CN117675009A - Dispersion compensation method based on reserve pool calculation - Google Patents
Dispersion compensation method based on reserve pool calculation Download PDFInfo
- Publication number
- CN117675009A CN117675009A CN202311435573.2A CN202311435573A CN117675009A CN 117675009 A CN117675009 A CN 117675009A CN 202311435573 A CN202311435573 A CN 202311435573A CN 117675009 A CN117675009 A CN 117675009A
- Authority
- CN
- China
- Prior art keywords
- signal
- reservoir
- optical
- reserve tank
- weight matrix
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 239000006185 dispersion Substances 0.000 title claims abstract description 33
- 238000004364 calculation method Methods 0.000 title claims abstract description 10
- 230000003287 optical effect Effects 0.000 claims abstract description 36
- 210000002569 neuron Anatomy 0.000 claims abstract description 24
- 239000013307 optical fiber Substances 0.000 claims abstract description 7
- 238000005070 sampling Methods 0.000 claims abstract description 5
- 238000001914 filtration Methods 0.000 claims abstract description 4
- 238000007493 shaping process Methods 0.000 claims abstract description 4
- 239000011159 matrix material Substances 0.000 claims description 31
- 238000012549 training Methods 0.000 claims description 17
- 230000008569 process Effects 0.000 claims description 10
- 238000009826 distribution Methods 0.000 claims description 9
- 230000003595 spectral effect Effects 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 7
- 238000013528 artificial neural network Methods 0.000 claims description 6
- 238000009827 uniform distribution Methods 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 3
- 230000009022 nonlinear effect Effects 0.000 claims description 3
- 238000004891 communication Methods 0.000 abstract description 2
- 230000006403 short-term memory Effects 0.000 abstract description 2
- 239000000835 fiber Substances 0.000 description 13
- 230000005540 biological transmission Effects 0.000 description 11
- 230000006870 function Effects 0.000 description 10
- 230000035945 sensitivity Effects 0.000 description 6
- 230000008901 benefit Effects 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 4
- 238000012417 linear regression Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 238000001228 spectrum Methods 0.000 description 3
- 125000004122 cyclic group Chemical group 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000006735 deficit Effects 0.000 description 1
- 238000005562 fading Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Landscapes
- Optical Communication System (AREA)
Abstract
The invention relates to a dispersion compensation method based on reserve pool calculation, which belongs to the technical field of optical communication and comprises the following steps: generating and sampling signals, and carrying out pulse shaping on the sampled signals through a root raised cosine filter; modulating the signal by an ideal intensity optical modulator, converting the electrical signal into an optical signal, and injecting the optical signal into an optical channel consisting of a standard single-mode optical fiber; the method comprises the steps of detecting the square rate of a photoelectric detector, converting an optical signal into an electric signal at a receiving end, converting the electric signal into a digital domain through an analog-to-digital converter, and carrying out matched filtering through a root raised cosine filter; and compensating the chromatic dispersion by using a trained reserve pool to obtain an output signal, and performing hard decision on the processed signal. The invention enables the reserve tank to have short-term memory capacity through random connection among neurons of the hidden layer, and the interior of the reserve tank has very rich dynamics characteristics, and the two characteristics of the reserve tank are utilized to compensate dispersion, so that the original signal is recovered.
Description
Technical Field
The invention relates to the technical field of optical communication, in particular to a dispersion compensation method based on reservoir calculation.
Background
Intensity Modulation Direct Detection (IMDD) has the advantages of simple structure and low power consumption in short-distance optical transmission, however, the serious limitation of the IMDD system is chromatic dispersion, because phase information is lost after direct detection, and in addition, the radio frequency power fading effect in the IMDD system can be caused by large signal bandwidth and long optical fiber transmission length; for IMDD transmission, the dispersion compensation technique mainly includes optical domain and electrical domain techniques, and the conventional optical technique requires optical devices such as a dispersion compensation optical fiber, a chirped and apodized fiber bragg grating, a ring resonator, and the like; optical domain dispersion compensation is not widely used in short-range optical transmission because it is either costly or is greatly affected by nonlinear effects caused by ambient temperature and optical power. The advantage of the electrical domain dispersion compensation technique over optical techniques is that it does not require changes in the structure of the transmitter and receiver, nor introduces optical losses. The most common equalizers in the electrical domain are Feed Forward Equalizer (FFE) and Decision Feedback Equalizer (DFE). FFE is a linear equalizer, but its compensation performance is poor. In a direct detection system, linear optical distortions can become nonlinear impairments in the electrical domain due to square-law detection, and linear FFE cannot compensate for these nonlinear distortions. The DFE is to add a prediction filter based on FFE to reduce the interference variance at the equalizer output, thereby improving the bit error rate performance. However, a major problem faced by DFEs is that they also suffer from error propagation when false equalization decisions are passed to the feedback process.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, provides a dispersion compensation method based on reservoir calculation, and solves the defects in the prior art.
The aim of the invention is achieved by the following technical scheme: a dispersion compensation method based on reservoir calculation, the dispersion compensation method comprising:
generating an OOK signal, sampling each symbol of the OOK signal, and pulse shaping the sampled OOK signal through a root raised cosine filter;
modulating the signal by an ideal intensity optical modulator, converting the electrical signal into an optical signal, and injecting the optical signal into an optical channel consisting of a standard single-mode optical fiber;
the method comprises the steps of detecting the square rate of a photoelectric detector, converting an optical signal into an electric signal at a receiving end, converting the electric signal into a digital domain through an analog-to-digital converter ADC, and performing matched filtering through a root raised cosine filter;
and compensating chromatic dispersion by using the trained reserve pool to obtain an output signal, and performing hard decision on the processed signal to obtain an OOK signal.
The reserve pool is based on a cyclic neural network, and a large-scale sparse random connection network is used for replacing a hidden layer of the traditional cyclic neural network. The training process of the algorithm is greatly simplified through the partial weight of the training network, and the defects that the traditional recurrent neural network structure is difficult to determine and the training process is too complex are overcome. Since the reserve tank has a neural network architecture with feedback connection, it can learn and memorize information observed in the past by forming an information circulation structure, so that it can compensate nonlinear effects, namely, dispersion effects in the IMDD system;
the reserve tank comprises an input layer, a reservoir layer and an output layer; obtaining weight matrix W connecting input layer and reservoir from uniform distribution U (-1, 1) in The dimension is (N+b) x M; obtaining probability of neuron interconnection in reservoir from binary distribution, and setting weight matrix W of connected neurons in reservoir according to standard normal distribution N (0, 1) res The dimension is MxM, and the state equation of the obtained reserve pool is x [ n ]]＝α·f(W in ·u[n]+W res ·x[n-1])+(1-α)·x[n-1]Wherein α represents the leak rate, f represents the activation function, u [ n ]]Representing the input signal, N representing the input of the network, b representing the bias component, M representing the number of neurons in the reservoir;
by usingRepresenting a linkThe weight matrix connected with the reservoir and the output is M multiplied by L in dimension>Representing a weight matrix connecting the input and output layers, the dimension being (n+b) x L, by changing the state x N of the reservoir]And input signal u [ n ]]Obtain the output signal->Where L represents the output of the network.
Training the reservoir by an equalization process comprising equalizing the number of reservoir neurons, the spectral radius and the leak rate of three key parameters defining the reservoir structure to achieve the optimal effect of the reservoir.
The over-equalization process training the reserve tank includes:
deriving a weight matrix W connecting the input layer and the storage layer from the uniform distribution U (-1, 1) in Generating a binary distributed reservoir weight matrix W according to the sparsity res Replacement of reservoir weight matrix W with normal distribution N (0, 1) res Is to rescale the reservoir weight matrix W according to the size of the spectral radius res ；
According to the input signal u [ n ]]Obtaining the state x [ n ] of the reserve tank by using the state equation of the reserve tank]＝α·f(W in ·u[n]+W res ·x[n-1])+(1-α)·x[n-1]And according to the formulaUsing the state x n of the reserve tank]Input signal u [ n ]]And a target signal y [ n ]]Training a weight matrix connecting the reservoir and the output layer +.>And a weight matrix connecting the input layer and the output layerWherein (1)> b represents the bias term and,lambda is a ridge regularization factor, and I is an identity matrix with the size of M+N+1;
input signal u [ n ]]According to the formula x [ n ]]＝α·f(W in ·u[n]+W res ·x[n-1])+(1-α)·x[n-1]Andcalculating to obtain output signal y n]。
The invention has the following advantages: a dispersion compensation method based on reservoir calculation, which enables the reservoir to have short-term memory capacity through random connection among neurons of a hidden layer, and the reservoir has very rich dynamics, and the dispersion is compensated by utilizing the two characteristics of the reservoir, so that an original signal is recovered.
Drawings
FIG. 1 is a schematic flow chart of the present invention;
FIG. 2 is a graph showing three key parameters as a function of fiber length, wherein (a) shows the change in bit error rate as a function of fiber length for different numbers of neurons in the reservoir; (b) The error rate is shown as a function of fiber length at different spectral radii; (c) The error rate is shown as a function of fiber length at different leak rates;
FIG. 3 is a graph showing the variation of bit error rate of a 32Gbaud OOK signal after SSMF transmission for 10km according to the received optical power;
FIG. 4 shows the error rate of 10 for different equalization techniques -3 A graph of the sensitivity cost of the received light power as a function of the length of the fiber is shown.
Detailed Description
For the purposes of making the objects, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, but not all embodiments. The components of the embodiments of the present application, which are generally described and illustrated in the figures herein, may be arranged and designed in a wide variety of different configurations. Accordingly, the following detailed description of the embodiments of the present application, provided in connection with the accompanying drawings, is not intended to limit the scope of the application, as claimed, but is merely representative of selected embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present application without making any inventive effort, are intended to be within the scope of the present application. The invention is further described below with reference to the accompanying drawings.
As shown in fig. 1, the present invention relates to a dispersion compensation method based on reservoir calculation, which specifically includes the following steps:
step 1, generating an OOK signal with a bandwidth of 32GBaud, and sampling 8 samples for each symbol of the OOK signal.
And step 2, pulse shaping is carried out on the OOK signals after sampling through a root raised cosine filter (SRRC) with the roll-off factor of 0.1.
And 3, modulating the signal by an ideal intensity light modulator, and converting the electric signal into an optical signal.
Step 4, injecting an optical signal into an optical channel consisting of a Standard Single Mode Fiber (SSMF), wherein the dispersion parameter is d=16.4 ps/nm/km, and the SSMF model does not contain nonlinearity due to consideration of short distance.
And 5, at a receiving end, detecting the square rate of a Photoelectric Detector (PD) to convert the optical signal into an electric signal.
Step 6, converting the signal into a digital domain through an analog-to-digital converter ADC, and then performing matched filtering through a root raised cosine filter (SRRC).
And 7, compensating the chromatic dispersion by using a reserve pool calculation method.
And 8, finally, performing hard Decision (DEC) on the processed signal to obtain an OOK signal.
Further, a reserve tankMainly consists of three parts, an input layer, a hidden layer called reservoir and an output layer. Weights (W) connecting the input layer and the reservoir in ) Derived from a uniform distribution U (-1, 1). The probability of neuron interconnection in a reservoir is derived from a binary distribution. Weights between neurons in reservoirs (W res ) Determines the connectivity of the reservoirs, these non-zero weights (W res ) Defined and kept fixed according to a standard normal distribution N (0, 1). Sparsity represents the proportion of non-zero elements in the reservoir matrix to total elements, where sparsity is set to 0.9. Neurons in the reservoir are leaky-integrate neurons, so the state equation of the reservoir is:
x[n]＝α·f(W in ·u[n]+W res ·x[n-1])+(1-α)·x[n-1]
where α is the leak rate, controlling the state of the reservoir at the last moment. f represents an activation function, here a hyperbolic tangent function is chosen. At the output level, the network reading y [ n ] is obtained by linearly changing the state of the reservoir x [ n ] and the input signal u [ n):
the evolution of the state of the pool depends only on the input and training is only used to optimize the output weightsThis optimization does not require complex and computationally expensive back propagation throughout the network as does RNN (time back propagation algorithm). Training may be performed by a single linear regression operation. The Mean Square Error (MSE) is used as the target loss function during the training phase and the least squares method is used to solve this linear regression problem. Computationally, this problem can be further handled as a pseudo-inverse problem.
The number of reservoir neurons, the spectral radius and the leak rate are key parameters in determining the reservoir structure, and therefore these parameters need to be explored to find the best equalization effect. In the simulation process, the principle of a control variable method is adopted, default values 300, 0.8 and 0.9 of the three parameters are respectively selected, and the transmission performance under different optical fiber distances is inspected.
As shown in fig. 2, the transmission variation of BER at different transmission distances is shown for a received optical power of-2 dBm. In order to be able to capture all sample features, the reservoir typically needs a sufficiently large network scale. As can be seen from graph (a) in fig. 2, reservoir performance increases with increasing number of neurons. It will be appreciated that the greater the number of neurons in the reservoir, the more neurons the signal will pass through, and the greater the nonlinear capability of the system. With the increasing number of neurons in the reservoir, the complexity of the network structure increases dramatically, and the phenomenon of "over-fitting" easily occurs, resulting in a decrease in the generalization ability of the network. Therefore, after increasing the number of neurons to 300, the system performance of 400 neurons and 500 neurons was hardly improved.
The spectrum radius is the maximum singular value of the connection weight matrix of the storage layer, and when the spectrum radius is smaller than 1 and larger than 0, the storage pool is guaranteed to have echo state characteristics. The spectral radius determines the storage capacity of the echo state network. Figure 2, panel (b), shows that the stability of the network is best at 0.8. In general, the greater the leak rate, the better the long-term memory capability of the network. As can be seen from fig. 2 (c), as the leak rate increases, the performance of the algorithm increases accordingly. When the leak rate exceeds 0.8, the performance of the reservoir tends to stabilize. Thus, the key parameter was chosen to be 300 neurons with a leak rate of 0.9 and a spectral radius of 0.8.
Further, the equalization process using the reservoir is:
training was performed using 20% of the symbol sequence, and 80% of the symbol sequence was used for testing.
The pool structure is initially set up. Deriving a weight matrix W connecting the input layer and the storage layer from the uniform distribution U (-1, 1) in . Generating a binary distributed reservoir weight matrix W according to the sparsity res Replacement of reservoir weight matrix W with normal distribution N (0, 1) res Is a non-zero element of (c). Finally, re-scaling the reservoir weight matrix W according to the size of the spectrum radius res 。
The pool structure is then trained.
The input signal is u n, and the state equation of the reserve tank is used to obtain the state x n of the reserve tank.
x[n]＝α·f(W in ·u[n]+W res ·x[n-1])+(1-α)·x[n-1]
By using the state x [ n ] of the reserve tank]Input signal u [ n ]]Target signal y [ n ]]Training a weight matrix connecting a reservoir and an output layerAnd a weight matrix connecting between the input layer and the output layer +.>
Wherein the method comprises the steps of b represents a bias term, ">Lambda is a ridge regularization factor; i is an identity matrix of size M+N+1. Training may be performed by a single linear regression operation. This linear regression problem is solved using the Mean Square Error (MSE) as the target loss function and using the least squares method. Computationally, this problem can be further handled as a pseudo-inverse problem.
And finally, predicting the trained reserve pool. The input signal u n is calculated to obtain the output signal y n by the following two equations.
x[n]＝α·f(W in ·u[n]+W res ·x[n-1])+(1-α)·x[n-1]
In order to numerically compare the performance of reservoir dispersion compensation, a conventional feedforward equalizer (FFE) and Decision Feedback Equalizer (DFE) are implemented in the receiver of the IMDD system. The number of taps of the FFE is set to 6, while the number of forward and reverse filter taps of the DFE are set to 4 and 2, respectively, and all of these taps are selected for optimal performance in the system. For both equalizers, the tap coefficients are adjusted using a Least Mean Square (LMS) algorithm. The equalizer is then modeled as a linear time invariant filter. The parameters of the reserve tank are set to the optimal parameters.
As shown in fig. 3, the relationship between the error rate and the received light power of three different equalization schemes at a transmission distance of 10km is shown. The reference system in the black curve represents an IMDD system without any dispersion compensation equalization. It can be seen that the reservoir equalizer is a better improvement over the other two equalization schemes at a fiber length of 10 km. Compared with a reference system, the bit error rate is 1 multiplied by 10- 3 When the power loss of the receiver sensitivity has been reduced to-4 dBm, more than 1dB is added than the other two equalizers.
To study the effect of dispersion compensation over different fiber lengths, it is shown in FIG. 4 that the bit error rate is 1X 10 over fiber lengths ranging from 0km to 38km -3 The cost of optical power sensitivity is received. It can be seen that the reservoir exhibits the lowest sensitivity penalty than the other two when the fiber length is less than 13 km. At fiber lengths of about 15km and 28km, the received optical power sensitivity of the reservoir had peaks indicating poor dispersion compensation in the system. This performance may be due to the limitation of the reservoir to compensate for dispersion at different transmission lengths, as shown by the PSD curves for different fiber lengths in fig. 1. The viable transmission distance of the reservoir can reach 38km in the range of 2dBm for the received optical power sensitivity penalty. When the length of the optical fiber exceeds 10km, the DFE exhibits better compensation performance than the FFE because the power attenuation is difficult to compensate with the linear equalizer.
The foregoing is merely a preferred embodiment of the invention, and it is to be understood that the invention is not limited to the form disclosed herein but is not to be construed as excluding other embodiments, but is capable of numerous other combinations, modifications and adaptations, and of being modified within the scope of the inventive concept described herein, by the foregoing teachings or by the skilled person or knowledge of the relevant art. And that modifications and variations which do not depart from the spirit and scope of the invention are intended to be within the scope of the appended claims.
Claims (4)
1. A dispersion compensation method based on reservoir calculation, characterized by: the dispersion compensation method comprises the following steps:
generating an OOK signal, sampling each symbol of the OOK signal, and pulse shaping the sampled OOK signal through a root raised cosine filter;
modulating the signal by an ideal intensity optical modulator, converting the electrical signal into an optical signal, and injecting the optical signal into an optical channel consisting of a standard single-mode optical fiber;
the method comprises the steps of detecting the square rate of a photoelectric detector, converting an optical signal into an electric signal at a receiving end, converting the electric signal into a digital domain through an analog-to-digital converter ADC, and performing matched filtering through a root raised cosine filter;
and compensating chromatic dispersion by using the trained reserve pool to obtain an output signal, and performing hard decision on the processed signal to obtain an OOK signal.
2. A reservoir-based calculated dispersion compensation method as claimed in claim 1, wherein: the reserve tank replaces a hidden layer of a traditional circulating neural network with a large-scale sparse random connection network, the training process is simplified by training partial weights of the network, the reserve tank is provided with a neural network framework with feedback connection, and information observed in the past is learned and memorized by forming an information circulation structure, so that nonlinear effects, namely dispersion effects in an IMDD system, can be compensated;
the reserve tank comprises an input layer, a reservoir layer and an output layer; deriving weights connecting input and reservoir from uniform distribution U (-1, 1)Matrix W in The dimension is (N+b) x M; obtaining probability of neuron interconnection in reservoir from binary distribution, and setting weight matrix W of connected neurons in reservoir according to standard normal distribution N (0, 1) res The dimension is MxM, and the state equation of the obtained reserve pool is x [ n ]]＝α·f(W in ·u[n]+W res ·x[n-1])+(1-α)·x[n-1]Wherein α represents the leak rate, f represents the activation function, u [ n ]]Representing the input signal, N representing the input of the network, b representing the bias component, M representing the number of neurons in the reservoir;
by usingRepresenting a weight matrix connecting reservoir and output, with dimensions MxL, < >>Representing a weight matrix connecting the input and output layers, the dimension being (n+b) x L, by changing the state x N of the reservoir]And input signal u [ n ]]Obtain the output signal->Where L represents the output of the network.
3. A reservoir-based calculated dispersion compensation method as claimed in claim 2, wherein: training the reservoir by an equalization process comprising equalizing the number of reservoir neurons, the spectral radius and the leak rate of three key parameters defining the reservoir structure to achieve the optimal effect of the reservoir.
4. A reservoir-based calculated dispersion compensation method as claimed in claim 3, wherein: the over-equalization process training the reserve tank includes:
deriving a weight matrix W connecting the input layer and the storage layer from the uniform distribution U (-1, 1) in Generating a binary distributed reservoir weight matrix W according to the sparsity res Replacement of reservoir weight matrix W with normal distribution N (0, 1) res Is to rescale the reservoir weight matrix W according to the size of the spectral radius res ；
According to the input signal u [ n ]]Obtaining the state x [ n ] of the reserve tank by using the state equation of the reserve tank]＝α·f(W in ·u[n]+W res ·x[n-1])+(1-α)·x[n-1]And according to the formulaUsing the state x n of the reserve tank]Input signal u [ n ]]And a target signal y [ n ]]Training a weight matrix connecting a reservoir and an output layerAnd a weight matrix connecting between the input layer and the output layer +.>Wherein (1)> b represents a bias term, ">Lambda is a ridge regularization factor, and I is an identity matrix with the size of M+N+1;
input signal u [ n ]]According to the formula x [ n ]]＝α·f(W in ·u[n]+W res ·x[n-1])+(1-α)·x[n-1]Andcalculating to obtain output signal y n]。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311435573.2A CN117675009B (en) | 2023-11-01 | Dispersion compensation method based on reserve pool calculation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311435573.2A CN117675009B (en) | 2023-11-01 | Dispersion compensation method based on reserve pool calculation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117675009A true CN117675009A (en) | 2024-03-08 |
| CN117675009B CN117675009B (en) | 2024-06-11 |
Family
ID=
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070206954A1 (en) * | 2006-03-03 | 2007-09-06 | Fishman Daniel A | Simultaneous electrical pre-compensation of self-phase modulation and chromatic dispersion |
| US20110305453A1 (en) * | 2009-02-20 | 2011-12-15 | Nokia Siemens Networks Gmbh & Co. Kg | Method and arrangement for adaptive dispersion compensation |
| CN107764046A (en) * | 2016-08-19 | 2018-03-06 | 郑州东方安彩耐火材料有限公司 | Refractory material of electric-arc furnace cools down safely production method |
| CN112839142A (en) * | 2021-01-08 | 2021-05-25 | 淮阴工学院 | Shading control system |
| CN113472443A (en) * | 2021-05-26 | 2021-10-01 | 中山大学 | Transmitting end iterative dispersion compensation method for intensity modulation direct detection system |
| US20210328671A1 (en) * | 2018-08-01 | 2021-10-21 | Accelink Technologies Co., Ltd | Remote Optical Fiber Dispersion Compensation Device And Method, Remote Access Device, Electronic Device And Non-Transient Computer-Readable Storage Medium |
| KR20220118182A (en) * | 2021-02-18 | 2022-08-25 | 한국과학기술원 | Nonlinear blind equalizer using artificial neural networks |
| CN115483980A (en) * | 2021-05-31 | 2022-12-16 | 华为技术有限公司 | Optical receiver and optical signal processing method |
| CN115765876A (en) * | 2022-11-03 | 2023-03-07 | 复旦大学 | Probability shaping signal equalization method based on improved LSTM nonlinear equalizer |
| WO2023109065A1 (en) * | 2021-12-15 | 2023-06-22 | 苏州浪潮智能科技有限公司 | Method and system for implementing recurrent neural network, and electronic device and storage medium |
| WO2023154534A1 (en) * | 2022-02-11 | 2023-08-17 | Smithsonian Institution | Gas-liquid falling film equilibration system and method |
| CN116629332A (en) * | 2023-03-08 | 2023-08-22 | 电子科技大学 | Signal compensation method based on optical reserve tank calculation |
| CN116827732A (en) * | 2023-06-25 | 2023-09-29 | 香港理工大学深圳研究院 | Optical fiber communication method and system based on nonlinear decision feedback equalization of neural network |
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070206954A1 (en) * | 2006-03-03 | 2007-09-06 | Fishman Daniel A | Simultaneous electrical pre-compensation of self-phase modulation and chromatic dispersion |
| US20110305453A1 (en) * | 2009-02-20 | 2011-12-15 | Nokia Siemens Networks Gmbh & Co. Kg | Method and arrangement for adaptive dispersion compensation |
| CN107764046A (en) * | 2016-08-19 | 2018-03-06 | 郑州东方安彩耐火材料有限公司 | Refractory material of electric-arc furnace cools down safely production method |
| US20210328671A1 (en) * | 2018-08-01 | 2021-10-21 | Accelink Technologies Co., Ltd | Remote Optical Fiber Dispersion Compensation Device And Method, Remote Access Device, Electronic Device And Non-Transient Computer-Readable Storage Medium |
| CN112839142A (en) * | 2021-01-08 | 2021-05-25 | 淮阴工学院 | Shading control system |
| KR20220118182A (en) * | 2021-02-18 | 2022-08-25 | 한국과학기술원 | Nonlinear blind equalizer using artificial neural networks |
| CN113472443A (en) * | 2021-05-26 | 2021-10-01 | 中山大学 | Transmitting end iterative dispersion compensation method for intensity modulation direct detection system |
| CN115483980A (en) * | 2021-05-31 | 2022-12-16 | 华为技术有限公司 | Optical receiver and optical signal processing method |
| WO2023109065A1 (en) * | 2021-12-15 | 2023-06-22 | 苏州浪潮智能科技有限公司 | Method and system for implementing recurrent neural network, and electronic device and storage medium |
| WO2023154534A1 (en) * | 2022-02-11 | 2023-08-17 | Smithsonian Institution | Gas-liquid falling film equilibration system and method |
| CN115765876A (en) * | 2022-11-03 | 2023-03-07 | 复旦大学 | Probability shaping signal equalization method based on improved LSTM nonlinear equalizer |
| CN116629332A (en) * | 2023-03-08 | 2023-08-22 | 电子科技大学 | Signal compensation method based on optical reserve tank calculation |
| CN116827732A (en) * | 2023-06-25 | 2023-09-29 | 香港理工大学深圳研究院 | Optical fiber communication method and system based on nonlinear decision feedback equalization of neural network |
Non-Patent Citations (2)
| Title |
|---|
| YINKE YANG; FENG WEN; FENG YANG; KUN QIU: "Optical Dispersion Compensation through a Nonlinear-optical Loop Mirror (NOLM)-based Optical Reservoir", 2022 PHOTONICS & ELECTROMAGNETICS RESEARCH SYMPOSIUM (PIERS), no. 06, 16 June 2022 (2022-06-16) * |
| 张俊文;邵宇丰;方武良;黄博;黄德修;迟楠;: "基于单边带调制单载波频域均衡技术的研究", 光学学报, no. 07, 15 July 2010 (2010-07-15) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109039472B (en) | Data center optical communication dispersion estimation and management method based on deep learning | |
| CN111010239B (en) | Nonlinear phase noise compensation method and system in coherent optical fiber communication system | |
| Sang et al. | Low complexity neural network equalization based on multi-symbol output technique for 200+ Gbps IM/DD short reach optical system | |
| CN112598072B (en) | Equalization method of improved Volterra filter based on weight coefficient migration of SVM training | |
| CN109347555A (en) | A kind of visible light communication equalization methods based on radial basis function neural network | |
| Huang et al. | Performance and complexity analysis of conventional and deep learning equalizers for the high-speed IMDD PON | |
| CN114301529B (en) | Volterra equalization method and system based on multi-symbol processing | |
| Chen et al. | 50-km C-band transmission of 50-Gb/s PAM4 using 10-G EML and complexity-reduced adaptive equalization | |
| CN114615499A (en) | Semantic optical communication system and method for image transmission | |
| Zhang et al. | 56-Gbit/s PAM-4 optical signal transmission over 100-km SMF enabled by TCNN regression model | |
| CN117675009B (en) | Dispersion compensation method based on reserve pool calculation | |
| CN117675009A (en) | Dispersion compensation method based on reserve pool calculation | |
| CN114204993B (en) | Nonlinear equalization method and system based on polynomial mapping feature construction | |
| Qin et al. | Low-complexity bi-directional recurrent neural network equalizer for short-range optical interconnect links | |
| Dong et al. | Reservoir Computing for Dispersion Compensation in IMDD Transmission | |
| CN113938198B (en) | Optical fiber transmission system, LDA-based method and module for simplifying nonlinear equalizer | |
| CN114978313B (en) | Compensation method of visible light CAP system based on Bayesian neurons | |
| Osadchuk et al. | Low-complexity samples versus symbols-based neural network receiver for channel equalization | |
| Tang et al. | Non-uniform EDFA power map monitoring via power-guided learned digital backpropagation | |
| Yan et al. | An optoelectronic joint equalization scheme based on photonic reservoir computing and TDNN for CD and nonlinearity compensation in IM/DD link | |
| CN114500189B (en) | Direct pre-equalization method, system, device and medium for visible light communication | |
| Osadchuk et al. | Multi-Symbol Reservoir Computing-based Equalization for PAM-4 IM/DD Transmission | |
| Ren et al. | Meta-learning accelerated Bi-LSTM for fiber nonlinearity compensation | |
| CN111988089B (en) | Signal compensation method and system for optical fiber communication system | |
| Yao et al. | MPT-Transformer based post equalizer utilized in underwater visible light communication system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |