормирован и имеет единичную энергию. Для созвездия 16QAM радиус R₀, вычисляемый для блока 65 в блоках 87 и 88 для поляризаций X и Y, соответственно, задается условиями (19), для созвездия 64QAM радиус R₀ задается условиями:calculated by one or more characters N≥1, where

are weights, r _X = | x _out |, r _Y = | y _out |, and the signal of each of the polarizations is normalized and has a unit energy. For the 16QAM constellation, the radius R ₀ calculated for block 65 in blocks 87 and 88 for the polarizations X and Y, respectively, is specified by conditions (19), for the 64QAM constellation, the radius R ₀ is specified by the conditions:

Information 89, 90 about the belonging of the signal to a specific group is sent together with the signals x _out and y _out to the frequency recovery units 34 and the carrier phase 35.

могут быть получены как при помощи декодирования мягких решений 42, 43 декодерами 44 и 45, так и вычислением жестких решений 42 и 43. Для формата 16QAM в управляемом решением компенсаторе жесткие оценки переданных символов 37 могут быть записаны в видеA compensator controlled by the solution can also be used for 16QAM and higher-level formats such as 64QAM and 256QAM, provided that the signal is first processed either by the blind alignment algorithm (13) or by the radius-controlled compensator algorithm (26). For the 16QAM format and higher-level formats such as 64QAM and 256QAM, in the decision-controlled equalizer, the estimates of the transmitted symbols 37 are sent to the decision loop 65 such that errors are minimized (14). As with the QPSK constellation, estimates

can be obtained both by decoding soft decisions 42, 43 by decoders 44 and 45, and by calculating hard decisions 42 and 43. For the 16QAM format in the decision-controlled compensator, hard estimates of the transmitted symbols 37 can be written as

for 64QAM format - in the form,

for 256QAM format - as

where the functions csgn _l6QAM (x), csgn _64QAM (x), and csgn _256QAM (x) are determined using standard rectangular lattices of constellation solution domains in the same way as (16). For an arbitrary high-level M-QAM modulation format such as 512QAM, 1024QAM and higher, in the decision-controlled equalizer, estimates of the transmitted symbols 37 can be written as

wherein csgn _MQAM (x) function are determined using standard rectangular grating regions constellation solutions analogously to (16), (29) - (31).

The radius-controlled compensator algorithm can also be constructed for any high-level M-QAM modulation format such as 512QAM, 1024QAM and higher. For this, for a constellation normalized to unit energy, it is necessary to divide the constellation points into N ^MQAM groups, g = 1, ..., N ^MQAM , so that the points of each group g lie on a circle of the same radius R ₀ [g] centered at the center of the constellation , and (R ₀ [g]) ² > (R ₀ [g + 1]) ² , and there are no other points of the constellation between circles of radii R ₀ [g] and R ₀ [g + 1]. Then, for each group of g points, the square of r [g] will be determined as the half-sum of squares of radii (r [g]) ² = 0.5 ((R ₀ [g]) ² + (R ₀ [g + 1]) ² ). For such an M-QAM constellation, the radius-controlled compensator algorithm will have the form

For the constellation QPSK, obviously, N ^QPSK = 1.

содержат по четыре точки, образующие QPSK созвездия с разными амплитудами. На фиг. 19 показано получающееся разбиение на группы точек созвездия 64QAM такое, что группа 93 включает точки

группа 94 включает точки

группа 95 включает точки

группа 96 включает точки

группа 97 включает точки

группа 98 включает точки

а так же точки

группа 99 включает такие точки, как

группа 100 включает точки созвездия

группа 101 включает точки

Таким образом, созвездие 64QAM разбивается на девять групп, точки каждой из групп лежат на окружности одного радиуса, причем три группы 93, 95 и 101, нумеруемые индексами

содержат по четыре точки, образующие QPSK созвездия с разными амплитудами. На фиг. 20 показано получающееся разбиение на группы точек созвездия 256QAM такое, что группа 102 содержит точки созвездия

группа 103 содержит точки

группа 104 содержит точки

группа 105 содержит точки

группа 106 содержит точки созвездия

группа 107 содержит точки созвездия

группа 108 содержит точки созвездия

как показано на фиг. 21. Таким образом, созвездие 256QAM разбивается на тридцать две группы, причем семь групп 102-108, нумеруемых индексами

содержат по четыре точки созвездия, образующие QPSK созвездия с разными амплитудами. Для произвольного созвездия M-QAM группы, которые содержат по четыре точки созвездия, образующие QPSK созвездия с разными амплитудами, нумеруются индексами

Для созвездия

For high-level modulation formats such as 64QAM and 256QAM, as well as for the 16QAM modulation format, the signal at the output of dynamic equalizer 33 is a polarized mode of the transmitted signal that is separated from each other. In FIG. 17 and 18 show diagrams of constellations of polarized modes of the signal 91 and 92 for the case of the modulation format DP-64QAM. FIG. 19 and FIG. 20 show the constellation points 64QAM and 256QAM, respectively. As a result of the action of the radius-controlled distortion compensator, the signals of the polarized modes at the output of the equalizer 33 are divided into groups. The 16QAM constellation points are divided into three groups 67-69, as shown in FIG. 10, moreover, two groups 67 and 69, numbered by indices

contain four points forming the QPSK constellation with different amplitudes. In FIG. 19 shows the resulting grouping of 64QAM constellation points such that group 93 includes points

group 94 includes dots

group 95 includes dots

group 96 includes dots

group 97 includes dots

group 98 includes dots

as well as points

group 99 includes points such as

group 100 includes constellation points

group 101 includes dots

Thus, the constellation 64QAM is divided into nine groups, the points of each group lie on a circle of the same radius, and the three groups 93, 95 and 101, numbered by indices

contain four points forming the QPSK constellation with different amplitudes. In FIG. 20 shows the resulting grouping of 256QAM constellation points such that group 102 contains constellation points

group 103 contains points

group 104 contains dots

group 105 contains points

group 106 contains constellation points

group 107 contains constellation points

group 108 contains constellation points

as shown in FIG. 21. Thus, the constellation 256QAM is divided into thirty-two groups, with seven groups 102-108 numbered by indices

contain four constellation points forming QPSK constellations with different amplitudes. For an arbitrary M-QAM constellation, groups that contain four constellation points each forming QPSK constellations with different amplitudes are numbered by indices

For constellation

для поляризации

для поляризации Y таких, что в каждой паре каждая из точек принадлежит одной из

групп,

содержащих по четыре точки созвездия, образующих QPSK созвездия с разными амплитудами. В блоке 111 по отобранным парам символов 110 частота несущей 112 оценивается по формуламIn some embodiments of the invention, the signal x _out , y _out from the output of the dynamic equalizer 33, together with information about splitting the signal into groups 89, 90, is then sent to block 34, schematically shown in FIG. 22, in which carrier frequency recovery is performed. The signals x _out , y _out for the polarizations X and Y are sent to block 109, in which N pairs of 110 consecutive symbols are selected for the M-QAM constellation

for polarization

for polarization Y such that in each pair each of the points belongs to one of

groups

containing four constellation points, forming QPSK constellations with different amplitudes. In block 111 on selected pairs of characters 110, the carrier frequency 112 is estimated by the formulas

on the basis of which the carrier frequency can be estimated, for example, as

- весовые коэффициенты, илиWhere

- weighting factors, or

or formulas with the reverse order of operations

on the basis of which the carrier frequency can be estimated, for example, as

- весовые коэффициенты, илиWhere

- weighting factors, or

For the QPSK constellation, the result coincides with (22), (23).

для поляризации X и сигналы

для поляризации Y такие, что каждый из них принадлежит одной из

групп,

содержащих по четыре точки созвездия, образующих QPSK созвездия с разными амплитудами. В блоке 111 выполняется компенсация частоты несущей по формуламIn some embodiments according to the invention, for the 16QAM format and higher level modulation formats such as 64QAM and 256QAM, in block 109, all signals are selected for the M-QAM constellation

for polarization X and signals

for polarization Y such that each of them belongs to one of

groups

containing four constellation points, forming QPSK constellations with different amplitudes. In block 111, the carrier frequency compensation is performed according to the formulas

и

концентрируются в областях точек созвездий, например 113 и 114 в случае созвездия 64QAM, как показано на фиг. 23 и фиг. 24 соответственно, изображенных на фиг. 15 в виде объектов 115 и 116 на фоне блока 34.selection of the carrier frequency Δƒ 112, which maximizes the concentration of symbols in the regions of the constellation points of the polarized modes of the transmitted signal, rotated by the value of the uncompensated carrier phase error so that the symbols

and

are concentrated in the regions of the constellation points, for example 113 and 114 in the case of the 64QAM constellation, as shown in FIG. 23 and FIG. 24, respectively, depicted in FIG. 15 in the form of objects 115 and 116 against the background of block 34.

In some embodiments of the invention, the signal from the output of the dynamic equalizer 33, together with the estimate of the carrier frequency 112, is fed to the carrier phase recovery unit 35.

символов 37, полученные по обратной связи от внешнего приемника 8, а так же пилотные символы x_plt, y_plt 119.In some embodiments of the invention for the QPSK format, as well as for the 16QAM modulation format and higher-level formats such as 64QAM and 256QAM, the carrier phase is restored according to the circuit shown in FIG. 25. In addition to the blocks shown in FIG. 13, it contains a character sampling unit 115, which sends to the phase 82 calculation unit information about the number of the group 116 to which the received signal belongs and the reference signal number 117 of the received signal, the calculation unit for the initial reference of phase 118 using estimates

characters 37, obtained by feedback from the external receiver 8, as well as pilot characters x _plt , y _plt 119.

для поляризации X, символы

для поляризации Y, пришедшие в моменты времени t[k], принадлежащие группам

содержащим по четыре точки созвездия, образующие QPSK созвездия с разными амплитудами. Как и в схеме на фиг. 13, эти символы в блоке 80 возводятся в четвертую степень (М=4) для устранения модуляции сигнала, в блоке 82 вычисляется фаза сигнала по формулеAt block 115, characters are selected

for polarization X, characters

for polarization Y, arriving at time t [k], belonging to the groups

containing four constellation points forming the QPSK constellation with different amplitudes. As in the circuit of FIG. 13, these symbols in block 80 are raised to the fourth degree (M = 4) to eliminate signal modulation, in block 82, the phase of the signal is calculated by the formula

- весовые коэффициенты, либо по формулеWhere

- weights, or by the formula

then, in block 83, a phase scan is performed. In addition to the phase raised to the fourth power, block 82 receives from block 115 information about the number of group 116 to which the received signal belongs and the reference number 117 of the received signal, necessary for calculating expressions (41), (42). For the QPSK constellation, expressions (41), (42) are reduced to (24), (25). The block for calculating the initial phase reference 118 uses averaging over 2N + 1 estimates of symbols 37 and pilots 119 to calculate the initial phase reference specified by the formula

or formula

- весовые коэффициенты. В результате оценка фазы, посылаемая из блока 83 на блок 84, задается выражениямиWhere

- weighting factors. As a result, the phase estimate sent from block 83 to block 84 is given by the expressions

Thus, the characters x _r [k], y _r [k] sent to the external receiver unit 8 and to the equalizer unit 33 via communication 41 are calculated by the formulas

выполняется компенсация частоты несущей (40), и фаза несущей вычисляется, какIn some embodiments of the invention, blocks 115 and 82 also receive an estimate of Δƒ of the carrier frequency 112. For each symbol

carrier frequency compensation (40) is performed, and the carrier phase is calculated as

either by the formulas

In some embodiments of the invention, symbols (40) are immediately calculated in the carrier frequency recovery unit 34, which in this case is shown in FIG. 26. The signal (40) indicated in the block diagram 34 in FIG. 26 by an arrow 120, is sent to the block 115 of the carrier phase recovery unit 35, shown schematically in this case in FIG. 27. In FIG. Figures 28 and 29 show diagrams of the constellations of the polarized modes of signal 121 and 122 for the case of the DP-64QAM modulation format at the output of block 35. The signal from the output of block 35 is then sent to an external receiver 8 and to the dynamic equalizer block 33. Feedback symbols received by block 8 37 are supplied to the dynamic equalizer unit 33 and the frequency and phase carrier recovery units 34 and 35.

The methods disclosed herein are not limited to the field of optical communication systems, and are used in other areas where a coherent receiver is used as a receiving / detecting device.

The above detailed description should be understood in all respects as explanatory and illustrative, but not as limiting, and the scope of the invention described here should not be determined from the description of the invention, but from the claims, interpreted in accordance with the full scope permitted by patent laws. It should be understood that the described embodiments are merely illustrative of the principles of the present invention, and that various modifications can be made by specialists without departing from the spirit and scope of the legal protection of this invention as set forth in the appended claims.

Claims (18)

1. An optical signal receiving device including an optical input optical signal processing unit including a polarization divider dividing the input optical signal into two components, a local oscillator, a power divider dividing the local oscillator signal into two components, two 90 ° hybrids, each of which mixes the corresponding component from the output of the polarizing divider of the input optical signal and the corresponding component from the output of the polarizing divider of the local oscillator signal, photodetectors, which are received from needles from the outputs of 90 ° hybrids, a unit that includes analog-to-digital converters, the inputs of which receive electrical signals from photodetectors, converting electrical signals with a sampling frequency into a set of time-quantized quantized signals, a digital demodulator, to which signals from the outputs of analog-digital converters, including blocks such as an orthogonalization block, a block of static compensation for signal distortions, such as chromatic dispersion, a block of dynamic compensation of time-varying In addition to effects such as polarization rotation and polarization mode dispersion, for both the DP-QPSK modulation format and the high-level DP M-QAM modulation formats, such as DP-16QAM, DP-64QAM and DP-256QAM, a carrier frequency recovery unit like for the DP-QPSK modulation format and for high-level DP M-QAM modulation formats such as DP-16QAM, DP-64QAM and DP-256QAM, a carrier phase recovery unit whose output signal is constellation signals for each of the polarized modes of the transmitted the signal with the restored frequency and phase of the carrier, as for DP-QPSK modulation format, as well as for high-level DP M-QAM modulation formats, such as DP-16QAM, DP-64QAM and DP-256QAM, an external receiver including an error correction block. 2. The device according to claim 1, characterized in that the signal from the symbol frequency recovery unit is fed to the dynamic equalizer unit, which can be implemented using a set of four custom filters. 3. The device according to claim 1, characterized in that the dynamic equalizer block contains a block in which for any high-level modulation format M-QAM, the constellation points are divided into groups so that the points of each group lie on a circle of the same radius with a center in the center of the constellation, and samples of signals of polarized modes are distributed among these groups. 4. The device according to claim 1, characterized in that the carrier frequency recovery unit includes a unit in which pairs of consecutive symbols are selected such that in each pair each of the points belongs to one of the groups containing four constellation points forming QPSK constellations with different amplitudes, as well as the unit in which these pairs are used to restore the carrier frequency. 5. The device according to claim 1, characterized in that the carrier frequency recovery unit includes a unit in which pairs of consecutive symbols are selected such that in each pair each of the points belongs to one of the groups containing four constellation points forming QPSK constellations with different amplitudes, as well as a unit in which these pairs are used to restore the carrier frequency and to compensate for the carrier frequency in the signal received from the dynamic equalizer unit, to use the received symbols in the phase recovery unit s. 6. The device according to claim 1, characterized in that the carrier phase recovery unit includes a block for selecting symbols belonging to one of the groups containing four points of the constellation, forming a QPSK constellation with different amplitudes. 7. A method of receiving an optical signal, including dividing the input optical signal into two components, dividing the local oscillator signal into two components, mixing the corresponding components of the input optical signal and the local oscillator signal, converting the electrical signals into a set of time-quantized quantized signals, demodulating the signal, including orthogonalization, compensation of signal distortions, such as chromatic dispersion, compensation of time-varying effects, such as time polarization enhancement and polarization mode dispersion for both the DP-QPSK modulation format and high-level DP M-QAM modulation formats such as DP-16QAM, DP-64QAM and DP-256QAM, carrier frequency recovery for both the DP-QPSK modulation format and for high-level DP M-QAM modulation formats such as DP-16QAM, DP-64QAM and DP-256QAM, carrier phase recovery for both the DP-QPSK modulation format and high-level DP M-QAM modulation formats, such as DP-16QAM, DP-64QAM and DP-256QAM, error correction of the demodulated signal. 8. The method according to claim 7, characterized in that the filter coefficients of the dynamic equalizer block are calculated using an adaptive algorithm in which a compensator based on the training sequence or CMA algorithm can be used to preprocess the signal. 9. The method according to p. 7, characterized in that for the QPSK format, after processing the signal to calculate the filter coefficients of the dynamic equalizer unit, a compensator controlled by the solution can be used. 10. The method according to claim 7, characterized in that for the 16QAM format and higher-level modulation formats, such as 64QAM and 256QAM, after the signal preprocessing, a radius-controlled compensator can be used to calculate the filter coefficients of the dynamic equalizer block. 11. The method according to claim 7, characterized in that for the 16QAM format and higher level modulation formats, such as 64QAM and 256QAM, after preprocessing the signal with a blind equalization algorithm or a radius-controlled compensator algorithm, a solution-controlled compensator can be used to calculate the filter coefficients of the dynamic equalizer block . 12. The method according to p. 7, characterized in that for any high-level modulation format M-QAM after compensation for chromatic dispersion and dynamic compensation for time-varying effects, such as polarization rotation and polarization mode dispersion, the signal samples are divided into groups such that each the group contains four constellation points forming the QPSK constellation with different amplitudes. 13. The method according to p. 7, characterized in that in the recovery unit of the carrier frequency, pairs of consecutive symbols are selected such that in each pair each of the points belongs to one of the groups containing four constellation points, forming QPSK constellations with different amplitudes, and these symbol pairs calculate the carrier frequency. 14. The method according to p. 7, characterized in that in the carrier frequency recovery unit, all symbols belonging to one of the groups containing four constellation points are selected, forming QPSK constellations with different amplitudes, and the carrier frequency is selected from these symbols to maximize the concentration of these symbols in the regions of the constellation points of the polarized modes of the transmitted signal, rotated by the value of the uncompensated phase of the carrier. 15. The method according to p. 7, characterized in that in the recovery unit of the carrier frequency in the signal received from the dynamic equalizer unit, the carrier frequency is compensated. 16. The method according to p. 7, characterized in that in the block recovery phase of the carrier are selected characters belonging to groups containing four points of the constellation, forming QPSK constellations with different amplitudes, which are raised to the fourth degree to eliminate signal modulation and are used to calculate the phase carrier. 17. Non-volatile computer readable medium that stores instructions of a computer program for restoring a received signal, mainly converted from optical to electrical signal, which (program), being executed by the processor, enables the device to perform the signal demodulation method according to any one of claims. 7-16. 18. A digital signal processor that stores instructions of a computer program for restoring a received signal, mainly converted from an optical signal to an electric signal, which, when executed by the processor, enables the device to perform the signal demodulation method according to any one of claims. 7-16.

Images