A method for simplifying coherent DSP

Abstract

This invention discloses a method for simplifying coherent DSP. A pilot signal (PT) is inserted outside the spectrum band of each dual-polarization signal at the transmitter. The receiver extracts the pilot signal through low-pass filtering. Frequency offset, phase noise, and polarization rotation are then estimated and compensated. A simple PAM equalization algorithm or a real 2x2 multiple-input, multiple-output (MIMO) equalization algorithm is then used to eliminate intersymbol interference (ISI), simplifying the coherent DSP. This method achieves not only fast polarization tracking capabilities and low complexity, but also low power consumption.

Description

Technical Field

[0001] The invention belongs to the technical field of DSP, and in particular relates to a method for simplifying coherent DSP. Background Art

[0002] With the rapid emergence and rollout of next-generation information technologies such as 5G communications, the Internet of Things, cloud computing, and artificial intelligence, global internet data traffic continues to grow rapidly in medium- and short-haul applications such as data centers, metropolitan area networks, and access networks. Ultra-high-speed (800Gb / s and higher) transmission system solutions are urgently needed. To address these industry trends and pressing needs, it is necessary to develop optical communication technologies that increase data capacity. Given the characteristics of medium- and short-haul applications, the available technical solutions must meet stringent requirements for low system power consumption and low implementation complexity. Traditional intensity modulation-direct detection solutions are simple to implement, but are currently unable to meet current data capacity requirements due to device bandwidth limitations. Coherent optical communication technologies based on polarization multiplexing and IQ modulation can significantly increase system transmission capacity and meet current industry demands. Therefore, coherent optical detection solutions have become a viable option for medium- and short-haul applications. However, the high complexity and power consumption of traditional coherent DSP algorithms make them unsuitable for medium- and short-haul applications. To address this demand, it is necessary to research new optical communication technologies that balance transmission capacity and power consumption to meet industry development needs.

[0003] Furthermore, in traditional coherent DSP, polarization demultiplexing is typically achieved using a multiple-input, multiple-output finite impulse response (MIMO) filter based on a constant modulus algorithm or a multimode algorithm. Following polarization demultiplexing, a fast Fourier transform (FFT) algorithm is used to estimate frequency offset. Carrier phase recovery is then achieved using a blind phase search (BPS) algorithm or a Viterbi-Viterbi phase estimation (VVPE) algorithm. However, this traditional coherent DSP approach is computationally complex and consumes a lot of power. Some have proposed using adaptive polarization controllers to implement polarization demultiplexing in optical fiber links to simplify DSP algorithms. However, current adaptive polarization control technology is not mature and suffers from issues such as slow polarization tracking. Furthermore, this approach offers limited DSP simplification. Some researchers have proposed using coherent coherence to avoid frequency offset estimation and carrier phase recovery, but this approach suffers from path mismatch and wastes channel resources. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a method for simplifying coherent DSP, which solves the problems of high power consumption and complicated processing in the prior art.

[0005] In order to solve the above technical problems, the technical solution of the present invention is implemented as follows: a method for simplifying coherent DSP, inserting a pilot PT outside the dual-polarization signal spectrum band at the transmitting end, extracting the pilot PT through low-pass filtering at the receiving end, and then simultaneously estimating and compensating for frequency offset, phase noise and polarization rotation. Afterwards, a simple PAM equalization algorithm or a real 2*2 multiple-input multiple-output (MIMO) equalization algorithm is used to eliminate inter-symbol interference (ISI) to simplify coherent DSP.

[0006] Preferably, the pilot PT and the signal do not overlap each other in the spectrum, and the transmission signal of the transmitting end is expressed as:

[0007]

[0008] Among them, E x / y (t), S x / y (t) represents the X / Y polarized transmission signal and SCM signal, respectively, and A and ω represent the amplitude and angular frequency of PT, respectively.

[0009] Preferably, the power of the pilot PT depends on the pilot signal power ratio PSR, specifically:

[0010] PSR(dB)=10log10(P pilot / P signal ).

[0011] PSR can usually be set to around -15dB.

[0012] Preferably, the transmission signal of the receiving end is expressed as:

[0013]

[0014] Among them, R x / y (t), J(t), Δω, n(t) represents the received X / Y polarization signal, the time-varying Jones matrix caused by random birefringence in the optical fiber, the frequency offset between the transmitting laser and the local oscillator light, the carrier phase noise, and the additive white Gaussian noise, respectively.

[0015] Preferably, in this method, since the pilot PT frequencies are symmetrical, the frequency offset Δω is estimated by calculating the deviation between the sum of the two pilot PT frequencies and the zero frequency, and then the pilot PT frequency is shifted to the zero frequency. The pilot PT is extracted using a low-pass filter, and the estimation matrix M of the polarization rotation and phase noise can be obtained. The matrix M is expressed as:

[0016]

[0017] Here, H{·} represents a low-pass filter operation.

[0018] Preferably, the low-pass filter is implemented by block averaging.

[0019] Preferably, the inverse matrix of signal compensation is:

[0020]

[0021] in,{·} -1 Represents the inverse of a matrix.

[0022] Preferably, when non-negligible dispersion exists in the optical fiber channel, after PT is used to compensate for frequency offset, polarization aliasing and phase noise, a real-valued 2*2 multiple-input multiple-output (MIMO) equalization algorithm is used to eliminate inter-symbol interference.

[0023] Preferably, when the optical fiber channel dispersion is negligible, after the frequency offset, polarization aliasing and phase noise are compensated by PT, a 4-way PAM equalization algorithm is used to eliminate inter-symbol interference (ISI).

[0024] Another technical solution of the present invention is achieved as follows: a coherent optical communication module applies the above-mentioned low-power simplified coherent DSP method.

[0025] Compared with the prior art, the method of the present invention not only achieves the purpose of fast polarization tracking capability and low complexity, but also achieves the purpose of low power consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] FIG1 is a spectrum of transmitted signals in a communication module provided in Embodiment 1 of the present invention, wherein:

[0027] Figure 1a is the X-polarization signal spectrum;

[0028] Figure 1b is the spectrum of the Y-polarization signal;

[0029] Figure 2 This is a flowchart of Example 1 of the present invention;

[0030] Figure 3 This is a block diagram of the PAM equalization principle of Example 1 of the present invention;

[0031] Figure 4 This is a flowchart of Example 2 of the present invention;

[0032] FIG5 is a spectrum of transmitted signals in a communication module provided in Embodiment 2 of the present invention, wherein:

[0033] Figure 5a is the X-polarization signal spectrum;

[0034] Figure 5b is the Y-polarization signal spectrum. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0036] An embodiment of the present invention provides a method for simplifying coherent DSP, specifically: a pilot PT is inserted outside the spectrum band of the dual-polarization signal at the transmitting end, and the pilot PT is extracted by low-pass filtering at the receiving end. Then, frequency offset, phase noise and polarization rotation are simultaneously estimated and compensated. After that, a simple PAM equalization algorithm or a real 2*2 multiple-input multiple-output (MIMO) equalization algorithm is used to eliminate inter-symbol interference (ISI) to simplify coherent DSP.

[0037] Furthermore, the pilot PT and the signal do not overlap each other in the spectrum, and the transmission signal of the transmitting end is expressed as:

[0038]

[0039] Among them, E x / y (t), S x / y (t) represents the X / Y polarized transmission signal and SCM signal, respectively, and A and ω represent the amplitude and angular frequency of PT, respectively.

[0040] Furthermore, the power of the pilot PT depends on the pilot signal power ratio PSR, specifically:

[0041] PSR(dB)=10log10(P pilot / P signal ).

[0042] PSR can usually be set to around -15dB.

[0043] Furthermore, the transmission signal of the receiving end is expressed as:

[0044]

[0045] Among them, R x / y (t), J(t), Δω, n(t) represents the received X / Y polarization signal, the time-varying Jones matrix caused by random birefringence in the optical fiber, the frequency offset between the transmitting laser and the local oscillator light, the carrier phase noise, and the additive white Gaussian noise, respectively.

[0046] Furthermore, since the pilot PT frequencies are symmetrical, the frequency offset Δω is estimated by calculating the deviation between the sum of the two pilot PT frequencies and zero frequency. The pilot PT frequency is then shifted to zero frequency and the pilot PT is filtered using a low-pass filter. This yields the estimated matrix M for polarization rotation and phase noise. The matrix M is expressed as:

[0047]

[0048] Here, H{·} represents a low-pass filter operation.

[0049] Furthermore, the low-pass filter is implemented by block averaging.

[0050] Furthermore, the inverse matrix of signal compensation is:

[0051]

[0052] in,{·} -1 Represents the inverse of a matrix.

[0053] Furthermore, when non-negligible dispersion exists in the optical fiber channel, after using PT to compensate for frequency offset, polarization aliasing and phase noise, a real-valued 2*2 multiple-input multiple-output (MIMO) equalization algorithm is used to eliminate inter-symbol interference.

[0054] Furthermore, when the fiber channel dispersion is negligible, after PT is used to compensate for frequency offset, polarization aliasing, and phase noise, a 4-way PAM equalization algorithm is used to eliminate inter-symbol interference (ISI).

[0055] Through the above-mentioned method of the embodiment of the present invention, not only the estimation and compensation of frequency offset, phase noise and polarization rotation are realized, but also the computational complexity of coherent DSP is greatly reduced. Moreover, when there is dispersion in the channel, since the two polarization signals are independent of each other and there is no crosstalk, two independent multi-tap real number 2*2 MIMO (multiple input multiple output) equalizers based on DDLMS (direct decision least mean square algorithm) can be used to simultaneously compensate for dispersion and eliminate inter-symbol crosstalk, which can further reduce the complexity compared to 4*4 MIMO in traditional coherent DSP. When there is no dispersion in the channel, after frequency offset, phase noise and polarization rotation compensation, the constant phase error introduced during PT down-conversion is first eliminated by a constant phase compensation algorithm. At this time, there is no polarization crosstalk and IQ crosstalk in the signal. The four signals are independent of each other and can be treated as PAM signals and equalized separately at a sampling rate of single baud rate, further reducing the computational complexity.

[0056] In addition, an embodiment of the present invention further provides a coherent optical communication model, which applies the above-mentioned method for simplifying coherent DSP.

[0057] The following are specific embodiments

[0058] Example 1

[0059] See also Figure 2 , Embodiment 1 of the present invention provides a coherent optical communication model.,This embodiment 1 is applicable to short-distance transmission scenarios (less than 10 km) and scenarios where dispersion can be ignored;

[0060] The specific flow chart is as follows:

[0061] The transmitter's DSP generates two independent random bit sequences, each mapped to the desired modulation format, such as QASK, 16QAM, or 64QAM. A root-raised cosine filter is used to reshape the signal. PTs are inserted at symmetrical out-of-band frequencies in the two signal spectra, leaving a guard bandwidth of at least 500 MHz between the PTs and the signal spectrum. The resulting signal spectrum is shown in Figure 1.

[0062] The signal is then converted into an analog signal by an arbitrary waveform generator and drives a dual-polarization optical IQ modulator. The input light of the modulator uses an O-band laser. The modulated signal is loaded into a standard single-mode optical fiber and transmitted to the receiving end. After passing through a bandpass optical filter, the signal light and the local oscillator light are received by a dual-polarization coherent optical receiver. The received signal is collected by a high-speed acquisition oscilloscope, and the collected electrical signal passes through a subsequent offline digital signal processing module. The digital signal processing module first compensates for the IQ delay and imbalance. Then, the frequency offset is estimated by calculating the deviation between the sum of the frequencies of the two PTs and the zero frequency. Subsequently, the PT is extracted by a low-pass filter according to formulas (4) and (5) to obtain the estimated signal polarization rotation and phase noise matrix M. The estimated matrix M is combined with the frequency offset to simultaneously compensate for the frequency offset, polarization and carrier phase noise. After the frequency offset, phase noise and polarization rotation are compensated, the constant phase error introduced during the PT down-conversion is first eliminated by a constant phase compensation algorithm. At this point, there is no polarization crosstalk or IQ crosstalk in the signal. The four signals are independent of each other and can be processed as PAM signals. PAM equalization algorithm is used to perform equalization at a sampling rate of single baud rate to eliminate inter-symbol crosstalk.

[0063] Figure 3 The block diagram of PAM equalization is shown in Figure 4. Finally, the bit error rate (BER) is calculated.

[0064] Example 2

[0065] Figure 4 The second embodiment of the present invention provides a coherent optical communication module. This embodiment is applicable to digital subcarrier multiplexing scenarios and can be used in point-to-multipoint (PTMP) scenarios in aggregation networks. The specific flow chart is as follows:

[0066] The transmitter's DSP generates six independent random bit sequences, each mapped to a desired modulation format such as QASK, 16QAM, or 64QAM. A root-raised cosine filter (RRC) is used to reshape the signal. The reshaped signals undergo subcarrier multiplexing (SCM), with each polarization signal consisting of three carriers (the number of subcarriers can be adjusted to meet specific requirements and can be greater than or equal to three). A guard bandwidth (above 500 MHz) is reserved between adjacent subcarriers in the SCM signal, and transmission time (PT) signals are inserted symmetrically within the guard bandwidth of the two polarization signals. The spectrum of the generated signal is shown in Figure 5.

[0067] The signal is then converted into an analog signal by an arbitrary waveform generator and drives a dual-polarization optical IQ modulator. The input light of the modulator uses an O-band laser, and the fiber dispersion can be ignored at this time. The modulated signal is loaded into a standard single-mode fiber and transmitted to the receiving end. The signal light and the local oscillator light are received by a dual-polarization coherent optical receiver, and the received signal is collected by a high-speed acquisition oscilloscope. The collected electrical signal passes through the subsequent offline digital signal processing module. The digital signal processing module first compensates for the IQ delay and imbalance. Then, the frequency offset is estimated by calculating the deviation between the sum of the frequencies of the two PTs and the zero frequency. Subsequently, the PT is extracted by a low-pass filter according to formulas (4) and (5) to obtain the estimated signal polarization rotation and phase noise matrix M. The estimated matrix M is combined with the frequency offset to simultaneously compensate for the frequency offset, polarization and carrier phase noise. The constant phase error introduced during PT down-conversion is first eliminated by a constant phase compensation algorithm. At this time, there is no polarization crosstalk and IQ crosstalk in the signal. The four signals are independent of each other and can be processed as PAM signals. The PAM equalization algorithm is used to equalize them separately at a sampling rate of single baud rate to eliminate inter-symbol crosstalk.

[0068] Figure 3 The block diagram of PAM equalization is shown in Figure 4. Finally, the bit error rate (BER) is calculated.

[0069] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for simplifying coherent DSP, characterized in that: At the transmitter, a pilot signal (PT) is inserted outside the spectrum band of each dual-polarization signal. At the receiver, low-pass filtering is used to extract the pilot signal. Frequency offset, phase noise, and polarization rotation are then estimated and compensated. A 4-way PAM equalization algorithm or a real 2x2 MIMO equalization algorithm is then used to eliminate inter-symbol interference (ISI), simplifying coherent DSP. The pilot PT and the signal do not overlap each other in the spectrum. The transmission signal of the transmitting end is expressed as: Among them, E x / y (t), S x / y (t) represents the X / Y polarized transmission signal and SCM signal, A and ω represent the amplitude and angular frequency of PT, respectively; The power of the pilot PT depends on the pilot signal power ratio PSR, specifically: PSR(dB)=10log10(P pilot / P signal ); The transmission signal of the receiving end is expressed as: Among them, R x / y (t), J(t), Δω, n(t) represents the received X / Y polarization signal, the time-varying Jones matrix caused by random birefringence in the optical fiber, the frequency offset between the transmitting laser and the local oscillator light, the carrier phase noise, and the additive white Gaussian noise, respectively; In this method, since the pilot PT frequencies are symmetrical, the frequency offset Δω is estimated by calculating the deviation between the sum of the two pilot PT frequencies and zero frequency. The pilot PT frequency is then shifted to zero frequency and the pilot PT is extracted using a low-pass filter. This yields an estimated matrix M for polarization rotation and phase noise. The matrix M is expressed as: Where, H{·} represents the low-pass filter operation; The low-pass filter is implemented by block averaging; The inverse matrix of signal compensation is: in,{·} -1 represents the inverse of a matrix; When non-negligible dispersion exists in the optical fiber channel, after using PT to compensate for frequency offset, polarization aliasing, and phase noise, a real-valued 2*2 MIMO equalization algorithm is used to eliminate inter-symbol interference. When fiber channel dispersion is negligible, after using PT to compensate for frequency offset, polarization aliasing, and phase noise, a 4-way PAM equalization algorithm is used to eliminate inter-symbol interference.

Images