A low-power, endpoint-tolerant DSP system

Abstract

This invention provides a low-power, transmitter-end-injury-tolerant DSP system, comprising a transmitter and a receiver. The transmitter includes a transmitter DSP module, a digital-to-analog converter module, and a dual-polarization optical IQ modulator. The receiver includes a dual-polarization coherent optical receiver, an analog-to-digital converter module, and a digital signal processing module connected in sequence. The advantages of this invention are: low power consumption, tolerance to transmitter-end IQ impairment, fast polarization tracking capability, and low complexity.

Description

Technical Field

[0001] This invention relates to the field of optical communication, and more particularly to a low-power, transmitter-end-damage-tolerant DSP system. Background Technology

[0002] With the emergence of various new technologies such as edge computing, IoT, mesh audio, and 5G / 6G, the growth rate of IP data traffic in metropolitan area networks (MANs) and access networks has surpassed that of backbone networks. The direct modulation-direct detection (IM-DD) and time division multiplexing (TDM) technologies commonly used in MANs and access networks cannot meet the ever-increasing traffic demands, leading to the decentralization of coherent reception solutions to MANs and access networks. Star topologies are commonly used in MANs and access networks to implement point-to-multipoint (P2MP) network architectures, thereby simplifying network design and maintenance. Against this backdrop, recently proposed coherent optical communication schemes based on digital subcarrier multiplexing (DSCM) can increase transmission capacity while achieving a flexible P2PM structure, which well matches the evolving needs of MANs and access networks. However, the high complexity and high power consumption of traditional coherent DSP algorithms make them unsuitable for direct application to DSCM coherent systems. Therefore, given this demand, it is necessary to conduct research on new optical communication technologies that balance transmission capacity and power consumption to meet the needs of industrial development.

[0003] In DSCM systems, to compensate for channel impairments, each subcarrier signal is typically treated as a single carrier signal after subcarrier demultiplexing, and traditional DSP algorithms are used to process each subcarrier separately. These algorithms include equalization, frequency offset estimation, and carrier phase recovery. However, since each subcarrier requires similar processing, this leads to significant computational redundancy and increases DSP power consumption. Previously, we proposed a scheme utilizing frequency domain FPT (Yang Yanfu, Fan Linsheng. A Low-Complexity Method for Co-recovery of Polarization Rotation and Carrier Phase [P]. Guangdong Province: CN114915350A, 2022-08-16.), which achieves the co-recovery of frequency offset, polarization aliasing, and carrier phase of all subcarriers before subcarrier demultiplexing, greatly reducing the computational complexity of the DSP.

[0004] However, DSCM systems are significantly more sensitive to imperfections in transceiver optoelectronic devices compared to single-carrier (SC) systems. These imperfections introduce in-phase quadrature (IQ) component damage, primarily including IQ amplitude and phase imbalances and IQ delays. These IQ impairments are particularly severe in DSCM systems, causing each subcarrier to generate conjugate symmetrical components, thus interfering with frequency-symmetrical subcarriers. For transceiver device impairments, compensation can be performed immediately after analog-to-digital conversion (ADC); however, for transceiver impairments, compensation can only be performed after polarization demultiplexing and carrier phase recovery. However, transceiver IQ impairments severely affect the demultiplexing and carrier phase recovery of DSCM signals, degrading system performance. To address the transceiver IQ impairment problem, a feasible solution is to use an 8×8 real-valued butterfly equalizer or a 4×4 complex wide-linearity butterfly equalizer to simultaneously equalize the four subcarrier signals (two polarization-symmetrical subcarriers), achieving polarization demultiplexing and compensating for transceiver IQ impairments. However, this approach has very low tolerance for phase noise, slow polarization tracking rate, and extremely high complexity. For power- and energy-sensitive applications, the computational complexity introduced by these two equalizers is unacceptable. Therefore, it is essential to find a low-complexity DSP technology suitable for DSCM signals that is highly tolerant of transmitter-end defects. Summary of the Invention

[0005] This invention provides a low-power, transmitter-end-impairment-tolerant DSP system, characterized by comprising a transmitter and a receiver. The transmitter includes a transmitter DSP module, a digital-to-analog converter module, and a dual-polarization optical IQ modulator. The transmitter DSP module generates independent random bit sequences loaded onto each subcarrier, which are mapped to signals of the desired modulation format. Root-raised cosine filters are used to shape the signals. The shaped signals are then multiplexed by row subcarriers to obtain SCM signals. A guard bandwidth is reserved between adjacent subcarriers of the SCM signal. One pilot frequency signal (FPT) is inserted into each of the two polarization signal guard bandwidths. The FPTs represent pilot signals in the frequency domain, and the frequencies of the two polarization FPTs are different and asymmetrical. Then, the digital-to-analog converter module converts the signals into analog signals and drives the dual-polarization optical IQ modulator. The modulated signals are loaded onto a standard single-mode optical fiber and transmitted to the receiver.

[0006] The receiving end includes a dual-polarization coherent optical receiver, an analog-to-digital converter, and a digital signal processing module connected in sequence. The dual-polarization coherent optical receiver is used to receive signal light and local oscillator light. The received signal is converted into a digital signal by the analog-to-digital converter, and the resulting electrical signal is processed by the digital signal processing module.

[0007] As a further improvement of the present invention, the mapping to the desired modulation format signal includes, but is not limited to, QPSK, 16QAM, and 64QAM.

[0008] As a further improvement of the present invention, in the transmitter DSP module, each polarization signal is composed of an even number of carriers.

[0009] As a further improvement of the present invention, a protection bandwidth is reserved between adjacent subcarriers of the SCM signal, and the protection bandwidth must ensure that the spectra of adjacent subcarriers do not overlap. 5. The DSP system according to claim 1, characterized in that the digital signal processing module first compensates for the IQ delay and imbalance of the receiving device, then estimates the frequency offset by calculating the deviation between the frequencies of the two FPTs transmitted at their respective frequencies, subsequently downconverts the two FPTs to 0 frequency, extracts the amplitude and phase of the FPTs by a low-pass filter, and obtains the estimated signal polarization rotation and phase noise matrix; the estimated matrix is ​​combined with the frequency offset to simultaneously compensate for the frequency offset, polarization, and carrier phase noise;

[0010] After compensating for frequency offset, phase noise, and polarization rotation, the signal is first demultiplexed by subcarriers, and then the two polarized signals are equalized separately. An equalizer is used to simultaneously equalize the two symmetrical subcarriers to compensate for transmission-end impairments. The signals of the two polarization states are equalized separately. Finally, the bit error rate is calculated.

[0011] As a further improvement of the present invention, in the digital signal processing module, the FPT is extracted from the low-pass filter according to the following formula:

[0012]

[0013]

[0014] Where H{·} represents the low-pass filter operation, and R x / y (t) represents the received X / Y polarized signal, ω1 and ω2 are the angular frequencies of the X and Y polarized FPTs respectively, and Δω and Δt are the angular frequencies of the FPTs. These represent the frequency offset, carrier phase noise, and additive white Gaussian noise between the transmitting laser and the local oscillator light, respectively. and These are the FPTs extracted from the low-pass filter, respectively.

[0015] As a further improvement of the present invention, the low-pass filter is implemented using a sliding window averaging method.

[0016] As a further improvement of the present invention, in the digital signal processing module, the equalizer is a 2*2 complex MIMO or a 4*4 real MIMO.

[0017] As a further improvement of the present invention, in the digital signal processing module, the tap coefficients of the filter are updated according to the minimum mean algorithm of direct decision, which can compensate for residual phase noise.

[0018] The beneficial effects of this invention are: it features low power consumption, tolerance to IQ impairment at the transmitter, fast polarization tracking capability, and low complexity. Attached Figure Description

[0019] Figure 1 The transmitted signal spectrum is shown in Figure 1, where (a) is the spectrum of the X-polarized signal and (b) is the spectrum of the Y-polarized signal.

[0020] Figure 2 These are the power spectra of signals under different damage conditions, where (a) no damage; (b) IQ delay: 20 ps; (c) amplitude imbalance: 3 dB, phase imbalance: 20°; (d) IQ delay: 20 ps, ​​amplitude imbalance: 3 dB, phase imbalance: 20°.

[0021] Figure 3 It is a 2x2 complex MIMO;

[0022] Figure 4 This is a system schematic diagram of the present invention;

[0023] Figure 5 This is a flowchart of the digital signal processing module;

[0024] Figure 6 This is a flowchart of the algorithm for frequency offset, polarization, and carrier phase noise recovery. Detailed Implementation

[0025] This invention discloses a low-power, transmitter-end-damage-tolerant DSP system. This DSP system is a DSP solution for DSCM coherent systems, featuring low power consumption, tolerance to transmitter-end IQ damage, and fast polarization tracking capabilities.

[0026] I. System Model and Principles:

[0027] Figure 1 (a) and (b) show the transmitted signal spectrum of one embodiment of the present invention. In the SCM signal, a guard bandwidth is provided between each subcarrier. A frequency-domain pilot signal (FPT) is inserted into the guard bandwidth in both polarization directions. To avoid crosstalk of the FPT caused by transmitter-end impairment, the frequencies of the two FPTs cannot be symmetrical. The transmitted signal can be represented as:

[0028]

[0029] Where E x / y (t),S x / y (t) represent the X / Y polarized transmitted signal and the SCM signal, respectively, and A and ω represent the amplitude and angular frequency of the FPT, respectively. The added power of the FPT depends on the pilot signal power ratio (PSR), defined as PSR(dB) = 10log10(P pilot / Psignal Under the influence of the initial IQ impairment effect, the output signal of Tx can be expressed as:

[0030]

[0031] Where, k ij and K ij (i∈{1,2}, j∈{1,2}) is the transfer function of IQ impairment. It can be seen that new conjugate components are generated under the influence of IQ impairment.

[0032] Ignoring polarization mode dispersion and polarization correlation loss, the received signal can be expressed as:

[0033]

[0034] Among them, R x / y (t),J(t),Δω, These represent the received X / Y polarization signals, the time-varying Jones matrix caused by random birefringence in the fiber, the frequency offset between the transmitting laser and the local oscillator light, carrier phase noise, and additive white Gaussian noise, respectively. The frequency offset Δω is estimated by comparing the frequency difference between the received signal FPT and the transmitted signal FPT. After estimating the frequency offset, the frequency of FPT is shifted to zero, and FPT is extracted using a low-pass filter, resulting in the following matrix M:

[0035]

[0036]

[0037]

[0038] Considering frequency offset, the inverse matrix of signal compensation can be obtained as: T(t)=conj{W T}e -jΔωt

[0039] Where H{·} represents the low-pass filter operation, conj{·} represents the conjugate operation, and {·} T This represents the transpose of a matrix.

[0040] To further reduce computational complexity, low-pass filters can be implemented using a block averaging approach. Here, the constant AK... 1x (ω1) and AK 1y (ω2) Eliminate by normalization: |J xx | 2 +|J yx | 2 =1 and |J xy | 2 +|J yy | 2=1.

[0041] Through the above process, even in the presence of transmitter damage and phase noise, it is still possible to simultaneously estimate and compensate for frequency offset, phase noise, and polarization rotation, which greatly reduces the computational complexity of coherent DSPs while ensuring the robustness of the algorithm to transmitter damage.

[0042] After compensating for frequency offset, phase noise, and polarization rotation, the signal is first demultiplexed by subcarriers. Then, the two polarized signals are equalized separately. A 2x2 complex MIMO is used to simultaneously equalize the two symmetrical subcarriers to compensate for transmission-end impairments. The equalizer structure diagram is shown below. Figure 3 As shown. Where E Ain and E Bin These are two subcarriers with mutually symmetrical frequencies. II. Specific Implementation Examples:

[0044] The DSP system of this invention is suitable for digital subcarrier multiplexing scenarios and can be used in point-to-multipoint (PTMP) scenarios in aggregation networks.

[0045] like Figure 4 As shown, the transmitter's DSP module generates independent random bit sequences loaded onto each subcarrier, which are mapped to the desired modulation format signal, such as QPSK, 16QAM, 64QAM, etc. Root-raised cosine filters are then used to shape the signals. After shaping, the signals undergo row subcarrier multiplexing (SCM), with each polarization signal consisting of four carriers (the number of subcarriers can be changed as needed, but is always an even number). A guard bandwidth (above 500MHz) is reserved between adjacent subcarriers in the SCM signal. One free-pole transducer (FPT) is inserted within each of the two polarization signal guard bandwidths. The frequencies of the two FPTs are different and asymmetrical. The spectrum of the generated signal is shown below. Figure 1 As shown. The signal is then converted to an analog signal by a digital-to-analog converter (DAC) module, driving a dual-polarization optical IQ modulator. The modulated signal is loaded onto a standard single-mode fiber and transmitted to the receiver. The signal light and local oscillator light are received by a dual-polarization coherent optical receiver. The received signal is converted to a digital signal by an DAC, and the resulting electrical signal is processed by a subsequent offline digital signal processing module. The digital signal processing module flow is as follows. Figure 5 As shown, the digital signal processing module first compensates for the IQ delay and imbalance of the receiving device. Then, it estimates the frequency offset by calculating the deviation between the frequencies transmitted by the two FPTs. Subsequently, the two FPTs are down-converted to 0 frequency according to the following formula. The amplitude and phase of the FPT are extracted using a low-pass filter to obtain the estimated signal polarization rotation and phase noise matrices. These estimated matrices, combined with the frequency offset, are used to simultaneously compensate for frequency offset, polarization, and carrier phase noise. The algorithm block diagram for frequency offset, polarization, and carrier phase noise recovery is shown below. Figure 6 As shown.

[0046] After compensating for frequency offset, phase noise, and polarization rotation, the signal is first demultiplexed by subcarriers. Then, the two polarized signals are equalized separately. A 2x2 complex MIMO is used to simultaneously equalize the two symmetrical subcarriers to compensate for transmission-end impairments. The two polarization states are equalized separately. The filter tap coefficients are updated according to the direct decision minimum mean algorithm to compensate for residual phase noise. The equalizer structure diagram is shown below. Figure 3 As shown. Finally, the bit error rate (BER) is calculated.

[0047] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A low-power, endpoint-tolerant DSP system, characterized in that, The system includes a transmitter and a receiver. The transmitter includes a transmitter DSP module, a digital-to-analog converter module, and a dual-polarization optical IQ modulator. The transmitter DSP module generates independent random bit sequences loaded onto each subcarrier, which are mapped to the desired modulation format signal. The signals are shaped using root-raised cosine filters. After shaping, the signals are multiplexed by row subcarriers to obtain SCM signals. A guard bandwidth is reserved between adjacent subcarriers of the SCM signal. One FPT is inserted into each of the two polarization signal guard bandwidths. The FPTs represent pilot signals in the frequency domain. The frequencies of the two polarization FPTs are different and asymmetrical. Then, the digital-to-analog converter module converts the signal into an analog signal and drives the dual-polarization optical IQ modulator. The modulated signal is loaded onto a standard single-mode fiber and transmitted to the receiver. The receiving end includes a dual-polarization coherent optical receiver, an analog-to-digital converter, and a digital signal processing module connected in sequence. The dual-polarization coherent optical receiver is used to receive signal light and local oscillator light. The received signal is converted into a digital signal by the analog-to-digital converter, and the resulting electrical signal is processed by the digital signal processing module. The digital signal processing module first compensates for the IQ delay and imbalance of the receiving device. Then, it estimates the frequency offset by calculating the deviation between the frequencies of the two FPTs transmitted at their respective frequencies. Subsequently, it downconverts the two FPTs to 0 frequency and extracts the amplitude and phase of the FPTs using a low-pass filter to obtain the estimated signal polarization rotation and phase noise matrix. The estimated matrix is ​​then combined with the frequency offset to simultaneously compensate for the frequency offset, polarization, and carrier phase noise. After compensating for frequency offset, phase noise, and polarization rotation, the signal is first demultiplexed by subcarriers, and then the two polarized signals are equalized separately. An equalizer is used to simultaneously equalize the two symmetrical subcarriers to compensate for transmission-end impairments. The signals of the two polarization states are equalized separately. Finally, the bit error rate is calculated.

2. The DSP system according to claim 1, characterized in that, The mapping to the desired modulation format signal includes QPSK, 16QAM, and 64QAM.

3. The DSP system according to claim 1, characterized in that, In the transmitter DSP module, each polarization signal consists of an even number of carrier waves.

4. The DSP system according to claim 1, characterized in that, A protection bandwidth is reserved between adjacent subcarriers of the SCM signal. The protection bandwidth must ensure that the spectra of adjacent subcarriers do not overlap.

5. The DSP system according to claim 1, characterized in that, In the digital signal processing module, the FPT is extracted from the low-pass filter according to the following formula: , , in, This indicates the operation of a low-pass filter, where This indicates the received X / Y polarization signal. and These are the angular frequencies of the X- and Y-polarized FPTs, respectively. and These represent the frequency offset, carrier phase noise, and additive white Gaussian noise between the transmitting laser and the local oscillator light, respectively. and These are the FPTs extracted from the low-pass filter, respectively.

6. The DSP system according to claim 5, characterized in that, The low-pass filter is implemented using a sliding window averaging method.

7. The DSP system according to claim 5, characterized in that, In the digital signal processing module, the equalizer is a 2*2 complex MIMO or a 4*4 real MIMO.

8. The DSP system according to claim 5, characterized in that, In the digital signal processing module, the tap coefficients of the filter are updated according to the minimum mean algorithm of direct decision, which can compensate for residual phase noise.

Images