A method and apparatus for optical communication signal optimization based on optical neural network

Abstract

The application discloses an optical communication signal optimization method and device based on an optical neural network, and relates to the technical field of optical communication.The method is characterized in that an optical neural network chip is designed to optimize optical transmission signals of an optical communication system on an optical path; the weights of a trained neural network are loaded into the optical neural network on an electrical path, the weights are continuously optimized through on-chip online training, and therefore the quality of optical signal optimization is improved.After training is completed, corresponding micro-ring modulators can be designed according to the weights of various devices to realize all-optical processing, and photoelectric conversion is not required, thereby reducing noise caused by photoelectric devices.The method processes optical communication signals through an optical neural network, avoids dependence on electronic computing resources, and improves the response speed and extinction ratio of an optical communication system.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of optical communication technology, in particular to a method and device for optimizing optical communication signals based on an optical neural network. BACKGROUND

[0002] With the rapid development of information technology, the demand for high-speed, large-capacity, long-distance data transmission is increasing. Traditional electrical communication systems have been unable to meet the needs of modern communication in terms of transmission rate and distance. Therefore, optical communication technology has become a research hotspot in the field of future communication due to its high bandwidth, long distance, low loss and other advantages.

[0003] In an optical communication system, signal modulation and transmission is one of the core technologies. Return to Zero (RZ), Non-Return to Zero (NRZ), and 4-level Pulse Amplitude Modulation (PAM4) are common digital signal modulation methods. In particular, in high-speed optical communication systems above 100 Gbit / s, PAM4 is widely used due to its high efficiency and high transmission rate.

[0004] However, there are many noise sources in high-speed optical communication systems, such as current noise, thermal noise, and crosstalk noise. In addition, signals are affected by dispersion and nonlinear effects during transmission, resulting in distortion and intersymbol interference. These factors can interfere with signal transmission quality and increase the bit error rate.

[0005] To solve this problem, researchers have been working to optimize signal processing methods, such as developing new pairs of signal equalization techniques to improve the transmission performance of optical communication systems. In the face of the shortcomings of existing high-speed optical communication systems, particularly the increase in bit error rate and the decrease in transmission efficiency, there is an urgent need to develop a new type of signal optimization chip. SUMMARY

[0006] The problem to be solved by the present application is to provide a method and device for optimizing optical communication signals based on an optical neural network, aiming to provide an efficient, low-power and all-optical optical communication signal optimization method that can improve the transmission quality of an optical communication system, reduce the bit error rate, and achieve data transmission reliability while significantly improving the overall performance and efficiency of the system.

[0007] The present application adopts the following technical solution: including receiving optical communication data, initializing the weight of the micro-ring modulator voltage signal, calculating the optical neural network gradient, updating the weight of the micro-ring modulator voltage signal, and outputting the optical communication data, the specific steps are as follows:

[0008] S1, receiving optical communication data:

[0009] S1.1, the optical communication system transmits various optical communication data formats, and a photodetector detects the light intensity of different optical communication data formats to construct an eye diagram and a data set I of the transmitted data;

[0010] S1.2, the optical signal input module of the optical chip is connected to the receiving end of the optical communication system through an optical fiber;

[0011] S2, initialize the weight of the micro-ring modulator voltage signal:

[0012] S2.1, construct an electrical neural network model;

[0013] S2.2, load the electrical neural network weight into the optical neural network, train the electrical neural network, update the weight constantly, determine the optimal weight, and convert the optimal weight of the electrical neural network into a voltage signal for driving the micro-ring modulator;

[0014] S3, calculate the gradient of the optical neural network:

[0015] S3.1, forward propagation, complete the input and output of the optical communication signal in the optical neural network;

[0016] S3.2, calculate the loss, calculate the loss between the output light intensity in the optical neural network and the ideal output light intensity;

[0017] S3.3, gradient descent, calculate the gradient between each layer in the optical neural network;

[0018] S4, update the weight of the micro-ring modulator voltage signal:

[0019] In the training process of the optical neural network system, if the actual communication signal index is not within the expected target signal index, the error is back propagated, and the weight parameter of the optical neural network is updated;

[0020] S5, output optical communication data:

[0021] The optical signal output module of the optical chip is connected out through an optical fiber and directly connected to the receiving end of the optical communication system.

[0022] Preferably, step S2.1, construct an electrical neural network model:

[0023] Define the input of the neural network model as X, the original RZ, NRZ, and PAM4 data format signals transmitted in the optical communication system as the ideal output Y' of the neural network model, and the actual output as Y. Construct an electrical neural network model f, then:

[0024] Y = f(X; w0)

[0025] In the formula, w0 is the initial weight of the electrical neural network, the loss of the network is the mean square error of the ideal output Y' and the actual output Y, and the size of the electrical neural network model f can be expanded or reduced according to the mode of the input signal.

[0026] Preferably, in step S2.2, the electrical neural network weight is loaded into the optical neural network.

[0027] The electrical neural network is trained, and the weight is constantly updated to determine the final weight w. The neural network weight is converted into a voltage signal v w driving the micro-ring modulator, and the steps are as follows:

[0028] Let the maximum output voltage of the N micro-ring modulators be V max , and the minimum output voltage be V min . Normalize the maximum weight of the electrical neural network to map its value to 0 to 1 to obtain w', wherein w ′ = w1, w2, …, w N . The weight corresponding to the N voltages can be obtained by the following formula:

[0029] v w = K w' (V max -V min )

[0030] In the formula, K = k1, k2, …, k N is the power coefficient of each micro-ring modulator, and v w contains a series of voltage values, i.e., tunable voltages V1, V2 to V N . Loading the obtained N voltages into the corresponding micro-ring modulators can realize the weight initialization of the micro-ring modulator voltage signal.

[0031] Preferably, in step S3.1, forward propagation is performed.

[0032] After initializing the weight of the N micro-ring modulator voltage signal, the receiver of the optical communication system sends RZ, NRZ, and PAM4 data format information to the optical neural network, and the power is recorded as P in . The data is transmitted into the neural network, and the final output produces a prediction signal, i.e., the optimized RZ, NRZ, and PAM4 data format information, and the power is recorded as P out . The original RZ, NRZ, and PAM4 data format information, and the power is recorded as P label . Due to the inherent error of the optical neural network system, the effect is worse than the result of the electrical neural network.

[0033] Preferably, in step S3.2, the loss is calculated.

[0034] The output of the optical neural network is P out , and the label is P label . Whether the output power is in the same order of magnitude is compared, and the label needs to be scaled to the same order of magnitude. Then the loss of the output P out and the label P label is calculated, and the loss function adopts mean square error:

[0035] L P = MSE P = (P out -P label ) 2

[0036] Preferably, step S3.3, gradient descent:

[0037] The calculation formula of the output layer gradient is:

[0038]

[0039] The output of the hidden layer is P Z = v w ·P in , and the calculation formula of the gradient of the hidden layer output to the weight is:

[0040]

[0041] The gradient of the intermediate neuron is:

[0042] ΔP Z = ΔP out *P in

[0043] The calculation formula of the hidden layer weight gradient is:

[0044] Δv w = ΔP Z *P in

[0045] According to the above formula, the gradient values of each part in the optical neural network can be obtained.

[0046] Preferably, step S4, updating the weight of the micro-ring modulator voltage signal:

[0047] In the training process of the optical neural network system, if the actual communication signal index is not in the expected target signal index, the error is back propagated. That is, the output error is transmitted back to the input layer through the hidden layer layer by layer, and the error is allocated to all micro-ring modulator voltage voltage controllers in each layer, and the error signal obtained from each layer is used as the basis for adjusting the weight of each unit.

[0048] The weight update formula of the optical neural network system is:

[0049]

[0050] wherein, θ is the learning rate of the optical neural network training.

[0051] The technical scheme of the present application also provides an optical communication signal optimization device based on an optical neural network, which is used to implement the optical communication signal optimization method described in any of the above, and comprises an optical signal input module, an optical neural network module, a first voltage source module, a second voltage source module and an optical signal output module.

[0052] The optical signal input module is used to output an optical signal in an optical communication system to the optical neural network module.

[0053] The optical neural network module comprises a plurality of micro-ring modulators, and is used to encode and decode the input optical signal, the encoding and decoding weights are controlled by the micro-ring modulators, the first voltage source module and the second voltage source module jointly drive the micro-ring modulators to work, adjust the weights of the optical neural network, and output the encoded optical signal.

[0054] The first voltage source module outputs a plurality of stabilized power sources, and is used to control the output optical power of the upper half of the micro-ring modulators in the optical neural network module.

[0055] The second voltage source module outputs a plurality of stabilized power sources, and is used to control the output optical power of the lower half of the micro-ring modulators in the optical neural network module.

[0056] The optical signal output module is used to receive the output optical signal of the optical neural network module and output the optical signal through an optical fiber.

[0057] Preferably, the modulation voltage range of the micro-ring modulator is 0V to 3V, and the resonance wavelength modulation range is 0nm to 2nm.

[0058] Compared with the prior art, the technical scheme of the present application has the following technical effects:

[0059] 1. The optical communication signal optimization method of the present application adopts an optical neural network chip and a matching algorithm, and focuses on optimizing the RZ, NRZ and PAM4 data format signals in optical communication transmission, thereby improving the extinction ratio of the transmission signal and reducing the bit error rate.

[0060] 2、The optical communication signal optimization method adopts an online training algorithm to realize training on an optical neural network chip, optimizes actual communication signals in an end-to-end mode, can further reduce errors of directly loading electrical neural network weights into the optical chip, improves the accuracy of the system, and the system is suitable for various complex electromagnetic environments, and can realize optimization of optical communication signals through deployment of the chip under strong signals within a line-of-sight range or weak signals at a super line-of-sight long distance, and has good practicability. BRIEF DESCRIPTION OF DRAWINGS

[0061] Figure 1 It is a schematic diagram of the optical neural network chip for optimizing optical communication signals of the application;

[0062] Figure 2 It is a flowchart of the method for optimizing optical communication signals of the application;

[0063] Figure 3 It is an online training schematic diagram of the optical neural network chip. DETAILED DESCRIPTION

[0064] In order to make the objects, technical solutions and advantages of the application clearer, the technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0065] The purpose of the application is to provide a method and device for optimizing optical communication signals based on an optical neural network, which can realize optical signal optimization of light to light at a receiving end of an optical communication system, improve the stability of the optical communication system, and reduce the bit error rate.

[0066] In one embodiment of the application, the device for optimizing optical communication signals based on an optical neural network, as shown in Figure 1 includes an optical signal input module, an optical neural network module, a first voltage source module, a second voltage source module and an optical signal output module.

[0067] The optical signal input module is used for outputting optical signals in an optical communication system to the optical neural network module.

[0068] The optical neural network module includes a plurality of microring modulators, is used for encoding and decoding input optical signals, the encoding and decoding weights are controlled by internal microring modulators, external voltage drives the microring modulators to work, and outputs the encoded optical signals.

[0069] The first voltage source module outputs a plurality of stabilized power sources for controlling the output light intensity of the upper half of the micro-ring modulators in the optical neural network module, i.e., the weights of the optical neural network.

[0070] The second voltage source module outputs a plurality of stabilized power sources for controlling the output light intensity of the lower half of the micro-ring modulators in the optical neural network module, i.e., the weights of the optical neural network.

[0071] The optical signal output module is used for receiving the optical signal of the optical neural network module and outputting through an optical fiber.

[0072] In this embodiment, the micro-ring modulator has 10, and the initial resonant wavelengths are λ1, λ2,..., λ 10 The first voltage source module outputs 5 stabilized power sources for controlling the micro-ring modulators with initial resonant wavelengths λ1, λ2,..., λ5, and the second voltage source module outputs 5 stabilized power sources for controlling the micro-ring modulators with initial resonant wavelengths λ6, λ7,..., λ 10 The input optical signal is input from the side of the micro-ring modulator with initial resonant wavelength λ1, and the optical signal is output from the side of the micro-ring modulator with initial resonant wavelength λ 10 .

[0073] Based on the above-mentioned optical communication signal optimization device based on an optical neural network, the corresponding optimization method of this embodiment includes receiving an optical communication data part, initializing a weight part of a micro-ring modulator voltage signal, calculating an optical neural network gradient part, updating the weight part of the micro-ring modulator voltage signal, and outputting an optical communication data part, as shown in the flowchart. Figure 2

[0074] The receiving optical communication data part includes:

[0075] (1) The optical communication system sends an optical communication data format, and the photodetector detects the light intensity of different data formats to construct an eye diagram and a data set I of the transmitted data.

[0076] (2) The optical signal input module of the optical chip is connected to the receiving end of the optical communication system through an optical fiber.

[0077] Specifically, in the process of constructing the data set I, it contains three types of data formats, RZ, NRZ and PAM4, and the input of each type of data is the light intensity received by the photodetector after passing through the optical communication system. The label of each type of data is the light intensity directly received by the photodetector from the optical communication system.

[0078] In this embodiment, the number of samples of each type of data is 10,000, and the number of test set samples is 2,000.

[0079] The initialization of the weight part of the micro-ring modulator voltage signal includes:​

[0080] (1) Constructing the electrical neural network model:

[0081] Define the neural network model input X, the original RZ, NRZ and PAM4 data format signals sent in the optical communication system as the ideal output Y' of the neural network model, and the actual output as Y. Construct the electrical neural network model f, then:

[0082] Y = f(X; w0)

[0083] In the formula, w0 is the initialization weight of the electrical neural network, and the loss of the network is the mean square error of the ideal output Y' and the actual output Y. The size of the electrical neural network model f can be expanded and reduced according to the mode of the input signal.

[0084] (2) Load the electrical neural network weight into the optical neural network:

[0085] Train the electrical neural network, and update the weight constantly. Determine the final weight as w. Convert the neural network weight into the voltage signal v w that drives the micro-ring modulator, and the steps are as follows:

[0086] In this embodiment, it is assumed that the maximum output of the 10 micro-ring modulators in the system corresponds to the voltage V max , and the minimum output corresponds to the voltage V min . Normalize the maximum weight of the electrical neural network, and map its value to 0 to 1 to obtain w', wherein w ′ = w1, w2, …, w 10 . The weight corresponding to the 10 voltages can be obtained by converting the following formula:

[0087] v w = K w' (V max -V min )

[0088] In the formula, K = k1, k2, …, k 10 is the power coefficient of each micro-ring modulator, and v w contains a series of voltage values, that is, the tunable voltages V1, V2 to V 10 .

[0089] Load the obtained 10 voltages into the corresponding micro-ring modulators, and the weight initialization of the micro-ring modulator voltage signal can be realized.

[0090] The calculation of the optical neural network gradient part includes:

[0091] (1) Forward propagation:

[0092] After initializing the weights of the 10 microring modulator voltage signals, the receiver of the optical communication system sends the RZ, NRZ, and PAM4 data format information to the optical neural network, and the power is recorded as P in The data is transmitted into the neural network, and the final output produces a predicted signal, i.e. the optimized RZ, NRZ, and PAM4 data format information, and the power is recorded as P out The original RZ, NRZ, and PAM4 data format information, and the power is recorded as P label Due to the inherent errors of the optical neural network system, the effect is worse than the results of the electrical neural network.

[0093] (2) Calculate the loss:

[0094] The output of the optical neural network is P out The label is P label Compare the power of the two outputs to see if they are in the same order of magnitude. The label needs to be scaled to the same order of magnitude. Then calculate the loss of the output P out and the label P label The loss function uses mean square error:

[0095] L P = MSE P = (P out -P label ) 2

[0096] (3) Gradient descent:

[0097] The calculation formula of the output layer gradient is:

[0098]

[0099] The output of the hidden layer is P Z = v w ·P in Then the calculation formula of the gradient of the hidden layer output to the weight is:

[0100]

[0101] The gradient of the intermediate neuron is:

[0102] ΔP Z = ΔP out *P in

[0103] The calculation formula of the hidden layer weight gradient is:

[0104] Δv w = ΔP Z *P in

[0105] According to the above formula, the gradient values of each part in the optical neural network can be obtained.

[0106] The weight part of the micro-ring modulator voltage signal is updated, including:

[0107] During the training process of the optical neural network system, if the actual communication signal indicator is not within the expected target signal indicator, the error is back-propagated. That is, the output error is transmitted back to the input layer through the hidden layer layer by layer, and the error is allocated to the voltage controllers of all micro-ring modulators in each layer, so that the error signals obtained from each layer are used as the basis for adjusting the weights of each unit.

[0108] The weight update formula of the optical neural network system is:

[0109]

[0110] wherein, theta is the learning rate of the optical neural network training. Then, the updated micro-ring modulator voltage signal is loaded to the corresponding voltage controller.

[0111] The output optical communication data part includes:

[0112] The optical signal output module of the optical chip is connected out through an optical fiber, and can be directly connected to the receiving end of the optical communication system.

[0113] The complete optical neural network chip online training scheme is as shown in Figure 3 .

[0114] In summary, the present application avoids the dependence on electronic computing resources through optical-to-optical computing, improves the anti-interference ability of the optical communication system, and improves the stability of the optical communication system through optimization of the communication light. The online training is realized in the deployment of the optical neural network, and the bit error rate is reduced.

[0115] Each embodiment in the specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between each embodiment can be referred to each other.

[0116] The principles and implementation modes of the present application are described by using specific examples in this paper. The above description of the embodiments is only used to help understand the method and core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range can be changed. In summary, the content of the specification should not be understood as a limitation of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be regarded as the protection scope of the present application.

Claims

1. A method of optical communication signal optimization based on optical neural networks, characterized in that, Comprise: Receiving optical communication data, initializing the weight of the micro-ring modulator voltage signal, calculating the optical neural network gradient, updating the weight of the micro-ring modulator voltage signal and outputting the optical communication data, the specific steps are as follows: S1, receiving optical communication data: S1.1, the optical communication system sends various optical communication data formats, and the photodetector detects the light intensity of different optical communication data formats to construct an eye diagram and a data set I of the transmitted data; S1.2, the optical signal input module of the optical chip is connected to the receiving end of the optical communication system through an optical fiber; S2, initializing the weight of the micro-ring modulator voltage signal: S2.1, constructing an electrical neural network model, specifically including: S2.1.1, defining the input of the neural network model as X and the actual output as Y, and the original RZ, NRZ and PAM4 data format signal sent in the optical communication system as the ideal output Y' of the neural network model; S2.1.2, constructing an electrical neural network model f, which has the following formula: Y=f(X;w0) Where w0 is the initialization weight of the electrical neural network, and the loss of the network is the mean square error of the ideal output Y' and the actual output Y; S2.1.3, the size of the electrical neural network model f is expanded and reduced according to the mode of the input signal; S2.2, load the electrical neural network weight into the optical neural network, train the electrical neural network, update the weight, determine the optimal weight, and convert the optimal weight of the electrical neural network into the voltage signal driving the micro-ring modulator; The optical neural network comprises a plurality of micro-ring modulators for encoding and decoding input optical signals, and the encoding and decoding weight is controlled by the micro-ring modulator. The first voltage source module and the second voltage source module jointly drive the micro-ring modulator to work, adjust the weight of the optical neural network, and output the encoded optical signal; the first voltage source module outputs a plurality of stabilized power supplies for controlling the output optical power of the upper half of the micro-ring modulator in the optical neural network module; the second voltage source module outputs a plurality of stabilized power supplies for controlling the output optical intensity of the lower half of the micro-ring modulator in the optical neural network module; S3, calculating the gradient of the optical neural network: S3.1, forward propagation, completing the input and output of the optical communication signal in the optical neural network; S3.2, loss calculation, calculating the loss between the output optical intensity and the ideal output optical intensity in the optical neural network; S3.3, gradient descent, calculating the gradient between each layer in the optical neural network; S4, updating the weight of the micro-ring modulator voltage signal: The error of the output voltage is transmitted back to the input layer through the hidden layer, and the error is distributed to the voltage controllers of all micro-ring modulator voltages in each layer. The error signal obtained from each layer is used as the basis for adjusting the weight of the micro-ring modulator voltage signal of the optical neural network; In the training process of the optical neural network system, if the actual communication signal index is not within the expected target signal index, the error is back propagated, and the weight parameters of the optical neural network are updated, and the formula is: wherein θ is the learning rate of the optical neural network training, and the updated microring modulator voltage signal v w is loaded onto the corresponding voltage controller, L P is the output power P out is the loss of the original power P label ; S5, outputting optical communication data: The optical signal output module of the optical chip is connected out through an optical fiber and directly connected to the receiving end of the optical communication system.

2. The method of optical neural network based optimization of optical communication signals according to claim 1, wherein, In step S1.1, the data set I includes three types of data formats of RZ, NRZ and PAM4, the input of each type of data is the optical intensity received by the photodetector after the optical communication system, and the label of each type of data is the optical intensity directly received by the photodetector from the optical communication system; a test set is constructed, and the number of test set samples is not less than 20% of the total samples of the data set I.

3. The method of optical neural network based optimization of optical communication signals according to claim 1, wherein, In step S2.2, the electrical neural network weight is loaded into the optical neural network, specifically including: S2.2.1, the optical neural network comprises N microring modulators, the voltage corresponding to the maximum output of the microring modulators is V max , and the voltage corresponding to the minimum output is V min , the maximum output of the optical neural network is normalized to map the value of the weight w to 0 to 1, to obtain w', w ′ = w1, w2,..., w N , wherein w i represents the output voltage weight of the i-th microring modulator, i = 1, 2,..., N; S2.2.2, the minimum weight of the output voltage of the N micro-ring modulators is calculated to convert the driving voltage signal of the micro-ring modulator, and the conversion formula is as follows: v w = Kw'(V max -V min ) where K = k1, k2,..., k N , k i is the power coefficient of the ith microring modulator, v w is the driving voltage signal, v w contains N tunable voltage values; S2.2.3, the obtained N tunable voltage values are loaded into the corresponding micro-ring modulator to realize the weight initialization of the N micro-ring modulator voltage signal.

4. The method of optical neural network based optimization of optical communication signals according to claim 3, wherein, In step S3.1, the forward propagation, the weights of the N microring modulator voltage signals are initialized, the receiver of the optical communication system sends the RZ, NRZ and PAM4 data format information to the optical neural network, the input power is P in ; the data is input into the optical neural network to output the generated prediction signal, and the optimized RZ, NRZ and PAM4 data format information is obtained, and the output power is P out ; the original RZ, NRZ and PAM4 data format information, and the original power is P label .

5. The method of optimizing optical communication signals based on optical neural networks according to claim 4, wherein, In step S3.2, the loss is calculated, the output power P out with the original power P label loss L P The loss function adopts mean square error, which is expressed as follows: L P = MSE P = (P out - P label ) 2 Wherein, MSE is the mean square error function of power.

6. The method of optimizing optical communication signals based on optical neural networks according to claim 5, wherein, In step S3.3, the gradient descent includes the following sub-steps: S3.3.1, calculating the output layer gradient ΔP out : Wherein, d is the gradient calculation function; S3.3.2, compute gradient P of hidden layer output with respect to weights in : where P Z is the output of the hidden layer, P Z = v w · P in ; S3.3.3, computing the gradient ΔP of the intermediate neuron Z : ΔP Z = ΔP out *P in S3.3.4, calculating the hidden layer weight gradient Δv w : Δv w = ΔP Z *P in .

7. An apparatus for optical communication signal optimization based on optical neural networks for implementing the method of any of claims 1-6, characterized by Including: The optical signal input module, the optical neural network module, the first voltage source module, the second voltage source module and the optical signal output module; The optical signal input module is used for outputting the optical signal in the optical communication system to the optical neural network module; The optical neural network module includes a plurality of micro-ring modulators, which are used for encoding and decoding the input optical signal, and the encoding and decoding weight is controlled by the micro-ring modulator, and the micro-ring modulator is driven by the first voltage source module and the second voltage source module to work, adjust the weight of the optical neural network, and output the encoded optical signal; The first voltage source module outputs a plurality of stabilized power supplies for controlling the output optical power of the upper half micro-ring modulator in the optical neural network module; The second voltage source module outputs a plurality of stabilized power supplies for controlling the output optical intensity of the lower half micro-ring modulator in the optical neural network module; The optical signal output module is used for receiving the output optical signal of the optical neural network module and outputting through an optical fiber.

8. The apparatus for optical neural network based optimization of optical communication signals according to claim 7, wherein, The modulation voltage range of the micro-ring modulator is 0V to 3V, and the resonance wavelength modulation range is 0nm to 2nm.

Images