Lithium niobate coherent modulator control method, device and equipment based on gradient algorithm

Abstract

The invention discloses a lithium niobate coherent modulator control method, device and equipment based on a gradient algorithm, and relates to the technical field of electro-optical modulators, and the method comprises the steps: setting the phases of an I path and a Q path of a lithium niobate IQ modulator to be the same based on a preprocessing algorithm; setting bias voltages of an I path, a Q path and a P path of the lithium niobate IQ modulator at optimal bias points through a bias voltage locking algorithm; and judging the working state of the lithium niobate IQ modulator under the current bias voltage based on a feedback algorithm, and performing error compensation adjustment to enable the working state of the lithium niobate IQ modulator to be optimal. According to the invention, precise control of the bias voltage of the modulator can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of electro-optic modulators, and in particular to a method, device and equipment for controlling a lithium niobate coherent modulator based on a gradient algorithm. Background Art

[0002] Lithium niobate optical modulators have the advantages of wide bandwidth, fast response speed, low insertion loss, high signal-to-noise ratio, low half-wave voltage, and good stability. They are one of the mainstream electro-optical modulator products and are widely used in optical communications, fiber optic sensors and other fields. In high-speed optical communication systems, lithium niobate IQ (in phase-quadrature) modulators have become key components because they can realize complex modulation formats. However, in order to ensure the efficient operation of the modulator, its bias voltage needs to be precisely controlled, which is a prerequisite for achieving high-quality optical signal transmission.

[0003] Traditional bias voltage control methods mainly rely on manual adjustment or simple automatic control strategies. These methods are often difficult to adapt to rapidly changing environmental conditions, resulting in unstable modulator performance and affecting communication quality. As optical communication systems continue to increase their requirements for signal quality, the requirements for modulator bias voltage control technology are also getting higher and higher. Therefore, how to achieve precise control of modulator bias voltage has become an urgent problem to be solved. Summary of the invention

[0004] The present application provides a lithium niobate coherent modulator control method, device and equipment based on a gradient algorithm, which can achieve precise control of the modulator bias voltage.

[0005] In a first aspect, an embodiment of the present application provides a lithium niobate coherent modulator control method based on a gradient algorithm, which is used to modulate a lithium niobate IQ modulator of an IQ modulation structure in coherent modulation. The lithium niobate coherent modulator control method based on a gradient algorithm includes:

[0006] Based on the preprocessing algorithm, the phases of the I and Q paths of the lithium niobate IQ modulator are set to be the same;

[0007] The bias voltage of I, Q and P paths of the lithium niobate IQ modulator is set at the optimal bias point through the bias voltage locking algorithm;

[0008] Based on the feedback algorithm, the working state of the lithium niobate IQ modulator under the current bias voltage is judged, and error compensation adjustment is performed to make the working state of the lithium niobate IQ modulator optimal.

[0009] In combination with the first aspect, in one implementation, the phases of the I path and the Q path of the lithium niobate IQ modulator are set to be the same based on the preprocessing algorithm, specifically including:

[0010] S101: Scan the entire working bias voltage range of the lithium niobate IQ modulator, output the optical power versus bias voltage curve, obtain the basic characteristic data of the lithium niobate IQ modulator, set the learning rate, convergence coefficient, convergence threshold, and the initial bias voltage of each channel, record the initial optical power, and go to S102;

[0011] S102: Calculate the gradient of the optical power of the I-path, Q-path, and P-path compared to the initial bias voltage, and move the bias voltages of the I-path, Q-path, and P-path in the direction of the maximum rising gradient of the optical power according to the set moving step length, obtain the bias voltages of the I-path, Q-path, and P-path after the movement, and record the optical power after the movement, and go to S103;

[0012] S103: Compare the current optical power after the move with the latest initial optical power. If the optical power after the move is greater than the initial optical power, go to S105. If the optical power after the move is not greater than the initial optical power, and the latest learning rate is greater than the convergence threshold, go to S104. If the optical power after the move is not greater than the initial optical power, and the latest learning rate is not greater than the convergence threshold, then before the last I-path, Q-path, and P-path bias voltage move, the bias voltages of the I-path, Q-path, and P-path obtained after the move are the bias voltages corresponding to the maximum optical power, indicating that the bias voltages of the I-path, Q-path, and P-path are set to the bias voltages corresponding to the maximum optical power, and the process ends.

[0013] S104: Divide the current learning rate by the convergence coefficient to obtain a value as the latest learning rate, and go to S105;

[0014] S105: taking the current optical power value after the shift as the latest initial optical power, and going to S102;

[0015] Among them, the set moving step size corresponding to the bias voltage of the I path is the latest learning rate multiplied by the gradient of the optical power of the I path compared to the initial bias voltage, the set moving step size corresponding to the bias voltage of the Q path is the latest learning rate multiplied by the gradient of the optical power of the Q path compared to the initial bias voltage, and the set moving step size corresponding to the bias voltage of the P path is the latest learning rate multiplied by the gradient of the optical power of the P path compared to the initial bias voltage.

[0016] In combination with the first aspect, in one embodiment, the bias voltage of the I-path, Q-path, and P-path of the lithium niobate IQ modulator is set at the optimal bias point through the bias voltage locking algorithm, specifically including: locking the I-path and Q-path bias voltages based on a gradient descent algorithm, locking the P-path bias voltage based on a sweep spectrum algorithm, and locking the I-path, Q-path, and P-path bias voltages based on a pilot algorithm.

[0017] In combination with the first aspect, in one implementation, the locking of the I-path and Q-path bias voltages based on the gradient descent algorithm specifically includes:

[0018] S201: when the bias voltage corresponding to the maximum optical power is obtained, the currently recorded optical power after movement is used as the first optical power, and the process goes to S202;

[0019] S202: Calculate the gradient of the optical power of the I and Q paths compared to the initial bias voltage, and move the bias voltages of the I and Q paths in the direction of the maximum decreasing gradient of the optical power according to the preset moving step length, obtain the bias voltages of the I and Q paths after the movement, and record the optical power after the movement, and go to S203;

[0020] S203: Compare the current optical power after the move with the latest first optical power. If the optical power after the move is less than the first optical power, go to S205. If the optical power after the move is not less than the first optical power and the latest learning rate is greater than the convergence threshold, go to S204. If the optical power after the move is not less than the first optical power and the latest learning rate is not greater than the convergence threshold, the bias voltages of the I and Q paths obtained before the last bias voltage movement of the I, Q and P paths are obtained are the bias voltages corresponding to the minimum optical power, indicating that the bias voltages of the I and Q paths are set to the bias voltages corresponding to the minimum optical power, and the process ends.

[0021] S204: Divide the current learning rate by the convergence coefficient to obtain a value as the latest learning rate, and go to S205;

[0022] S205: taking the current optical power value after the movement as the latest first optical power, and going to S202;

[0023] Among them, the preset moving step size corresponding to the bias voltage of the I path is the latest learning rate multiplied by the gradient of the optical power of the I path compared to the initial bias voltage, and the preset moving step size corresponding to the bias voltage of the Q path is the latest learning rate multiplied by the gradient of the optical power of the Q path compared to the initial bias voltage.

[0024] In combination with the first aspect, in one implementation, the locking of the P-circuit bias voltage based on the spectrum sweep algorithm specifically includes:

[0025] Set the bias voltage of the I and Q paths to the bias voltage corresponding to the minimum optical power, and set the bias voltage of the P path to the bias voltage corresponding to the maximum optical power;

[0026] Sweep the bias voltage of the P path and store the optical power, obtain any pair of adjacent optical power maximum and optical power minimum values ​​in the stored optical power, and calculate the target optical power of the Quad point based on the optical power maximum and optical power minimum values;

[0027] In the bias voltage interval corresponding to the maximum optical power and the minimum optical power, the bias voltage point closest to the Quad point target optical power is obtained, determined as the Quad point of the P path, and the bias voltage of the P path is locked as the Quad point.

[0028] In combination with the first aspect, in one implementation,

[0029] The pilot algorithm is used to lock the bias voltages of the I-channel, Q-channel and P-channel, wherein the sine wave required by the pilot is simulated by the DAC of the bias control circuit, and the sine wave of the specified frequency is simulated by the MCU timer trigger;

[0030] The simulation of the pilot signal includes:

[0031] Set up the discrete sine signal array, add pilot duration, DC bias voltage, pilot sine amplitude, and set the counting threshold according to the pilot sine frequency;

[0032] Start the MCU timer interrupt, start the timing from 0, and set the MCU timer count to 0. The internal clock of the MCU timer triggers the count to increase, and an interrupt is generated when the count reaches the count threshold;

[0033] Perform interrupt processing and generate pilot bias voltage, specifically:

[0034] VH=VDC+Sin[i]*A

[0035] Wherein, VH represents the pilot bias voltage, VDC represents the DC bias voltage, A represents the pilot sine amplitude, Sin[i] represents the discrete sine signal array generated in real time, and when the length i of the discrete sine signal array generated in real time exceeds the set length of the discrete sine signal array, i is reset to 0, and when i exceeds the length of the discrete sine signal array, i is reset to 0, and when the timing reaches the pilot duration, the MCU timer interrupt is turned off, and the simulation of generating the pilot signal ends.

[0036] In combination with the first aspect, in one implementation, the locking of the bias voltages of the I-path, Q-path, and P-path based on the pilot algorithm specifically includes:

[0037] S221: Load frequency f on path I I After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I The normalized optical power at , goes to S222;

[0038] S222: Determine frequency f IIs the normalized optical power at less than the set threshold? If so, go to S223; if not, adjust the bias voltage of the I path and go to S222;

[0039] S223: Load frequency f on Q path Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. Q The normalized optical power at , goes to S224;

[0040] S224: Determine frequency f Q Whether the normalized optical power at is less than the set threshold, if so, go to S225, if not, adjust the bias voltage of the Q path, and go to S223;

[0041] S225: Load frequency f on I path I A sinusoidal pilot signal with a frequency of f is loaded on the Q path. Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I +f Q The normalized optical power and |f I -f Q |, go to S226;

[0042] S226: frequency f I +f Q The normalized optical power and |f I -f Q | whether the normalized optical power at the locations is less than the set threshold value; if not, adjust the bias voltage of the P path and go to S225; if so, go to S227;

[0043] S227: Execute judgment: If the frequency loaded on path I is f I After the sinusoidal pilot signal is sent, the bias voltage of the I path is not adjusted, and the frequency f is loaded on the Q path. Q After the sinusoidal pilot signal is sent, the bias voltage of the Q circuit is not adjusted, and at the same time, the frequency f is loaded on the I circuit. I The sinusoidal pilot signal and the frequency f loaded on the Q path Q If the bias voltage of the P circuit is not adjusted after the sinusoidal pilot signal, the process ends; otherwise, the process goes to S221.

[0044] In combination with the first aspect, in one implementation, judging the working state of the lithium niobate IQ modulator under the current bias voltage based on the feedback algorithm and performing error compensation adjustment to optimize the working state of the lithium niobate IQ modulator specifically includes:

[0045] S301: Setting the optical power change threshold and the normalized optical power threshold, and setting the bias voltages of the I, Q, and P paths at the optimal bias points to obtain the optical power corresponding thereto as the optimal optical power, and then proceeding to S302;

[0046] S302: Read the current optical power regularly, calculate the change between the current optical power and the optimal optical power, and go to S303;

[0047] S303: Determine whether the calculated change exceeds the optical power change threshold, if so, go to S304, if not, go to S302;

[0048] S304: Loading frequency f on path I I After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I The normalized optical power at the Q path is loaded with a frequency of f Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. Q The normalized optical power at the location, and the frequency f is loaded on the I path I The sinusoidal pilot signal and the frequency f loaded on the Q path Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I +f Q The normalized optical power and |f I -f Q The normalized optical power at |, according to the recorded frequency f I Normalized optical power at frequency f Q Normalized optical power at frequency f I +f Q Normalized optical power at frequency |f I -f Q | Whether the normalized optical power at exceeds the normalized optical power threshold to determine whether the working point is offset, if so, go to S305, if not, go to S306;

[0049] S305: Lock the bias voltages of the I-channel, Q-channel, and P-channel based on the pilot algorithm, perform offset correction, and then go to S306;

[0050] S306: based on the change of the optimal optical power caused by the re-locking of the working point or the change of the overall optical power, update the optimal optical power and go to S302.

[0051] In a second aspect, an embodiment of the present application provides a lithium niobate coherent modulator control device based on a gradient algorithm, and the lithium niobate coherent modulator control device based on a gradient algorithm includes:

[0052] A first setting module, which is used to set the phases of the I path and the Q path of the lithium niobate IQ modulator to be the same based on a preprocessing algorithm;

[0053] A second setting module, which is used to set the bias voltages of the I-path, Q-path, and P-path of the lithium niobate IQ modulator at an optimal bias point through a bias voltage locking algorithm;

[0054] The execution module is used to judge the working state of the lithium niobate IQ modulator under the current bias voltage based on the feedback algorithm, and perform error compensation adjustment to make the working state of the lithium niobate IQ modulator optimal.

[0055] In a third aspect, an embodiment of the present application provides a lithium niobate coherent modulator control device based on a gradient algorithm, wherein the lithium niobate coherent modulator control device based on a gradient algorithm comprises a processor, a memory, and a lithium niobate coherent modulator control program based on a gradient algorithm stored in the memory and executable by the processor, wherein when the lithium niobate coherent modulator control program based on a gradient algorithm is executed by the processor, the steps of the lithium niobate coherent modulator control method based on a gradient algorithm are implemented.

[0056] The beneficial effects brought by the technical solution provided in the embodiments of the present application include:

[0057] (1) The gradient ascent algorithm is used for preprocessing, and the gradient descent algorithm is used to find the optimal bias voltage point of the modulator. The bias voltages of different paths can be adjusted synchronously, and the influence between paths can be compensated in real time. Compared with the ordinary thermal spectrum scanning algorithm, the time and complexity of iterative adjustment are greatly reduced. In addition, the dynamic nature of the gradient algorithm enables it to adapt to IQ modulators with similar bias voltage ranges, and has good compatibility.

[0058] (2) Using DAC to simulate the sine wave required by the pilot method, and using digital FFT filtering instead of hardware filtering, no additional chip devices are required, and the detection quality meets the detection requirements of the pilot method; compared with the method of using a function generator to generate a sine signal and hardware filtering detection, it can save material costs, reduce the size of the control circuit board, and improve the locking accuracy without complicating the circuit design, which is conducive to integrated applications;

[0059] (3) By designing an automatic optimization algorithm, it is effective in reducing the impact of external interference and heat accumulation effects on the modulator during long-term operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1Design schematic diagram for bias voltage system of lithium niobate IQ modulator;

[0061] Figure 2 This is a flow chart of a lithium niobate coherent modulator control method based on a gradient algorithm in this application;

[0062] Figure 3 This is a functional module diagram of a lithium niobate coherent modulator control device based on a gradient algorithm in this application;

[0063] Figure 4 This is a schematic diagram of the hardware structure of the lithium niobate coherent modulator control device based on the gradient algorithm of this application. DETAILED DESCRIPTION

[0064] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0065] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0066] In the first aspect, an embodiment of the present application provides a lithium niobate coherent modulator control method based on a gradient algorithm, which is used to modulate the lithium niobate IQ modulator of the IQ modulation structure in coherent modulation, monitor the optical power by measuring and controlling the photodiode at the output end of the lithium niobate IQ modulator, find the optimal working point through the bias voltage control algorithm, and lock each bias voltage to the optimal working point to achieve high-speed signal modulation.

[0067] The lithium niobate coherent modulator control method of the present application can reduce iteration time and complexity, quickly and accurately lock the optimal bias point, and combine the pilot method in the control process to help achieve accurate locking and real-time optimization of the optimal working point. In the traditional pilot method, function generator chips, operational amplifiers and other devices are required to form a sine wave generation circuit. The present application uses a DAC (digital-to-analog converter) to simulate the sine wave required by the pilot method, and uses a digital FFT (Fast Fourier Transform) for filtering. This method reuses multiple DACs and multiple ADCs (analog-to-digital converters), and there is no need to design a sine wave generation circuit and a pilot detection circuit. Compared with the method of using a function generator to generate a sine signal and hardware filtering, the circuit design is simplified while ensuring the detection quality, which can save material costs, reduce the size of the control circuit board, and is conducive to integrated applications.

[0068] First of all, it should be noted that the lithium niobate IQ modulator in this application adopts a typical IQ modulation structure in coherent modulation, see Figure 1 As shown in FIG. 1 , the lithium niobate IQ modulator consists of two sub-MZI (multi-mode interferometer) structures and one parent MZI structure. The upper and lower sub-MZI structures in the structure are called I path and Q path respectively, and the parent MZI structure is called P path. When the modulator is working, the phase difference between the I path and the Q path is 90 degrees.

[0069] The common operating points of Mach-Zehnder modulators include the slope quadrature point (Quad point for short), the optical power minimum point (Null point for short), and the optical power maximum point (Max point for short). For common modulation code types such as QPSK and PAM-4, the bias voltage points of the I and Q paths need to be set at the Null point, and the bias voltage point of the P path needs to be set at the Quad point.

[0070] The bias voltage control system of lithium niobate IQ modulator includes a lithium niobate IQ modulator chip, a bias voltage control circuit and a bias voltage control algorithm. The design diagram is shown in FIG. Figure 1 shown. Figure 1 In the figure, number 1 is a lithium niobate IQ modulator; number 2 is an optical waveguide; number 3 is a multimode interferometer (MMI), which is used as a beam splitter and combiner in the modulator; number 4 is a thermal phase shifter of the I path, which is used to control the phase of the I path; number 5 is a thermal phase shifter of the Q path, which is used to control the thermal phase shifter of the Q path; number 6 is a thermal phase shifter of the P path, which is used to control the phase difference between the I path and the Q path; number 7 is a monitoring photodiode (MPD), which is used to detect the optical power and convert it into photocurrent. It should be noted that the MPD can be an integrated MPD on the chip or an external MPD device. In actual applications, at least one of the two forms can be selected. Figure 1 Regarding the MPD settings, the MPD on the left corresponds to the MPD form integrated on the chip, and the MPD on the right corresponds to the external MPD device form.

[0071] For the bias voltage control circuit: the photocurrent is converted into an analog voltage signal through a transimpedance amplifier (TIA) circuit; the analog voltage signal is converted into a digital signal through an analog-to-digital converter (ADC) and sent to the main microcontroller unit (MCU); after the main control runs the control algorithm, the digital bias voltage signal is sent to a multi-channel digital-to-analog converter (DAC); the multi-channel DAC converts the digital bias voltage signal into an analog bias voltage and loads it onto the I-channel, Q-channel and P-channel thermal phase shifters.

[0072] In one embodiment, referring to Figure 2 , Figure 2 This is a flow chart of the control method of the lithium niobate coherent modulator based on the gradient algorithm of this application. Figure 2As shown, the control method of lithium niobate coherent modulator based on gradient algorithm includes:

[0073] S1: Based on the preprocessing algorithm, the phases of the I and Q paths of the lithium niobate IQ modulator are set to be the same;

[0074] The lithium niobate coherent modulator control method of the present application specifically includes three parts: preprocessing, bias voltage locking and automatic optimization.

[0075] In step S1, the purpose of the preprocessing algorithm is to set the phase of the I-path and Q-path of the lithium niobate IQ modulator to be the same, so as to facilitate the subsequent locking. First, the entire working bias voltage range of the lithium niobate IQ modulator is scanned, and the output optical power versus bias voltage curve is stored to obtain the basic characteristic data of the lithium niobate IQ modulator, and the initial bias voltage is set, and then the gradient rise method is used to lock the I-path, Q-path, and P-path bias voltages at the maximum optical power point, at which time the I-path and Q-path of the lithium niobate IQ modulator are in the same phase.

[0076] S2: By using a bias voltage locking algorithm, the bias voltages of the I, Q and P paths of the lithium niobate IQ modulator are set at the optimal bias point;

[0077] In step S2, the bias voltage locking algorithm is used to achieve the optimal bias point locking of I-way, Q-way, and P-way. First, a preliminary locking is performed, and the bias voltages of I-way and Q-way are locked at the Null point using the gradient algorithm, and then the bias voltage of P-way is locked at the Quad point by scanning the spectrum analysis data, and then precise locking is performed, and pilot signals are added to I-way and Q-way, and the optical power is sampled and FFT transformation is performed. The spectrum analysis can accurately determine whether the bias voltages of I-way, Q-way, and P-way are at the optimal working point. If not, fine-tune it, and repeat this step until I-way, Q-way, and P-way are all at the optimal working point. The present application also provides a method for using DAC to simulate the generation of the sine wave required by the pilot method. This method does not require additional equipment. Compared with the method of using a function generator to generate a sine signal, it can save material costs, reduce the size of the control circuit board, and improve the locking accuracy without complicating the circuit design, which is conducive to integrated applications.

[0078] S3: Based on the feedback algorithm, the working state of the lithium niobate IQ modulator under the current bias voltage is judged, and the error compensation adjustment is performed to make the working state of the lithium niobate IQ modulator optimal. That is, through the automatic optimization algorithm, based on the principle of feedback control, the lithium niobate IQ modulator is guaranteed to be in the best working state under the influence of long-term operation and environmental factors.

[0079] Step S3 uses a feedback algorithm to determine whether the working state of the lithium niobate IQ modulator is optimal under the current bias voltage, and makes adjustments to compensate for errors caused by long-term operation of the modulator and environmental influences.

[0080] Further, in one embodiment, based on the preprocessing algorithm, the phases of the I path and the Q path of the lithium niobate IQ modulator are set to be the same, specifically including:

[0081] S101: Scan the entire working bias voltage range of the lithium niobate IQ modulator, output the optical power versus bias voltage curve, obtain the basic characteristic data of the lithium niobate IQ modulator, set the learning rate, convergence coefficient, convergence threshold, and the initial bias voltage of each channel, record the initial optical power, and go to S102;

[0082] It should be noted that, since the value of the learning rate may change later, the learning rate set in step S101 is the initial learning rate;

[0083] S102: Calculate the gradient of the optical power of the I-path, Q-path, and P-path compared to the initial bias voltage, and move the bias voltages of the I-path, Q-path, and P-path in the direction of the maximum rising gradient of the optical power according to the set moving step length, obtain the bias voltages of the I-path, Q-path, and P-path after the movement, and record the optical power after the movement, and go to S103;

[0084] S103: Compare the current optical power after the move with the latest initial optical power. If the optical power after the move is greater than the initial optical power, it means that the bias voltage is moving in the direction of increasing the optical power. Then go to S105. If the optical power after the move is not greater than the initial optical power, and the latest learning rate is greater than the convergence threshold, it means that the optical power has exceeded the maximum value point and the gradient algorithm has not converged. Go to S104. If the optical power after the move is not greater than the initial optical power, and the latest learning rate is not greater than the convergence threshold, it means that the gradient algorithm has converged. At this time, retract the last bias voltage, that is, before the last bias voltage movement of the I-way, Q-way, and P-way, the bias voltages of the I-way, Q-way, and P-way after the move are obtained, which are the bias voltages corresponding to the maximum optical power, indicating that the bias voltages of the I-way, Q-way, and P-way are set to the bias voltages corresponding to the maximum optical power. End;

[0085] It should be noted that when the optical power after the first move is compared with the latest initial optical power, the optical power after the move refers to the optical power after the move recorded after the first move, and the initial optical power is the initial optical power recorded for the first time. For the comparison between the subsequent optical power after the move and the latest initial optical power, if the value of the initial optical power has changed, the latest initial optical power refers to the initial optical power after the change; the current optical power after the move refers to the optical power after the move recorded after the latest move;

[0086] S104: The value obtained by dividing the current learning rate by the convergence coefficient is used as the latest learning rate, and the process goes to S105; that is, the value obtained by dividing the current learning rate by the convergence coefficient is used as the latest learning rate value;

[0087] S105: taking the current value of the optical power after the movement as the latest initial optical power, and going to S102; that is, replacing the value of the initial optical power with the current value of the optical power after the movement;

[0088] Among them, the set moving step size corresponding to the bias voltage of the I path is the latest learning rate multiplied by the gradient of the optical power of the I path compared to the initial bias voltage, the set moving step size corresponding to the bias voltage of the Q path is the latest learning rate multiplied by the gradient of the optical power of the Q path compared to the initial bias voltage, and the set moving step size corresponding to the bias voltage of the P path is the latest learning rate multiplied by the gradient of the optical power of the P path compared to the initial bias voltage.

[0089] Furthermore, in one embodiment, the bias voltages of the I, Q, and P paths of the lithium niobate IQ modulator are set at the optimal bias point through a bias voltage locking algorithm, specifically including: locking the bias voltages of the I and Q paths based on a gradient descent algorithm, locking the bias voltage of the P path based on a sweep spectrum algorithm, and locking the bias voltages of the I, Q, and P paths based on a pilot algorithm.

[0090] Furthermore, in one embodiment, the I-path and Q-path bias voltages are locked based on a gradient descent algorithm, specifically including:

[0091] S201: when the bias voltage corresponding to the maximum optical power is obtained, the currently recorded optical power after movement is used as the first optical power, and the process goes to S202;

[0092] In step S1, the bias voltages of the I path, Q path, and P path are set to the bias voltages corresponding to the maximum optical power;

[0093] S202: Calculate the gradient of the optical power of the I and Q paths compared to the initial bias voltage, and move the bias voltages of the I and Q paths in the direction of the maximum decreasing gradient of the optical power according to the preset moving step length, obtain the bias voltages of the I and Q paths after the movement, and record the optical power after the movement, and go to S203;

[0094] S203: Compare the current optical power after the move with the latest first optical power. If the optical power after the move is less than the first optical power, it means that the bias voltage is moving in the direction of reducing the optical power. Go to S205. If the optical power after the move is not less than the first optical power and the latest learning rate is greater than the convergence threshold, it means that the optical power has crossed the minimum point and the gradient algorithm has not converged. Go to S204. If the optical power after the move is not less than the first optical power and the latest learning rate is not greater than the convergence threshold, it means that the gradient algorithm has converged. At this time, retract the last bias voltage, that is, before the last bias voltage movement of the I-way, Q-way and P-way, the bias voltages of the I-way and Q-way after the move are obtained, which are the bias voltages corresponding to the minimum optical power, indicating that the bias voltages of the I-way and Q-way are set to the bias voltages corresponding to the minimum optical power. End;

[0095] S204: Divide the current learning rate by the convergence coefficient to obtain a value as the latest learning rate, and go to S205;

[0096] S205: taking the current optical power value after the movement as the latest first optical power, and going to S202;

[0097] Among them, the preset moving step size corresponding to the bias voltage of the I path is the latest learning rate multiplied by the gradient of the optical power of the I path compared to the initial bias voltage, and the preset moving step size corresponding to the bias voltage of the Q path is the latest learning rate multiplied by the gradient of the optical power of the Q path compared to the initial bias voltage.

[0098] Furthermore, in one embodiment, locking the P-circuit bias voltage based on a spectrum sweep algorithm specifically includes:

[0099] S211: setting the bias voltages of the I path and the Q path to the bias voltages corresponding to the minimum optical power, and setting the bias voltage of the P path to the bias voltage corresponding to the maximum optical power;

[0100] In steps S201 to S205, the bias voltages of the I and Q paths have been set to the bias voltages corresponding to the minimum optical power, i.e., the Null point; in step S1, the bias voltage of the P path has been set to the Max point;

[0101] S212: Sweep the bias voltage of the P path and store the optical power, obtain any pair of adjacent optical power maximum values ​​and optical power minimum values ​​in the stored optical power, and calculate the target optical power of the Quad point according to the optical power maximum value and the optical power minimum value;

[0102] Half of the sum of the maximum optical power and the minimum optical power is the target optical power at the Quad point;

[0103] S213: In the bias voltage interval corresponding to the maximum optical power and the minimum optical power, obtain the bias voltage point closest to the Quad point target optical power, determine it as the Quad point of the P path, and lock the bias voltage of the P path to the Quad point. At this time, the I path and the Q path are orthogonal.

[0104] Furthermore, in one embodiment, the bias voltages of the I-channel, Q-channel, and P-channel are locked based on a pilot algorithm, wherein the sine wave required for the pilot is simulated by a DAC of a bias control circuit, and a sine wave of a specified frequency is simulated by triggering an MCU timer;

[0105] The simulation of the pilot signal includes:

[0106] S21: Set the discrete sine signal array, add the pilot duration, DC bias voltage, pilot sine amplitude, and set the counting threshold according to the pilot sine frequency;

[0107] Sine wave signals with different amplitudes and frequencies can be generated by modifying the amplitude and counting threshold;

[0108] S22: Start MCU timer interrupt, start timing from 0, and set the MCU timer count to 0. The internal clock of the MCU timer triggers the count to increase, and an interrupt is generated when the count reaches the count threshold;

[0109] S23: Perform interrupt processing and generate a pilot bias voltage. Specifically:

[0110] VH=VDC+Sin[i]*A

[0111] Among them, VH represents the pilot bias voltage, VDC represents the DC bias voltage, A represents the pilot sine amplitude, Sin[i] represents the discrete sine signal array generated in real time, and the length i of the discrete sine signal array generated in real time exceeds the length of the set discrete sine signal array, and i is set to 0, and when the timing reaches the pilot duration, the MCU timer interrupt is turned off, and the simulation of the pilot signal generation ends. The DAC simulates the pilot signal through the MCU timer interrupt signal, which does not require additional components and power supply circuits, simplifies the circuit design, and saves space.

[0112] Furthermore, in one embodiment, the I-path, Q-path, and P-path bias voltages are locked based on a pilot algorithm, specifically including:

[0113] S221: Load frequency f on path I I After collecting the optical power of a set period of time (i.e. collecting the optical power of a short period of time), stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I The normalized optical power at , goes to S222;

[0114] S222: Determine frequency f I Whether the normalized optical power at is less than the set threshold, if so, go to S223, if not, adjust the bias voltage of the I path (specifically, fine-tune the bias voltage of the I path), and go to S222;

[0115] S223: Load frequency f on Q path Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. Q The normalized optical power at , goes to S224;

[0116] S224: Determine frequency f Q Whether the normalized optical power at is less than the set threshold, if so, go to S225, if not, adjust the bias voltage of the Q path (specifically, fine-tune the bias voltage of the Q path), and go to S223;

[0117] S225: Load frequency f on I path I A sinusoidal pilot signal with a frequency of f is loaded on the Q path. Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I +f Q The normalized optical power and |f I -f Q |, go to S226;

[0118] S226: frequency f I +f Q The normalized optical power and |f I -f Q | whether the normalized optical power at the locations is less than the set threshold value; if not, adjust the bias voltage of the P path (specifically, fine-tune the bias voltage of the P path), and go to S225; if yes, go to S227;

[0119] S227: Execute judgment: If the frequency loaded on path I is f I After the sinusoidal pilot signal is sent, the bias voltage of the I path is not adjusted, and the frequency f is loaded on the Q path. Q After the sinusoidal pilot signal is sent, the bias voltage of the Q circuit is not adjusted, and at the same time, the frequency f is loaded on the I circuit. I The sinusoidal pilot signal and the frequency f loaded on the Q path Q If the bias voltage of the P circuit is not adjusted after the sinusoidal pilot signal, the process ends; otherwise, the process goes to S221.

[0120] That is, after loading the pilot signal three times, no fine-tuning operation is performed, which means that the I, Q, and P paths are all at the optimal working point and the algorithm ends; if fine-tuning is performed, in order to eliminate inter-path interference, it is necessary to jump to step S221 for further detection and verification.

[0121] The collected optical power is sent to the MCU timer for FFT to find the corresponding pilot frequency. Compared with the hardware circuit filtering method to collect the optical power at the pilot, the circuit is simpler and the detection is more flexible. At this point, the bias voltage locking algorithm ends, the I path and the Q path are at the Null point, the P path is at the Quad point, the I path and the Q path are orthogonal and the optical power is the smallest. At this time, the loading signal effect is best.

[0122] Further, in one embodiment, the working state of the lithium niobate IQ modulator under the current bias voltage is judged based on the feedback algorithm, and error compensation adjustment is performed to optimize the working state of the lithium niobate IQ modulator, specifically including:

[0123] S301: Setting the optical power change threshold and the normalized optical power threshold, and setting the bias voltages of the I, Q, and P paths at the optimal bias points to obtain the optical power corresponding thereto as the optimal optical power, and then proceeding to S302;

[0124] S302: Read the current optical power regularly, calculate the change between the current optical power and the optimal optical power, and go to S303;

[0125] S303: Determine whether the calculated change exceeds the optical power change threshold, if so, go to S304, if not, go to S302;

[0126] S304: Loading frequency f on path I I After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I The normalized optical power at the Q path is loaded with a frequency of f Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. Q The normalized optical power at the location, and the frequency f is loaded on the I path I The sinusoidal pilot signal and the frequency f loaded on the Q path Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I +f Q The normalized optical power and |f I -f Q The normalized optical power at |, according to the recorded frequency f INormalized optical power at frequency f Q Normalized optical power at frequency f I +f Q Normalized optical power at frequency |f I -f Q | Whether the normalized optical power at exceeds the normalized optical power threshold value to determine whether the working point is offset. If so, it means that the overall optical power has changed, and the process goes to S305. If not, it means that the working state of the lithium niobate IQ modulator is still optimal, and the process goes to S306.

[0127] S305: Lock the bias voltages of the I-channel, Q-channel, and P-channel based on the pilot algorithm, perform offset correction, and then go to S306;

[0128] S306: based on the change of the optimal optical power caused by the re-locking of the working point or the change of the overall optical power, update the optimal optical power and go to S302.

[0129] It should be noted that the gradient algorithm is an ideal choice for bias voltage control of lithium niobate IQ modulators due to its efficient locking speed. The gradient algorithm calculates the gradient of the modulator output signal and quickly locates the bias voltage with the best signal quality to achieve efficient automatic adjustment. The efficiency of this method is reflected in its ability to synchronously adjust the bias voltages of different paths and compensate for the influence between paths in real time, greatly reducing the time and complexity of iterative adjustment. In addition, the dynamic nature of the gradient algorithm enables it to adapt to IQ modulators with different bias voltage ranges.

[0130] In order to further improve the accuracy of locking, the present application adopts the pilot method for precise locking. By adding pilot signals to the I and Q paths, performing FFT transformation after sampling the optical power, and analyzing the spectrum, it can accurately determine whether the bias voltages of the I, Q, and P paths are at the optimal working point. After the gradient algorithm quickly locks the bias voltage, it can further refine the adjustment to ensure the continuous operation of the modulator in the optimal working state, thereby improving the working stability of the modulator and the overall performance of the optical communication system.

[0131] This application utilizes an efficient gradient algorithm combined with precise locking of the pilot method to achieve dynamic and precise control of the bias voltage. Compared with traditional control methods, the control algorithm proposed in this application has better adaptability and stability, and can effectively cope with problems such as environmental changes and equipment aging, ensuring that the modulator can maintain the best working state under various conditions, thereby improving the reliability and transmission efficiency of the optical communication system.

[0132] In a second aspect, an embodiment of the present application also provides a lithium niobate coherent modulator control device based on a gradient algorithm.

[0133] In one embodiment, referring to Figure 3 , Figure 3 This is a functional module diagram of the lithium niobate coherent modulator control device based on the gradient algorithm of this application. Figure 3 As shown, the lithium niobate coherent modulator control device based on the gradient algorithm includes: a first setting module, a second setting module, and an execution module.

[0134] The first setting module is used to set the phase of the I and Q paths of the lithium niobate IQ modulator to be the same based on a preprocessing algorithm; the second setting module is used to set the bias voltages of the I, Q and P paths of the lithium niobate IQ modulator to the optimal bias point through a bias voltage locking algorithm; the execution module is used to judge the working state of the lithium niobate IQ modulator under the current bias voltage based on a feedback algorithm, and perform error compensation adjustment to make the working state of the lithium niobate IQ modulator optimal.

[0135] In a third aspect, an embodiment of the present application provides a lithium niobate coherent modulator control device based on a gradient algorithm. The lithium niobate coherent modulator control device based on a gradient algorithm can be a personal computer (PC), a laptop computer, a server, or other device with data processing capabilities.

[0136] Reference Figure 4 , Figure 4 The schematic diagram of the hardware structure of the lithium niobate coherent modulator control device based on the gradient algorithm involved in the embodiment of the present application is shown in FIG. In the embodiment of the present application, the lithium niobate coherent modulator control device based on the gradient algorithm may include a processor, a memory, a communication interface and a communication bus.

[0137] The communication bus may be of any type and is used to interconnect the processor, the memory, and the communication interface.

[0138] The communication interface includes an input / output (I / O) interface, a physical interface, and a logical interface, etc., which are used to realize the interconnection of devices inside the lithium niobate coherent modulator control device based on the gradient algorithm, and an interface for realizing the interconnection between the lithium niobate coherent modulator control device based on the gradient algorithm and other devices (such as other computing devices or user devices). The physical interface can be an Ethernet interface, an optical fiber interface, an ATM interface, etc.; the user device can be a display (Display), a keyboard (Keyboard), etc.

[0139] The memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0140] The processor may be a general-purpose processor, which may call the lithium niobate coherent modulator control program based on the gradient algorithm stored in the memory, and execute the lithium niobate coherent modulator control method based on the gradient algorithm provided in the embodiment of the present application. For example, the general-purpose processor may be a central processing unit (CPU). The method executed when the lithium niobate coherent modulator control program based on the gradient algorithm is called may refer to the various embodiments of the lithium niobate coherent modulator control method based on the gradient algorithm of the present application, which will not be repeated here.

[0141] Those skilled in the art will understand that Figure 4 The hardware structure shown in the figure does not constitute a limitation on the present application, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0142] The terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices. The terms "first", "second" and "third" are used to distinguish different objects, etc., and do not represent a sequence, nor do they limit "first", "second" and "third" to different types.

[0143] In the description of the embodiments of the present application, "exemplary", "for example" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary", "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary", "for example" or "for example" is intended to present related concepts in a specific way.

[0144] In the description of the embodiments of the present application, unless otherwise specified, “ / ” means or, for example, A / B can mean A or B; the “and / or” in the text is merely a description of the association relationship of associated objects, indicating that three relationships may exist, for example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiments of the present application, “multiple” refers to two or more than two.

[0145] In some processes described in the embodiments of the present application, multiple operations or steps that appear in a specific order are included, but it should be understood that these operations or steps may not be executed in the order in which they appear in the embodiments of the present application or in parallel, and the sequence number of the operation is only used to distinguish the different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed in sequence or in parallel, and these operations or steps may be combined.

[0146] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, disk, CD) as described above, and includes a number of instructions for a terminal device to execute the methods described in each embodiment of the present application.

[0147] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A lithium niobate coherent modulator control method based on a gradient algorithm, used to modulate a lithium niobate IQ modulator with an IQ modulation structure in coherent modulation, characterized in that: The lithium niobate coherent modulator control method based on the gradient algorithm includes: Based on the preprocessing algorithm, the phases of the I and Q paths of the lithium niobate IQ modulator are set to be the same; The bias voltage of I, Q and P paths of the lithium niobate IQ modulator is set at the optimal bias point through the bias voltage locking algorithm; Based on the feedback algorithm, the working state of the lithium niobate IQ modulator under the current bias voltage is judged, and error compensation adjustment is performed to make the working state of the lithium niobate IQ modulator optimal.

2. The control method of a lithium niobate coherent modulator based on a gradient algorithm as claimed in claim 1, characterized in that: The method of setting the phases of the I path and the Q path of the lithium niobate IQ modulator to be the same based on the preprocessing algorithm specifically includes: S101: Scan the entire working bias voltage range of the lithium niobate IQ modulator, output the optical power versus bias voltage curve, obtain the basic characteristic data of the lithium niobate IQ modulator, set the learning rate, convergence coefficient, convergence threshold, and the initial bias voltage of each channel, record the initial optical power, and go to S102; S102: Calculate the gradient of the optical power of the I-path, Q-path, and P-path compared to the initial bias voltage, and move the bias voltages of the I-path, Q-path, and P-path in the direction of the maximum rising gradient of the optical power according to the set moving step length, obtain the bias voltages of the I-path, Q-path, and P-path after the movement, and record the optical power after the movement, and go to S103; S103: Compare the current optical power after the move with the latest initial optical power. If the optical power after the move is greater than the initial optical power, go to S105. If the optical power after the move is not greater than the initial optical power, and the latest learning rate is greater than the convergence threshold, go to S104. If the optical power after the move is not greater than the initial optical power, and the latest learning rate is not greater than the convergence threshold, then before the last I-path, Q-path, and P-path bias voltage move, the bias voltages of the I-path, Q-path, and P-path obtained after the move are the bias voltages corresponding to the maximum optical power, indicating that the bias voltages of the I-path, Q-path, and P-path are set to the bias voltages corresponding to the maximum optical power, and the process ends. S104: Divide the current learning rate by the convergence coefficient to obtain a value as the latest learning rate, and go to S105; S105: taking the current optical power value after the shift as the latest initial optical power, and going to S102; Among them, the set moving step size corresponding to the bias voltage of the I path is the latest learning rate multiplied by the gradient of the optical power of the I path compared to the initial bias voltage, the set moving step size corresponding to the bias voltage of the Q path is the latest learning rate multiplied by the gradient of the optical power of the Q path compared to the initial bias voltage, and the set moving step size corresponding to the bias voltage of the P path is the latest learning rate multiplied by the gradient of the optical power of the P path compared to the initial bias voltage.

3. A lithium niobate coherent modulator control method based on gradient algorithm as claimed in claim 2, characterized in that: The bias voltage locking algorithm is used to set the bias voltages of the I, Q and P paths of the lithium niobate IQ modulator at the optimal bias point, specifically including: locking the bias voltages of the I and Q paths based on a gradient descent algorithm, locking the bias voltage of the P path based on a spectrum sweep algorithm, and locking the bias voltages of the I, Q and P paths based on a pilot algorithm.

4. The control method of a lithium niobate coherent modulator based on a gradient algorithm as claimed in claim 3, characterized in that: The locking of the I-path and Q-path bias voltages based on the gradient descent algorithm specifically includes: S201: when the bias voltage corresponding to the maximum optical power is obtained, the currently recorded optical power after movement is used as the first optical power, and the process goes to S202; S202: Calculate the gradient of the optical power of the I and Q paths compared to the initial bias voltage, and move the bias voltages of the I and Q paths in the direction of the maximum decreasing gradient of the optical power according to the preset moving step length, obtain the bias voltages of the I and Q paths after the movement, and record the optical power after the movement, and go to S203; S203: Compare the current optical power after the move with the latest first optical power. If the optical power after the move is less than the first optical power, go to S205. If the optical power after the move is not less than the first optical power and the latest learning rate is greater than the convergence threshold, go to S204. If the optical power after the move is not less than the first optical power and the latest learning rate is not greater than the convergence threshold, the bias voltages of the I and Q paths obtained before the last bias voltage movement of the I, Q and P paths are obtained are the bias voltages corresponding to the minimum optical power, indicating that the bias voltages of the I and Q paths are set to the bias voltages corresponding to the minimum optical power, and the process ends. S204: Divide the current learning rate by the convergence coefficient to obtain a value as the latest learning rate, and go to S205; S205: taking the current optical power value after the movement as the latest first optical power, and going to S202; Among them, the preset moving step size corresponding to the bias voltage of the I path is the latest learning rate multiplied by the gradient of the optical power of the I path compared to the initial bias voltage, and the preset moving step size corresponding to the bias voltage of the Q path is the latest learning rate multiplied by the gradient of the optical power of the Q path compared to the initial bias voltage.

5. The control method of a lithium niobate coherent modulator based on a gradient algorithm as claimed in claim 4, characterized in that: The locking of the P-circuit bias voltage based on the spectrum sweep algorithm specifically includes: Set the bias voltage of the I and Q paths to the bias voltage corresponding to the minimum optical power, and set the bias voltage of the P path to the bias voltage corresponding to the maximum optical power; Sweep the bias voltage of the P path and store the optical power, obtain any pair of adjacent optical power maximum and optical power minimum values ​​in the stored optical power, and calculate the target optical power of the Quad point based on the optical power maximum and optical power minimum values; In the bias voltage interval corresponding to the maximum optical power and the minimum optical power, the bias voltage point closest to the Quad point target optical power is obtained, determined as the Quad point of the P path, and the bias voltage of the P path is locked as the Quad point.

6. The control method of a lithium niobate coherent modulator based on a gradient algorithm as claimed in claim 3, characterized in that: The pilot algorithm is used to lock the bias voltages of the I-channel, Q-channel and P-channel, wherein the sine wave required by the pilot is simulated by the DAC of the bias control circuit, and the sine wave of the specified frequency is simulated by the MCU timer trigger; The simulation of the pilot signal includes: Set up the discrete sine signal array, add pilot duration, DC bias voltage, pilot sine amplitude, and set the counting threshold according to the pilot sine frequency; Start the MCU timer interrupt, start the timing from 0, and set the MCU timer count to 0. The internal clock of the MCU timer triggers the count to increase, and an interrupt is generated when the count reaches the count threshold; Perform interrupt processing and generate pilot bias voltage, specifically: VH=VDC+Sin[i]*A Wherein, VH represents the pilot bias voltage, VDC represents the DC bias voltage, A represents the pilot sine amplitude, Sin[i] represents the discrete sine signal array generated in real time, and when the length i of the discrete sine signal array generated in real time exceeds the set length of the discrete sine signal array, i is reset to 0, and when i exceeds the length of the discrete sine signal array, i is reset to 0, and when the timing reaches the pilot duration, the MCU timer interrupt is turned off, and the simulation of generating the pilot signal ends.

7. A gradient algorithm-based lithium niobate coherent modulator control method as claimed in claim 6, characterized in that: The locking of the bias voltages of the I-path, Q-path, and P-path based on the pilot algorithm specifically includes: S221: Load frequency f on path I I After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I The normalized optical power at , goes to S222; S222: Determine frequency f I Is the normalized optical power at less than the set threshold? If so, go to S223; if not, adjust the bias voltage of the I path and go to S222; S223: Load frequency f on Q path Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. Q The normalized optical power at , goes to S224; S224: Determine frequency f Q Whether the normalized optical power at is less than the set threshold, if so, go to S225, if not, adjust the bias voltage of the Q path, and go to S223; S225: Load frequency f on I path I A sinusoidal pilot signal with a frequency of f is loaded on the Q path. Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I +f Q The normalized optical power and |f I -f Q |, go to S226; S226: frequency f I +f Q The normalized optical power and |f I -f Q | whether the normalized optical power at the locations is less than the set threshold value; if not, adjust the bias voltage of the P path and go to S225; if so, go to S227; S227: Execute judgment: If the frequency loaded on path I is f I After the sinusoidal pilot signal is sent, the bias voltage of the I path is not adjusted, and the frequency f is loaded on the Q path. Q After the sinusoidal pilot signal is sent, the bias voltage of the Q circuit is not adjusted, and at the same time, the frequency f is loaded on the I circuit. I The sinusoidal pilot signal and the frequency f loaded on the Q path Q If the bias voltage of the P circuit is not adjusted after the sinusoidal pilot signal, the process ends; otherwise, the process goes to S221.

8. The control method of a lithium niobate coherent modulator based on a gradient algorithm as claimed in claim 7, characterized in that: The method of judging the working state of the lithium niobate IQ modulator under the current bias voltage based on the feedback algorithm and performing error compensation adjustment to optimize the working state of the lithium niobate IQ modulator specifically includes: S301: Setting the optical power change threshold and the normalized optical power threshold, and setting the bias voltages of the I, Q, and P paths at the optimal bias points to obtain the optical power corresponding thereto as the optimal optical power, and then proceeding to S302; S302: Read the current optical power regularly, calculate the change between the current optical power and the optimal optical power, and go to S303; S303: Determine whether the calculated change exceeds the optical power change threshold, if so, go to S304, if not, go to S302; S304: Loading frequency f on path I I After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I The normalized optical power at the Q path is loaded with a frequency of f Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. Q The normalized optical power at the location, and the frequency f is loaded on the I path I The sinusoidal pilot signal and the frequency f loaded on the Q path Q After collecting the optical power for a set period of time, stop loading the sinusoidal pilot signal, perform FFT transformation on the optical power, and record the frequency f on the spectrum. I +f Q The normalized optical power and |f I -f Q The normalized optical power at |, according to the recorded frequency f I Normalized optical power at frequency f Q Normalized optical power at frequency f I +f Q Normalized optical power at frequency |f I -f Q | Whether the normalized optical power at exceeds the normalized optical power threshold to determine whether the working point is offset, if so, go to S305, if not, go to S306; S305: Lock the bias voltages of the I-channel, Q-channel, and P-channel based on the pilot algorithm, perform offset correction, and then go to S306; S306: based on the change of the optimal optical power caused by the re-locking of the working point or the change of the overall optical power, update the optimal optical power and go to S302.

9. A lithium niobate coherent modulator control device based on a gradient algorithm, characterized in that: The lithium niobate coherent modulator control device based on the gradient algorithm comprises: A first setting module, which is used to set the phases of the I path and the Q path of the lithium niobate IQ modulator to be the same based on a preprocessing algorithm; A second setting module, which is used to set the bias voltages of the I-path, Q-path, and P-path of the lithium niobate IQ modulator at an optimal bias point through a bias voltage locking algorithm; The execution module is used to judge the working state of the lithium niobate IQ modulator under the current bias voltage based on the feedback algorithm, and perform error compensation adjustment to make the working state of the lithium niobate IQ modulator optimal.

10. A lithium niobate coherent modulator control device based on a gradient algorithm, characterized in that: The lithium niobate coherent modulator control device based on the gradient algorithm includes a processor, a memory, and a lithium niobate coherent modulator control program based on the gradient algorithm stored in the memory and executable by the processor, wherein the lithium niobate coherent modulator control program based on the gradient algorithm, when executed by the processor, implements the steps of the lithium niobate coherent modulator control method based on the gradient algorithm as described in any one of claims 1 to 8.

Images