Method and apparatus for channel carrier phase recovery

Abstract

The present application relates to a method and apparatus for channel carrier phase recovery. Compared with the traditional signal hard decision based carrier phase recovery method and apparatus, the carrier phase recovery method and apparatus of the present application is based on signal soft decision, effectively improving the tolerance of the phase estimation process to system noise.

Description

Technical Field

[0001] This application relates to the field of communications, and more particularly to methods and apparatus for channel carrier phase recovery. Background Technology

[0002] With the rapid development of internet technology, high-speed coherent optical communication technology based on high-order modulation techniques has attracted much attention. In practical applications, high-order modulation methods are highly sensitive to phase noise that may exist in the channel. Therefore, channel phase estimation is one of the core steps in receiver data processing in classical coherent optical communication.

[0003] Traditional communication systems often employ the Viterbi-Viterbi phase estimation method to recover the channel carrier phase. However, this method is extremely limited in its application in amplitude-phase dual modulation systems such as quadrature amplitude modulation. Furthermore, to simplify the phase estimation process, both the traditional decision-assisted maximum likelihood (DA-ML) phase estimation method and the traditional Viterbi-Viterbi method assume that the phases of adjacent signals in the received signal remain constant or change slowly. This assumption significantly limits the use of Viterbi-Viterbi and DA-ML methods in systems with high phase noise. Additionally, in the traditional DA-ML phase estimation method, the phase estimate at the current moment is based on the hard decision results of signals from past moments. Therefore, in practical applications of DA-ML modules, accumulated errors are easily generated, which in turn affects the estimation accuracy of subsequent signals. Summary of the Invention

[0004] According to one aspect, the present invention provides a method for channel carrier phase recovery, the method comprising: receiving a signal from a pre-channel estimation module; performing time-domain decomposition on the received signal to construct an estimation window; calculating the carrier phase at the current moment using soft-decision information of sample signals within the estimation window to recover the carrier phase; wherein if the received signal has not yet ended, the estimation window is slid backward by one signal period, such that the center position of the estimation window falls at the next moment, and the carrier phase at the next moment is calculated, until the received signal ends.

[0005] Preferably, calculating the carrier phase at the current moment includes calculating the phase noise estimate of the current signal based on the soft-decision maximum likelihood phase estimate of the sample signal within the estimation window. To obtain the carrier phase at the current moment.

[0006] Preferably, the estimation window is a symmetrical estimation window with a length of (2L+1) centered on the received signal at the current time, where L is a positive integer greater than 0, and the estimated phase noise value of the current signal is... It is determined by the following formula:

[0007]

[0008] Where k is the signal sequence number, This indicates the originating signal within the current estimation window. Denotes the real part of x. Denotes the imaginary part of x.

[0009]

[0010] Where K represents the received signal sample The cross-correlation matrix has a length of (2L+1)×(2L+1), and the elements in the cross-correlation matrix are determined by the following formula:

[0011]

[0012] Where N represents the set and Size of the intersection δ(x) represents the phase noise variance, δ(x) represents the Dirac function with respect to x, and N0 is the two-sided power spectral density of additive white Gaussian noise.

[0013] Preferably, calculating the carrier phase at the current moment further includes obtaining a phase estimation result through iteration and performing phase recovery.

[0014] Preferably, the iteration includes:

[0015] (a1) Set the threshold ε for determining the end of the iteration. θ ;

[0016] (a2) Determine the iteration starting point and select the final phase estimate of the sample corresponding to the previous moment of the currently received signal sample. As the initial value for the current phase estimation iteration;

[0017] (a3) Based on the phase estimation of the (n+1)th iteration Phase estimation with the nth iteration The relationship is used to start the iteration, and the phase estimate obtained in the (n+1)th iteration is calculated. in

[0018] as well as

[0019] (a4) when When the time is reached, the iteration ends; the current iteration phase estimate is set. As the final estimate of the phase Right now

[0020] Preferably, the initial pilot phase estimate of the received signal is...

[0021]

[0022] Where m(l) is the pilot signal, N p It represents the number of pilot signals.

[0023] According to another aspect, the present invention provides an apparatus for channel carrier phase recovery, the apparatus being configured to perform the aforementioned method for channel carrier phase recovery.

[0024] According to another aspect, the present invention provides a receiver for a communication system, the receiver comprising a pre-channel estimation module, a phase recovery module and a demodulation module connected in sequence, wherein the phase recovery module is the aforementioned device for channel carrier phase recovery. Attached Figure Description

[0025] Figure 1 This is a block diagram of a coherent optical communication receiver;

[0026] Figure 2 This is a flowchart of a phase recovery method according to an embodiment of this application;

[0027] Figure 3 It is an estimation window diagram constructed according to an embodiment of this application; and

[0028] Figure 4 This is a schematic diagram of the phase recovery result according to an embodiment of the present application, illustrating the impact of different levels of phase noise and signal-to-noise ratio on phase estimation performance. Detailed Implementation

[0029] Figure 1 This is a schematic block diagram of a coherent optical communication receiver. For example... Figure 1As shown, signal light 111 and local oscillator light 112 are split into polarized light signals 130A and 130B with two polarization directions after passing through polarization beam splitter (PBS) 110. Polarized light signals 130A and 130B are converted into corresponding electrical signals 131A and 131B by photodetector 140; the first electrical signal 131A and the second electrical signal 131B are processed by pre-channel estimation module 120 into corresponding electrical signals 133A and 133B; electrical signals 133A and 133B are received and processed by phase recovery modules 190A and 190B respectively to recover the carrier phase, and then the phase-recovered signals 135A and 135B are transmitted to signal demodulation module 192 by phase recovery modules 190A and 190B, and output after demodulation by demodulation module 192. Generally, the pre-channel estimation module 120 may include a clock recovery module 150, a dispersion compensation module 160, a polarization demultiplexing module 170, and a frequency offset compensation module 180 connected in sequence.

[0030] Indicative, and not restrictive, in Figure 1 In the diagram, clock recovery module 150 is shown as including 150A and 150B, dispersion compensation module 160 is shown as including 160A and 160B, and frequency offset compensation module 180 is shown as including 180A and 180B. This is primarily to demonstrate that the pre-channel estimation module 120 processes two electrical signals 131A and 131B simultaneously. Similarly, phase recovery modules 190A and 190B are also primarily to demonstrate that phase recovery is performed simultaneously on two electrical signals 133A and 133B. In other words, the data processing for the polarized optical signals 130A and 130B in two polarization directions is consistent.

[0031] Figure 2 This is a flowchart illustrating a phase recovery method 200 performed by phase recovery modules 190A and 190B according to an embodiment of this application. The phase recovery method 200 includes:

[0032] In step 210, a signal from the front-end channel estimation module is received.

[0033] Next, in step 220, the received signal is decomposed in the time domain to construct an estimation window.

[0034] In one example, a symmetric estimation window sample of length (2L+1) centered on the received signal at the current time can be constructed. Where r(l) is the sample value of the received signal at time l, and L is a positive integer greater than 0. Figure 3 An estimation window constructed in this way is shown. This is an example, not a limitation, of... Figure 3 In this context, L is 2. For example... Figure 3 As shown, the transmission frame structure contains N p Pilot signals 310 and N mFor a data signal 320, at time k, a symmetrical estimation window of length 5 centered on the currently estimated signal 330 is 340; at time k+1, the estimation window of length 350 remains 5, but is symmetrically distributed centered on the data signal received at time k+1. Here, for ease of analysis, this application introduces the received signal as shown in formula (1):

[0035] r(k)=m(k)e jθ(k) +n(k) Formula (1)

[0036] Where r(k) represents the received signal of phase recovery modules 190A and 190B. Represents the transmitter information, ρ(k) represents the amplitude modulation coefficient, and E S Let φ(k) represent the signal power, φ(k) represent the transmitter phase modulation, θ(k) = θ(k-1) + v(k) represent the carrier phase, and v(k) represent the carrier phase noise, which has been shown to follow a mean of 0 and a variance of 0. The signal follows a normal distribution, Δv is the 3-dB linewidth of the laser at the transmitting and receiving ends, T is the period of the received signal, n(k) represents additive white Gaussian noise, and the two-sided power spectral density of the noise is N0.

[0037] To further simplify the analysis, this application introduces a linearly simplified received signal as shown in formula (2):

[0038]

[0039] Because it is a linearly simplified signal, given the transmitted signal m(l) and the carrier phase θ(k) at time k, the l-th received signal r(l) within the estimation window at time k follows a Gaussian distribution with mean m(l).

[0040] Based on all possible combinations of transmitted signals within the current estimation window, this application uses the maximum likelihood phase estimation algorithm to construct the theoretical model formula (3) for phase estimation:

[0041]

[0042] Where M represents the modulation coefficient. Based on the conclusions drawn from the simplified linear model above, the probability density function of the sample is as shown in formula (4):

[0043]

[0044] Formula (4) and Substituting the element expressions in the cross-correlation matrix into formula (3), we obtain formula (5):

[0045]

[0046] By solving formula (5), the maximum likelihood phase estimate based on signal soft decision can be obtained, that is, the phase calculated by formula (6) below.

[0047]

[0048] Where k is the signal sequence number, This is the estimated phase noise value of the current signal. This indicates the originating signal within the current estimation window. Denotes the real part of x. Denotes the imaginary part of x. It can be obtained from formula (7).

[0049]

[0050] Where K represents the received signal sample The cross-correlation matrix has a length of (2L+1)×(2L+1), and the elements in this matrix can be written as formula (8):

[0051]

[0052] Where N represents the set and Size of the intersection δ(x) represents the phase noise variance, δ(x) represents the Dirac function with respect to x, and N0 is the two-sided power spectral density of additive white Gaussian noise.

[0053] Now turn back Figure 2 and combined Figure 3 In step 230, the carrier phase at the current moment is calculated using the soft decision information of the sample signals within the estimation window to recover the carrier phase. In one example, the phase at the current moment is estimated using the maximum likelihood estimation method (ML) based on the soft decision of the received signal within the current estimation window, and carrier phase recovery is performed.

[0054] Specifically, the above-mentioned symmetric estimation window samples can be used as a reference. Substitute the values ​​into formula (6) to calculate the phase estimation result. Simultaneously, obtain the phase estimation result through an iterative method to perform phase recovery.

[0055] In one embodiment of the present invention, the iterative method includes the following steps (a1) to (a4):

[0056] (a1) Set the threshold ε for determining the end of the iteration. θ .

[0057] (a2) Determine the iteration starting point and select the final phase estimate of the sample corresponding to the previous moment of the currently received signal sample. This serves as the initial value for the current phase estimation iteration.

[0058] (a3) Based on the phase estimation of the (n+1)th iteration Phase estimation with the nth iteration The relationship is used to start the iteration, and the phase estimate obtained in the (n+1)th iteration is calculated. That is, the following formula (10):

[0059] Specifically, the phase estimate obtained in the (n+1)th iteration can be calculated as follows: That is, the following formula (10)

[0060] First, transform formula (6) into formula (6B):

[0061]

[0062] From the iterative formula (9)

[0063] x n+1 =f(x) n ) Formula (9)

[0064] It can be seen that phase estimation That is, x in the step iteration formula. This is f(x) in the step iteration formula.

[0065] Substituting formulas (6) and (6B) into formula (9), we obtain the iterative estimation formula (10) for the phase estimation model:

[0066]

[0067] (a4) when When the time is reached, the iteration ends; the current iteration phase estimate is set. As the final estimate of the phase, the final estimate of the iterative phase is written as... Right now

[0068] Due to calculation Need to get And calculation Need to get And so on, the initial pilot phase estimate is... This can be derived from the following formula (11):

[0069]

[0070] Where m(l) is the pilot signal, N p It represents the number of pilot signals.

[0071] Then, in step 240, it is determined whether the signal has ended. If the signal has not ended, step 250 is executed, and the estimation window is shifted backward by one signal cycle, so that the center of the estimation window falls at the next time step. For example, as... Figure 3 As shown, the estimation window 340 at time k is slid backward by one signal cycle to the estimation window 350 at time k+1.

[0072] Next, step 230 is executed for the carrier phase at the next moment until the signal is determined to be finished in step 240, at which point the process jumps to step 260 to end.

[0073] Figure 4 This is a schematic diagram of the phase recovery result according to an embodiment of the present application, illustrating the impact of different levels of phase noise and signal-to-noise ratio on phase estimation performance.

[0074] As an example, and not a limitation, in this embodiment, the transmitted signal is a 50 GSPS QPS modulated signal, and the number of pilots is N. p =10, Number of data signals N m =40, phase noise variance Signal-to-noise ratio (SNR) = 10 dB, L = 2. Based on the above conditions, according to... Figure 2 and 3 The method shown yields a phase estimate to recover the phase. Specifically, the steps are as follows:

[0075] Step (1): Obtain pilot signal samples Specifically as follows:

[0076] 4.4721i 4.4721 4.4721i -4.4721 4.4721 -4.4721i 4.4721 4.4721i 4.4721 4.4721

[0077] Step (2) Obtain the received signal sample Specifically as follows:

[0078] -0.3897+5.2373i 4.4052+0.4815i -0.2955+4.3930i -3.7478+0.4784i 4.7609+1.6145i -0.0893-4.4752i 3.9467+0.6508i -1.2575+4.7281i 4.8004+0.7594i 4.6705+1.1398i 3.2629+0.5041i 3.0452+1.3401i 4.3343+1.7239i -0.6372+4.4524i 4.4634+1.1707i 3.2315+0.2577i -5.5219-1.4794i 1.0031-5.4884i 0.5548-5.3033i 4.1697+0.4011i 0.3470+3.2919i -1.4801+5.9653i -4.3178-0.1982i -0.2125+2.9811i 5.0341-0.4182i -1.1006+4.3686i 4.7020-0.7603i 6.4095-0.0929i -1.1900-3.6990i 0.5100+4.3516i -0.4151-4.5125i -0.9572-5.1648i 3.8630-0.8238i -4.0688-0.2810i -2.8004+0.8101i -4.9127+0.7124i -3.9762+1.5545i 3.9945+0.0209i 0.3577-3.5878i 4.7316-1.0769i 0.8911+4.0051i -0.6842-5.1944i -1.1294-4.2443i 1.8187+5.0289i 1.1077+4.0560i 1.1077+4.0560i -5.1381+0.9355i -0.6438-3.9608i 0.3565-3.7983i -5.2559-0.2579i

[0079] Step (3): Take the sample obtained in step (1) and the first N obtained in step (2) p Substituting each sample into formula (11), we obtain the initial values ​​for the iteration.

[0080]

[0081] Step (4): Take the sample obtained in step (2) Substituting into formula (6), the phase estimation result is obtained by iterative method. The iterative method is as follows:

[0082] (a1) Set the threshold for determining the end of the iteration.

[0083] ε θ =0.01

[0084] (a2) Calculate the phase estimate after the first iteration

[0085]

[0086] because Continue iterating and calculate the phase estimate after the second iteration.

[0087]

[0088] When the number of iterations is 2 Iteration stopped, selection As the final phase estimate.

[0089] Step (5): Perform phase compensation to obtain the Nth... p +1 phase-compensated received signal, i.e.

[0090]

[0091] Step (6): Shift the estimation window one position to the right, and repeat steps (4) to (5) to obtain the Nth step. p The received signal after +2 phase compensation is r′(12)≈2.5908+2.0875i.

[0092] Step (7): Shift the estimation window one position to the right, and repeat steps (4) to (5), ..., until the Nth window is obtained. p +N m -L received signals have been received.

[0093] like Figure 4 As shown, the phase recovery module and method of this application, by employing soft-decision phase estimation 410 compared to hard-decision phase estimation 420, have a higher tolerance for signal-to-noise ratio and phase noise. This is mainly due to the fact that the phase recovery module and method described in this application fully consider the signal phase changes within the estimation window, effectively compensate for signal phase jitter within the estimation window, and the channel phase estimation based on soft-decision effectively avoids the accumulation of errors, thus improving the stability of phase estimation.

[0094] For the sake of brevity, not all possible substitutions and / or combinations of the various technical features in the above embodiments are described. However, as long as there is no contradiction in the substitutions and / or combinations of these technical features, they should be considered to be covered by the scope of this specification.

[0095] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be understood that various modifications, substitutions, additions, and deletions can be made to the various embodiments and their features within the spirit and principles of this application without exceeding the scope of protection claimed in this application.

Claims

1. A method for channel carrier phase recovery, characterized in that, The method includes: Receive signals from the front-end channel estimation module; The received signal is decomposed in the time domain to construct an estimation window; The carrier phase at the current moment is calculated using the soft decision information of the sample signals within the estimation window, in order to recover the carrier phase; The calculation of the carrier phase at the current moment includes calculating the phase estimate of the current signal based on the soft-decision maximum likelihood phase estimate of the sample signal within the estimation window. To obtain the carrier phase at the current moment, The estimation window is a symmetrical estimation window with a length of 2L+1 centered on the received signal at the current time, where L is a positive integer greater than 0. The phase estimation of the current signal It is determined by the following formula: Where k is the signal sequence number, This indicates the originating signal within the current estimation window. Denotes the real part of x. Denotes the imaginary part of x. ， Where K represents the received signal sample The cross-correlation matrix has a length of (2L+1)×(2L+1), and the elements in the cross-correlation matrix are determined by the following formula: Where N represents the set and Size of the intersection Here, δ(x) represents the phase noise variance, δ(x) represents the Dirac function with respect to x, and N0 is the two-sided power spectral density of additive white Gaussian noise. If the received signal has not yet ended, the estimation window is shifted backward by one signal cycle so that its center position falls at the next time moment, and the carrier phase at the next time moment is calculated, until the received signal ends. The calculation of the carrier phase at the current moment also includes obtaining a phase estimation result through iteration and performing phase recovery, wherein the iteration includes: (a1) Set the threshold ε for determining the end of the iteration. θ ; (a2) Determine the iteration starting point and select the final phase estimate of the sample corresponding to the previous moment of the currently received signal sample. As the initial value for the current phase estimation iteration; (a3) Based on the phase estimation of the (n+1)th iteration Phase estimation with the nth iteration The relationship is used to start the iteration, and the phase estimate obtained in the (n+1)th iteration is calculated. in And (a4) when When the time is reached, the iteration ends; the current iteration phase estimate is set. As the final estimate of the phase Right now 2. The method as described in claim 1, characterized in that, The initial pilot phase estimate of the received signal is the initial pilot phase estimate of the received signal. Where m(l) is the pilot signal, N p It represents the number of pilot signals.

3. An apparatus for channel carrier phase recovery, the apparatus being configured to perform the method as described in any one of claims 1 to 2.

4. A receiver for a communication system, comprising a pre-channel estimation module, a phase recovery module, and a demodulation module connected in sequence, wherein the phase recovery module is the device for channel carrier phase recovery as described in claim 3.

Images