Bandwidth limitation and electric dispersion cooperative pre-compensation method based on improved GS algorithm

Abstract

The invention relates to a bandwidth limitation and electric dispersion cooperative pre-compensation method based on an improved GS algorithm, and belongs to the technical field of optical fiber communication and compensation. According to the method, a filter for balancing bandwidth limitation is designed and integrated into an improved GS pre-EDC method, and the method specifically comprises the steps that 1, the improved GS pre-EDC is adopted at a transmitting end, balancing is conducted after FFE is used at a receiving end, and the tap coefficient of the FFE is stored as hBL; 2, initializing a signal; step 3, executing receiving end constraint on the amplitude of the odd number samples of the signal; step 4, converting the signal to a transmitting end through an inverse dispersion frequency domain transmission function and an inverse filter; step 5, executing sending end constraint on the phase of the signal; step 6, converting the signal back to a receiving end by using a dispersion frequency domain transmission function and a filter; and step 7, repeating the steps 3-6 to complete N iterations. The pre-compensation performance of the method is superior to that of an improved GS pre-EDC method, the bandwidth requirement for a photoelectric device in a high-speed system is reduced, and post-equalization is not needed.

Description

Technical Field

[0001] The present invention relates to a method for co-precompensation of bandwidth limitation and electrical dispersion based on an improved GS algorithm, belonging to the technical field of optical fiber communication and compensation. Background Art

[0002] Driven by applications such as big data, cloud computing, artificial intelligence, and short videos, the traffic of data center interconnection has increased exponentially. Intensity Modulation / Direct Detection (IM / DD) systems are widely used in cost-sensitive short-distance optical interconnections due to their low cost, low power consumption, and simple structure. In the context of high-speed transmission, the bandwidth limitation (BL) of low-cost devices will cause the loss of high-frequency information of the signal, resulting in inter-symbol interference. In addition, the periodic power fading caused by the combined action of chromatic dispersion (CD) of the optical fiber and the square-law detection of the receiver will cause serious distortion of the signal. Therefore, the impairments caused by bandwidth limitation and dispersion are still the main challenges for improving the transmission performance in the C-band of IM / DD systems.

[0003] Recently, the electrical dispersion pre-compensation method (Electronic Dispersion pre-compensation, pre-EDC) based on the Gerchberg-Saxton (GS) algorithm has become a research hotspot. The basic GS pre-EDC regards the amplitude at the transmitter and the phase information before direct detection at the receiver as degrees of freedom, and uses the GS iterative algorithm to generate a pre-distorted signal with the ability to counteract dispersion. On this basis, the improved GS pre-EDC method (Modified GS pre-EDC) introduces amplitude and phase errors, as well as related error inversion factors, to accelerate the algorithm convergence and improve the performance. A pre-EDC method based on a polynomial non-linear filter (PNLF) designs a PNLF to replace the improved GS pre-EDC to achieve signal pre-distortion, and a feed-forward equalizer (FFE) needs to be combined at the receiver for post-equalization. This method only uses the odd-sample information of the signal during the pre-compensation process. In order to make full use of the degrees of freedom in the improved GS pre-EDC method, a pre-EDC method based on a trained dual finite impulse response (FIR) filter further improves the compensation performance by designing FIR filters for odd and even samples respectively. However, few works have studied the impact of bandwidth limitation on the improved GS pre-EDC.

[0004] The present invention designs a filter for equalizing bandwidth limitation and incorporates it into an improved GS pre-EDC method. The present invention is dedicated to achieving pre-Bandwidth Limitation and Electrical Dispersion Co-Precompensation (pre-BL-EDC) according to an optimized GS algorithm, and improving the compensation performance of the improved GS pre-EDC in a bandwidth-limited system. Summary of the Invention

[0005] The object of the present invention is to solve the problem that the improved GS pre-EDC is affected by bandwidth limitation, and a method for pre-Bandwidth Limitation and Electrical Dispersion Co-Precompensation (pre-BL-EDC) based on an improved GS algorithm is proposed. Based on the improved GS pre-EDC, the method introduces a filter for equalizing bandwidth limitation to perform co-pre-distortion of bandwidth limitation and electrical dispersion on the signal. It is specifically divided into two stages.

[0006] To achieve the above object, the present invention is realized through the following technical solutions:

[0007] The method for co-pre-distortion of bandwidth limitation and electrical dispersion includes the following steps (where step 1 is stage one, and steps 2-7 are stage two):

[0008] Step 1: At the transmitter, use the improved GS pre-EDC to pre-distort the signal, and at the receiver, use FFE for post-equalization, and store the tap coefficients of the FFE as h BL ;

[0009] Step 2: Initialize the signal. Initialize the amplitude of the signal to the target amplitude A Target , and initialize the phase to zero. Among them, the target amplitude A Target is the amplitude after shaping the On-Off Keying (OOK) signal through a Raised Cosine (RC) filter with a roll-off factor of 1.0;

[0010] Step 3: Perform receiver-side constraints on the amplitudes of the odd samples of the signal, that is, replace them with A Target(odd) (n)-α·error(n). Among them, α is the amplitude error inversion factor, and its value range is 0≤α≤1, error is the amplitude error of the odd samples, and n is the sample index;

[0011] Step 4: Convert the signal to the transmitter side through the inverse dispersion frequency-domain transfer function and the inverse filter ;

[0012] Step 5: Perform transmitter-side constraints on the phase of the signal, that is, use to replace. Among them, β is the phase error inversion factor, and its value range is 0≤β≤1;

[0013] Step 6: Use the dispersion frequency-domain transfer function H CD and the filter H BL to convert the signal back to the receiving end;

[0014] Step 7: Repeat Steps 3 - 6 for N iterations. The amplitude A Tx (n) of the signal in Step 5 is used as the final output after the completion of this algorithm iteration. This pre-distorted signal can cancel the impairments caused by bandwidth limitation and dispersion, eliminating the need for post-equalization using FFE at the receiving end.

[0015] Among them, the dispersion frequency-domain transfer function H CD is expressed as:

[0016]

[0017] where D is the dispersion coefficient; λ is the optical carrier wavelength; ω is the angular frequency; l is the transmission distance; c is the speed of light in vacuum. The Z-domain expression of the filter H BL is:

[0018]

[0019] where k is the number of taps.

[0020] Beneficial Effects

[0021] A method for co-precompensation of bandwidth limitation and electrical dispersion based on an improved GS algorithm according to the present invention realizes co-precompensation of bandwidth limitation and electrical dispersion in a cost-sensitive IM / DD optical fiber transmission system. Compared with the improved GS pre-EDC method, it has the following beneficial effects:

[0022] 1. The method has a higher tolerance for system bandwidth limitation;

[0023] 2. The method improves the achievable symbol rate under a limited system bandwidth;

[0024] 3. The method can improve the compensation performance, enhance the receiving sensitivity, and eliminate the need for post-equalization. Description of the Drawings

[0025] Figure 1 is the schematic diagram of a method for co-precompensation of bandwidth limitation and electrical dispersion based on an improved GS algorithm according to the present invention. Among them, Figure 1 (a) is the overall schematic diagram, Figure 1 (b) is the schematic diagram of Stage 1, Figure 1 (c) is the schematic diagram of Stage 2;

[0026] Figure 2It is the spectrum comparison of the pre-distorted signal generated by a bandwidth limitation and electrical dispersion collaborative pre-compensation method based on an improved GS algorithm and an improved GS pre-EDC method of the present invention;

[0027] Figure 3 It is the spectrum comparison of the received signal of a bandwidth limitation and electrical dispersion collaborative pre-compensation method based on an improved GS algorithm and an improved GS pre-EDC method of the present invention;

[0028] Figure 4 It is the schematic diagram of the C-band 20 / 28 / 32 Gbit / s 50 km OOK IM / DD experimental system and the flowchart of digital signal processing (DSP) in Embodiment 1 of a bandwidth limitation and electrical dispersion collaborative pre-compensation method based on an improved GS algorithm of the present invention. Among them, the inset is the frequency response of the optical back-to-back channel;

[0029] Figure 5 It is the compensation result diagram of Embodiment 1 of a bandwidth limitation and electrical dispersion collaborative pre-compensation method based on an improved GS algorithm of the present invention at different baud rates;

[0030] Figure 6 It is the received sensitivity result diagram of Embodiment 1 of a bandwidth limitation and electrical dispersion collaborative pre-compensation method based on an improved GS algorithm of the present invention at different baud rates. Among them, Figure 6 (a) is 20 Gbit / s, Figure 6 (b) is 28 Gbit / s, Figure 6 (c) is 32 Gbit / s. Detailed implementation manners

[0031] The present invention will be described below in conjunction with the accompanying drawings and specific embodiments. The following embodiments are only used to describe the present invention in more detail, but do not limit the scope of use of a bandwidth limitation and electrical dispersion collaborative pre-compensation method based on an improved GS algorithm of the present invention.

[0032] Embodiment 1

[0033] In specific implementation, for example, when some operators interconnect the data centers in Haidian District, Beijing with the data centers in Yizhuang District, signal distortion problems caused by device bandwidth limitation and fiber dispersion effects may be encountered. Using our method can effectively improve the signal quality, reduce the bit error rate, and ensure that the high-speed data transmission meets the design requirements. OOK signals of 20 Gbit / s, 28 Gbit / s, and 32 Gbit / s are respectively sent in a 50 km optical fiber transmission system, and the transmitted signals all adopt a bandwidth limitation and electrical dispersion collaborative pre-compensation method based on an improved GS algorithm of the present invention. The pre-compensation method is divided into two stages, and the flowchart is as shown in the appendix Figure 1As shown below. The specific implementation steps are as follows:

[0034] Step 1: At the transmitter, use the improved GS pre-EDC to pre-distort the signal, and at the receiver, use the FFE for post-equalization. Store the tap coefficients of the FFE as h BL ;

[0035] Step 2: Initialize the signal. Initialize the amplitude of the signal to the target amplitude A Target , and initialize the phase to zero. Among them, the target amplitude A Target is the amplitude after shaping the On-Off Keying (OOK) signal through a Raised Cosine (RC) filter with a roll-off factor of 1.0;

[0036] Step 3: Apply the receiver constraint to the amplitude of the odd samples of the signal, that is, replace it with A Target(odd) (n)-α·error(n). Among them, α is the amplitude error inversion factor, and its value range is 0 ≤ α ≤ 1. error is the amplitude error of the odd samples, and n is the sample index;

[0037] Step 4: Convert the signal to the transmitter side through the inverse dispersion frequency domain transfer function and the inverse filter ;

[0038] Step 5: Apply the transmitter constraint to the phase of the signal, that is, replace it with . Among them, β is the phase error inversion factor, and its value range is 0 ≤ β ≤ 1;

[0039] Step 6: Use the dispersion frequency domain transfer function H CD and the filter H BL to convert the signal back to the receiver side;

[0040] Step 7: Repeat steps 3 - 6 to complete N iterations. The amplitude A Tx (n) of the signal in step 5 is used as the final output after the iteration of this algorithm. This pre-distorted signal can cancel the impairments caused by bandwidth limitation and dispersion, eliminating the need for post-equalization using FFE at the receiver.

[0041] Among them, the dispersion frequency domain transfer function H CD is expressed as:

[0042]

[0043] Among them, D is the dispersion coefficient; λ is the optical carrier wavelength; ω is the angular frequency; l is the transmission distance; c is the speed of light in vacuum. The Z-domain expression of the filter H BL is:

[0044]

[0045] Among them, k is the number of taps.

[0046] Appendix Figure 2 And Appendix Figure 3 The pre-distortion signal spectrum and the received signal spectrum generated by a method for collaborative pre-compensation of bandwidth limitation and electrical dispersion based on an improved GS algorithm according to the present invention and the improved GS pre-EDC method are respectively compared. The improved GS pre-EDC method requires the use of FFE at the receiving end to perform post-equalization on the loss caused by bandwidth limitation, which also amplifies the high-frequency noise of the signal. However, the pre-distortion signal generated by a method for collaborative pre-compensation of bandwidth limitation and electrical dispersion based on an improved GS algorithm according to the present invention is slightly higher in the high-frequency part than that of the improved GS pre-EDC method, which can reduce the influence of system bandwidth limitation, thereby making the received signal spectrum flatter and eliminating the need for post-equalization.

[0047] In specific implementation, an experimental system is built as shown in Appendix Figure 4 . The system bandwidth is fixed, so different system bandwidth limitations are achieved by changing the baud rate of the OOK signal. At the sending end, first 2 15A random data stream is mapped to OOK symbols using Gray coding, upsampled by a factor of 2, and pulse-shaped by a raised cosine filter with a roll-off factor of 1. Then, a pre-distortion method for bandwidth limitation and electrical dispersion co-precompensation based on an improved GS algorithm of the present invention is used to pre-distort the signal. The pre-distorted digital signal is successively fed into an arbitrary waveform generator (AWG) with a bandwidth of 20 GHz and a sampling rate of 65 GSa / s, an electrical amplifier (EA) with a gain of 19 dB and a bandwidth of 12 GHz, and a Mach-Zehnder modulator (MZM) with a bandwidth of 40 GHz. An optical carrier with a wavelength of 1500 nm and a linewidth of 10 MHz is generated by an external cavity laser (ECL). The first erbium-doped optical fiber amplifier (EDFA) amplifies the optical power of the signal to 5 dBm. After transmission through 50 km of optical fiber, the second EDFA is used to compensate for the transmission loss, and an optical bandpass filter (OBPF) with a bandwidth of 0.4 nm is used to filter out out-of-band noise. The received optical power is adjusted by a variable optical attenuator (VOA). Subsequently, the signal passes through a positive intrinsic-negative (PIN) photodiode with a transimpedance amplifier (TIA) and a bandwidth of 40 GHz, and a digital phosphor oscilloscope (DPO) with a bandwidth of 36 GHz and a sampling rate of 80 GSa / s. Finally, offline DSP is performed on the signal, including resampling, frame synchronization, downsampling, OOK demapping, and bit error ratio (BER) calculation. The inset shows the frequency response of the optical back-to-back channel, which reveals that the 3 dB bandwidth of the system is approximately 9 GHz and the 10 dB bandwidth is approximately 16 GHz. This indicates that the available bandwidth is insufficient for 20 Gbit / s, 28 Gbit / s, and 32 Gbit / s OOK signals.

[0048] Appendix Figure 5The BER performances obtained by three methods at different baud rates are compared: the method without pre - equalization, the improved GS pre - EDC method, and a co - pre - compensation method for bandwidth limitation and chromatic dispersion based on an improved GS algorithm of the present invention. Among them, the method without pre - equalization and the improved GS pre - EDC method need to be combined with FFE post - equalization, and the number of taps of FFE is optimized to the best according to BER. For the method without pre - equalization, the signals at all baud rates cannot reach the Hard - Decision Forward Error Correction (HD - FEC) threshold. The improved GS pre - EDC method can reduce BER, but for the 32 Gbit / s signal, due to more severe bandwidth limitation, the BER cannot reach the HD - FEC threshold. However, a co - pre - compensation method for bandwidth limitation and chromatic dispersion based on an improved GS algorithm of the present invention shows better BER performance, enabling the 32 Gbit / s signal to meet the HD - FEC threshold without post - equalization. The results show that in an IM / DD system with limited bandwidth, a co - pre - compensation method for bandwidth limitation and chromatic dispersion based on an improved GS algorithm of the present invention improves the achievable symbol rate.

[0049] Appendix Figure 6 The received sensitivities of the three methods at different baud rates are further compared. For the 20 Gbit / s signal shown in Appendix Figure 6 a, it can be observed that compared with the improved GS pre - EDC method, a co - pre - compensation method for bandwidth limitation and chromatic dispersion based on an improved GS algorithm of the present invention provides a received sensitivity gain of 1.0 dB at the HD - FEC threshold. For the 28 Gbit / s signal shown in Appendix Figure 6 b, the bandwidth limitation of the system becomes severe, while a co - pre - compensation method for bandwidth limitation and chromatic dispersion based on an improved GS algorithm of the present invention still has an advantage in received sensitivity of 1.0 dB. For the 32 Gbit / s signal shown in Appendix Figure 6 c, the improved GS pre - EDC method can hardly be lower than the HD - FEC threshold, while a co - pre - compensation method for bandwidth limitation and chromatic dispersion based on an improved GS algorithm of the present invention can reach the HD - FEC threshold at an ROP of - 8.0 dBm, providing a received sensitivity gain of 2.0 dB. The results show that a co - pre - compensation method for bandwidth limitation and chromatic dispersion based on an improved GS algorithm of the present invention significantly reduces the bandwidth requirements for optoelectronic devices in high - speed systems.

[0050] In summary, the above are only the preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A bandwidth limitation and electronic dispersion cooperative pre-compensation method based on an improved GS algorithm is proposed. A filter for equalizing bandwidth limitation is designed and integrated into an improved GS pre-EDC. The generated pre-distortion signal is used for bandwidth limitation and electronic dispersion cooperative pre-compensation. The method is characterized by: The steps include: Step 1: Use the improved GS pre-EDC to predistort the signal at the transmitter, use FFE for post-equalization at the receiver, and store the FFE tap coefficients as h BL ; Step 2: Initialize the signal and set the amplitude of the signal to the target amplitude A. Target , the phase is initialized to zero; Step 3: Perform receiver constraints on the amplitude of the odd samples of the signal, that is, use A Target(odd) (n)-α·error(n), where α is the amplitude error inversion factor, error is the amplitude error of odd samples, and n is the sample index; Step 4: Use the anti-dispersion frequency domain transfer function and anti-filter Convert the signal to the sender; Step 5: Perform transmitter constraints on the phase of the signal, i.e., use Instead, where β is the phase error inversion factor; Step 6: Use the dispersion frequency domain transfer function H CD and filter H BL Convert the signal back to the receiving end; Step 7: Repeat steps 3-6 to complete N iterations.

2. According to claim 1, a bandwidth limitation and electric dispersion coordinated pre-compensation method based on an improved GS algorithm is characterized by: In step 2, the target amplitude A Target It is the amplitude of the OOK signal after being shaped by a raised cosine filter with a roll-off factor of 1.

0.

3. The method for coordinated pre-compensation of bandwidth limitation and electric dispersion based on improved GS algorithm according to claim 1, characterized in that: In step 3, the value range of α is 0≤α≤1.

4. The method for coordinated pre-compensation of bandwidth limitation and electric dispersion based on improved GS algorithm according to claim 1, characterized in that: In step 5, the value range of β is 0≤β≤1.

5. The method for coordinated pre-compensation of bandwidth limitation and electric dispersion based on improved GS algorithm according to claim 1, characterized in that: In step 6, the dispersion frequency domain transfer function H CD It is expressed as: Among them, D is the dispersion coefficient, λ is the wavelength of the optical carrier, ω is the circular frequency, l is the transmission distance, and c is the speed of light in vacuum.

6. The method for coordinated pre-compensation of bandwidth limitation and electric dispersion based on improved GS algorithm according to claim 1, characterized in that: In step 6, filter H BL The Z domain expression is: Among them, k is the number of taps.

7. The method for coordinated pre-compensation of bandwidth limitation and electric dispersion based on improved GS algorithm according to claim 1, characterized in that: In step 5, the amplitude of the signal A Tx (n) As the final output after the algorithm is iterated, the predistorted signal can offset the impairment caused by bandwidth limitation and dispersion, eliminating the need for post-FFE equalization at the receiving end.

Images