A multi-channel optical routing system module

Abstract

The present invention relates to the technical field of optical communication between chips, and discloses a multi-channel optical routing system module, comprising multiple MEMS lenses, reflectors, multiple vertical cavity surface emitting lasers (VCSELs), and multiple detectors. A VCSEL and a detector form a transceiver channel; a set of MEMS lenses is positioned above the channel, one of which reflects the laser beam emitted by the VCSEL to a transfer reflector in the channel, and the other reflects the laser beam from the transfer reflector to the detector. The transfer reflector and MEMS lenses work together to achieve real-time information transmission between different channels. Ultimately, optical signal communication is adjusted and optimized through real-time detection and feedback of optical signals.

Description

Technical Field

[0001] The invention relates to the technical field of optical communication between chips and discloses a multi-channel optical routing system module. Background Art

[0002] With the rapid development of information technology, the demand for high-performance computing, data centers, and Internet of Things (IoT) devices continues to increase. In these applications, data transmission speed, bandwidth, and energy efficiency have become critical metrics. Traditional electronic interconnect technologies have gradually reached their physical limits, especially in the field of short-distance, high-density interconnection. To overcome these limitations, inter-chip optical interconnect technology has emerged as a key component of next-generation communication architectures. Traditional electronic interconnect technology, based on copper wire signal transmission, has several limitations: As data rates increase, signal attenuation and distortion in copper wires increase, resulting in a decrease in effective bandwidth; high-speed signals transmitted over copper wires are susceptible to electromagnetic interference (EMI), affecting signal integrity and reliability; high-speed electronic interconnects generate high power consumption during signal transmission, posing challenges to device heat dissipation management and energy efficiency; and electronic interconnect technology is limited in high-density integration, making it difficult to meet the growing demand for integration. Optical interconnect technology uses light waves as an information carrier to establish high-speed communication links between chips. Compared with traditional electronic interconnects, optical interconnects offer significant advantages.

[0003] For example, a Chinese patent with the authorization publication number CN104849812A discloses an optical multi-channel router, which includes a housing, an input distributor, an output distributor, a switching fiber, a fiber input collimator, a fiber output collimator, multiple input collimators, multiple output collimators, multiple output fibers, multiple photodetectors, and multiple output ports. One end of the switching fiber is connected to a fiber input collimator, and the other end is connected to a fiber output collimator; the input distributor can drive the fiber input collimator to align with the multiple input collimators one by one; the output distributor can drive the fiber output collimator to align with the multiple output collimators one by one; and the multiple output fibers are respectively connected between the multiple output collimators and the multiple photodetectors. Because the multiple photodetectors are centrally arranged, a unified hardware configuration can be used, so that the photoelectric conversion efficiency is consistent when analyzing different physical and optical properties, reducing the measurement error caused by photoelectric conversion on optical analysis, and the experimental operation is simple and convenient.

[0004] The aforementioned patent relies on input and output splitters to drive the alignment of fiber collimators with multiple input / output collimators, resulting in a relatively complex mechanical structure. During frequent optical path switching or long-term operation, wear and looseness of mechanical components, as well as minor errors in manufacturing and assembly, may accumulate, resulting in a decrease in optical path alignment accuracy and affecting the transmission efficiency and accuracy of optical signals. Although the centralized arrangement of multiple photodetectors and the use of a unified hardware configuration ensure a certain degree of consistency in photoelectric conversion efficiency, they may not meet the requirements for high-sensitivity and specificity in detecting optical signals of different wavelengths and power ranges. The subsequent photoelectric signal processing is not elaborated in detail. Summary of the Invention

[0005] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0006] To solve the above technical problems, the main purpose of the present invention is to provide a multi-channel optical routing system module, comprising:

[0007] A transmitting module, comprising a plurality of vertical cavity surface emitting lasers;

[0008] The MEMS lens module includes a plurality of MEMS lenses, wherein the first MEMS lens is used to transfer and reflect the laser beam emitted by the transmitting laser and reflect it to the transfer reflector, and the second MEMS lens reflects the laser beam from the transfer reflector to the detector;

[0009] A transfer reflector, comprising an upper reflector and a lower reflector, wherein the upper reflector is used to receive the laser beam reflected by one MEMS lens, and the lower reflector is used to reflect the laser beam reflected by the upper reflector to the second MEMS lens;

[0010] a detector module, configured to receive the laser beam reflected by the second MEMS lens;

[0011] An optical signal detection module includes a data unit, a processing unit, and an optical analysis unit, wherein the data unit is used to collect optical power and spectrum, the processing unit is used to pre-process the collected optical power and spectrum optical signals, and the optical analysis unit is used to calculate the performance indicators of the optical signal;

[0012] The feedback module is used to obtain optical performance indicators and provide feedback control compensation.

[0013] As a preferred solution of a multi-channel optical routing system module of the present invention, wherein:

[0014] Multiple vertical cavity surface emitting lasers are integrated into one chip, and the detector module is integrated into another chip;

[0015] A vertical cavity surface emitting laser and a detector form a transceiver channel;

[0016] Multiple MEMS lenses are integrated into one MEMS module;

[0017] The MEMS mirror can reflect the laser beam emitted by the vertical cavity surface emitting laser into the transfer mirror or reflect the laser beam from the transfer mirror to the detector.

[0018] As a preferred solution of a multi-channel optical routing system module of the present invention, wherein:

[0019] In the transfer reflectors, every two reflectors form a group of transfer reflectors to realize the transfer transmission of laser beams in different channels, including a lower reflector and an upper reflector;

[0020] Furthermore, two opposite transmitting and receiving channels share a set of transfer reflectors;

[0021] The transfer reflector simultaneously transfers the light beam information emitted in the channel and the opposite channel, and transfers the light beam information received by the two channels.

[0022] As a preferred solution of a multi-channel optical routing system module of the present invention, wherein:

[0023] The data unit includes an optical power acquisition subunit and a spectrum acquisition subunit;

[0024] The processing unit includes a filtering and noise reduction module and a data calibration and normalization module;

[0025] The optical analysis unit calculates the signal optical power by:

[0026] S1. Detect the signal light peak, determine the preliminary area of ​​the light signal, and obtain spectral data;

[0027] S2. Determine whether the optical signal conforms to a preset distribution by fitting the spectral shape. If so, determine the optimal standard deviation through iterative optimization and further calculate the optical power of the optical signal. If not, the optical signal deviates and an early warning is issued.

[0028] S3, dividing the optical signal into regions and calculating the optical power of the optical signal in each region;

[0029] S4. Obtain the final optical power of all areas and add them up;

[0030] S5. Calculate the noise optical power and deduct the noise.

[0031] As a preferred solution of a multi-channel optical routing system module of the present invention, wherein:

[0032] If the spectral shape of the signal light in the initially determined signal light area conforms to the preset distribution, the optical power of the optical signal in each area is calculated by first obtaining the standard deviation of the spectrum in each area and then calculating the optical power of the optical signal in each area based on the obtained standard deviation.

[0033] The calculation expression for the total power density of each region spectrum is as follows:

[0034]

[0035] Where P(γ) is the optical power density of the optical signal at wavelength γ in the initial region, P max is the optical power density corresponding to the spectral peak in the region, γ is the wavelength of the optical signal in the input region, in nm, γ max is the maximum wavelength of the optical signal in the region, in nm, σ is the standard deviation of the spectrum in the region, in nm, and e is the exponential constant;

[0036] The optical power of the optical signal in each area is calculated using the obtained standard deviation. The calculation expression is as follows:

[0037]

[0038] Among them, P z is the total optical power of all optical signal areas, σ i is the standard deviation of the optical signal power in the i-th region, i is the optical preference segmentation region, and n is the total number of regions.

[0039] As a preferred solution of a multi-channel optical routing system module of the present invention, wherein:

[0040] The noise optical power calculation method includes:

[0041] S101, selecting a wavelength range far from the signal light region, calculating the average value and standard deviation of the optical power in the region, and using the average value as an estimated value of the background noise;

[0042] S102: After excluding the determined signal light area within the spectral range, the remaining area is the noise light area. A noise threshold is set, and points above the noise threshold are regarded as abnormal points, which are corrected and excluded.

[0043] S103. Calculate the noise light power for the determined noise light area. Divide the noise light area into multiple small wavelength intervals. The light power in each interval is approximately equal to the light power in the jth small wavelength interval of the noise light area. Calculate the noise light power.

[0044] The calculation expression of noise optical power is as follows:

[0045]

[0046] Among them, P noi is the noise optical power, P j is the optical power in the jth small wavelength interval of the noise light area, Δλ j is the width of the jth small wavelength interval divided by the noise light area, j is the optical signal area, and m is the total number of areas.

[0047] As a preferred solution of a multi-channel optical routing system module of the present invention, wherein:

[0048] Collect the data series of optical power changes over time and spectral data, and extract features from the optical power data;

[0049] Build a light performance index prediction model, use the extracted features as input, and use the known light performance indicators as output labels for model training;

[0050] The optical performance index prediction model consists of an input layer, several hidden layers, and an output layer. The number of neurons in the input layer is equal to the number of input optical signal data features, n, and the number of neurons in the output layer is equal to the number of output optical performance indicators, m. Assume there are L hidden layers, and the number of neurons in the Zth layer (Z = 1, 2, 3, ..., L) is h.

[0051] For the connection weights between neurons, let the weight matrix from the input layer to the first hidden layer be W1, with a dimension of h1×(n+1), then the weight matrix from the Z-1th hidden layer to the lth hidden layer is W Z , dimension is h Z ×(h Z-1 +1);

[0052] The forward propagation includes inputting the optical signal data feature into the input layer, calculating the optical performance index vector to the first hidden layer through the activation function, the optical performance index vector performing propagation calculation between multiple hidden layers, the last hidden layer receiving the optical performance index vector, and propagating the optical performance index vector received through the last hidden layer to the output layer, and calculating and outputting the optical performance index vector through the output layer.

[0053] As a preferred solution of a multi-channel optical routing system module of the present invention, wherein:

[0054] The difference between the predicted value of the optical performance index vector and the true value of the optical performance index vector is quantified by the loss function. The calculation expression is as follows:

[0055]

[0056] Among them, T(W) is the loss value, is the input optical performance index vector prediction value, y i1 is the true value of the input optical performance index vector, I is the total number of data, and i1 is the number of input optical performance index vector groups.

[0057] As a preferred solution of a multi-channel optical routing system module of the present invention, wherein:

[0058] Feedback compensation includes compensation strategy generation, compensation execution and feedback adjustment;

[0059] Compare and analyze the predicted optical performance indicators with the preset target performance indicators;

[0060] Based on the analysis results, a compensation strategy is generated. If it is an attenuation problem, the amount of gain to increase is determined; if it is a transmission power problem, the drive current or voltage of the VCSEL is adjusted.

[0061] Determine the length of the dispersion-compensating fiber and the parameter adjustment amount in the digital dispersion compensation algorithm based on the dispersion amount of the optical performance index prediction model;

[0062] For nonlinear effect compensation, pre-compensation parameters of the optical signal are adjusted.

[0063] A computer device includes: a memory for storing instructions; a processor for executing the instructions, so that the device implements the above-mentioned multi-channel optical routing system module.

[0064] A computer-readable storage medium stores a computer program, which, when executed, implements the multi-channel optical routing system module.

[0065] Beneficial effects of the present invention:

[0066] Multiple vertical-cavity surface-emitting lasers and detector modules are integrated into a single chip. This integrated design helps reduce the physical size of the entire system, improving its compactness and facilitating deployment in space-constrained applications. It also helps mitigate signal transmission loss and electromagnetic interference issues that may arise from the dispersed layout of components, improving the overall stability and reliability of the system. An optical performance prediction model is constructed. The model is trained based on the collected time-varying optical power data series and feature extraction from spectral data. With a rational architecture of input, hidden, and output layers, the inherent relationship between optical signal characteristics and optical performance indicators can be learned based on large amounts of data, enabling effective prediction of optical performance indicators and proactively understanding performance trends. The feedback compensation process encompasses compensation strategy generation, compensation execution, and feedback adjustment. Based on the comparison and analysis results of the predicted optical performance indicators with the preset target performance indicators, targeted compensation strategies are generated. Whether addressing attenuation, transmit power issues, dispersion, or nonlinear effects, appropriate adjustments can be made to ensure that the optical signal consistently maintains optimal performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0068] in:

[0069] Figure 1 This is a system composition diagram of a multi-channel optical routing system module of the present invention;

[0070] Figure 2 This is a working diagram of a multi-channel optical routing system module of the present invention;

[0071] Figure 3 This is a schematic diagram of communication between two channels of a multi-channel optical routing system module of the present invention;

[0072] Figure 4 This is a schematic diagram of an array chip of lasers and detectors for the transmitting and receiving channels on the same side of a multi-channel optical routing system module of the present invention. DETAILED DESCRIPTION

[0073] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0074] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0075] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0076] Example 1

[0077] like Figure 1 As shown, a multi-channel optical routing system module includes:

[0078] A transmitting module, comprising a plurality of vertical cavity surface emitting lasers;

[0079] Among them, multiple vertical cavity surface emitting lasers and detector modules are each integrated into one chip;

[0080] Furthermore, a vertical cavity surface emitting laser and a detector form a transceiver channel;

[0081] The MEMS lens module includes a plurality of MEMS lenses, wherein the first MEMS lens is used to transfer and reflect the laser beam emitted by the transmitting laser and reflect it to the transfer reflector, and the second MEMS lens reflects the laser beam from the transfer reflector to the detector;

[0082] Furthermore, multiple MEMS lenses are integrated into one MEMS module;

[0083] The MEMS lens can reflect the laser beam emitted by the vertical cavity surface emitting laser into the transfer reflector or reflect the laser beam from the transfer reflector to the detector;

[0084] Multiple MEMS lenses and transfer mirrors work together to achieve real-time communication between multiple channels;

[0085] A transfer reflector, comprising an upper reflector and a lower reflector, wherein the upper reflector is used to receive the laser beam reflected by one MEMS lens, and the lower reflector is used to reflect the laser beam reflected by the upper reflector to the second MEMS lens;

[0086] Among them, in the transfer reflectors, every two reflectors form a group of transfer reflectors to realize the transfer transmission of laser beams in different channels, including a lower reflector and an upper reflector;

[0087] Among them, two opposite transmitting and receiving channels share a set of transfer reflectors;

[0088] Furthermore, the transfer reflector can simultaneously transfer the light beam information emitted in the channel and the opposite channel and the light beam information required to be received by the two channels;

[0089] a detector module, configured to receive the laser beam reflected by the second MEMS lens;

[0090] An optical signal detection module includes a data unit, a processing unit, and an optical analysis unit, wherein the data unit is used to collect optical power and spectrum, the processing unit is used to pre-process the collected optical power and spectrum optical signals, and the optical analysis unit is used to calculate the performance indicators of the optical signal;

[0091] The feedback module is used to obtain optical performance indicators and provide feedback control compensation.

[0092] Example 2

[0093] like Figure 2 As shown, the multi-channel optical routing system module schematic diagram of the embodiment, the markings in the figure include

[0094] Vertical cavity surface emitting laser 1 in the transceiver channel; MEMS lens 2 corresponding to the vertical cavity surface emitting laser; lower reflector 3 of the transfer reflector; upper reflector 4 of the transfer reflector; MEMS lens 5 corresponding to the detector; detector 6 in the transceiver channel; different transceiver channels Chan.1, Chan.1'...Chan.N+1, Chan.N+1'.

[0095] like Figure 3 As shown, including:

[0096] The information transmission of the transceiver channel is controlled by multiple MEMS lenses integrated in two MEMS modules, such as Figure 1As shown, the vertical cavity surface emitting laser in Chan.1 emits a laser beam carrying information, which is deflected by its corresponding MEMS lens and transmitted to the transfer reflector corresponding to Chan.3', and then transmitted to the detector in Chan.3'; at the same time, the vertical cavity surface emitting laser in Chan.N' emits a laser beam carrying information, which is deflected by its corresponding MEMS lens and transmitted to the transfer reflector corresponding to Chan.1, and then transmitted to the detector in Chan.1; in this way, real-time communication between different channels can be achieved. The vertical cavity surface emitting laser and detector in the transceiver channel on the same side are a highly integrated vertical cavity surface emitting laser array chip and detector array chip, and the scale of the array chip can be set according to actual needs. The scale of the MEMS lens module on the upper side of the transceiver channel is designed according to the array scale of the laser and detector below it. The array scale of the transfer reflector is also set according to the array scale of the transceiver channel;

[0097] Furthermore, specific application scenarios include fast data exchange between different servers in a data center; and on-chip integration can achieve fast communication between different chips, reduce communication latency, and increase communication bandwidth;

[0098] Furthermore, the high-speed VCSEL chip converts the electrical signal into a laser beam carrying information, and the high-speed detector in the receiving channel converts the received optical signal carrying information into an electrical signal. The optical signal is transmitted in the optical router, and both the input and output ends of the optical router transmit electrical signals.

[0099] The optical router converts the electrical signal received by the detector from the chip or server into a laser beam (i.e., optical signal) that carries the information through the internal high-speed VCSEL.

[0100] The specific implementation includes: the high-speed VCSEL device 7 in the transmitting channel receives the external electrical signal, converts it into an optical signal carrying information, and transmits it to the MEMS lens 8 thereon. The MEMS lens 8 deflects a specific angle to reflect the optical signal to the lower reflector 9 in the relay reflective module. The lower reflector reflects the optical signal to the upper reflector 10 in the relay module. The upper reflector transmits the optical signal again to the corresponding MEMS lens 11 in the high-speed detector in the receiving channel. The MEMS lens then transmits the optical signal to the high-speed detector 12 in the receiving channel, which converts it into an electrical signal and transmits it to the receiving device. Figure 1 As shown, Chan.1, Chan.1', Chan.2, Chan.2', Chan.3, Chan.3'...Chan.N, Chan.N', Chan.N+1, Chan.N+1' can all be used as transmitters and receivers, and in this way, simultaneous communication between different devices can be achieved.

[0101] Example 3

[0102] An optical signal detection module includes a data unit, a processing unit, and an optical analysis unit, wherein the data unit is used to collect optical power and spectrum, the processing unit is used to pre-process the collected optical power and spectrum optical signals, and the optical analysis unit is used to calculate the performance indicators of the optical signal;

[0103] The data unit includes an optical power acquisition subunit and a spectrum acquisition subunit;

[0104] The optical power acquisition subunit uses a high-sensitivity, wide-dynamic-range photodetector array, which is distributed at key nodes of the optical communication link, such as the optical transmitter, before and after the optical amplifier, and at the optical receiver, to achieve multi-point real-time acquisition of optical power.

[0105] The photodetector converts the received optical power signal into an electrical signal, which is amplified by a low-noise amplifier circuit to increase the signal amplitude for subsequent analog-to-digital conversion (ADC).

[0106] The ADC module converts the amplified analog electrical signal into a digital signal. The sampling rate needs to be reasonably set according to the bandwidth of the optical communication system and the signal change rate. For example, for high-speed optical communication systems, the sampling rate can be set to several Gsps or even higher to ensure that the details of the optical power changes can be accurately captured.

[0107] The spectrum acquisition subunit uses a spectroscopic element such as a diffraction grating or an interferometer to decompose the optical signal according to different wavelengths.

[0108] Furthermore, the spectrum acquisition subunit uses a linear array or planar array of photoelectric detectors to detect the decomposed light components of different wavelengths and convert the light intensity signal into an electrical signal.

[0109] After amplification and ADC conversion, a digital signal representing the spectrum is obtained. To improve spectral resolution, the splitting capability of the spectrometer and the number and accuracy of the detector pixels need to be carefully designed and selected. For example, for dense wavelength division multiplexing (DWDM) systems, the spectral resolution may need to reach the picometer level.

[0110] The processing unit includes a filtering and noise reduction module and a data calibration and normalization module;

[0111] The filtering and noise reduction module applies digital filters to the collected optical power and spectrum digital signals. For optical power signals, a low-pass filter removes high-frequency noise interference. For spectral signals, a bandpass filter or a filter with a specific shape can be used to reduce the impact of noise on spectral analysis, depending on the spectral characteristics.

[0112] Furthermore, the filtering and noise reduction module uses signal averaging or adaptive noise reduction algorithms to further reduce noise. For example, signal averaging improves the signal-to-noise ratio by sampling the same signal multiple times and averaging it. Adaptive noise reduction algorithms dynamically adjust filter parameters based on the statistical characteristics of the signal and noise to achieve optimal noise reduction.

[0113] Furthermore, because the response characteristics of photodetectors may exhibit nonlinearities and temperature drift, data calibration is required. By pre-measuring the detector's response curve at different optical powers and wavelengths and storing it in a lookup table, the optical power and spectral data are calibrated during actual operation based on the collected signal and the lookup table data to ensure data accuracy.

[0114] Through data calibration and normalization, the calibrated optical power and spectral data are normalized to a unified numerical range or reference standard, which facilitates subsequent analysis and comparison. For example, the optical power data is normalized to a specific reference power value, and the spectral data is normalized to a specific wavelength range and intensity standard;

[0115] The optical analysis unit calculates the signal optical power by:

[0116] S1. Detect the signal light peak, determine the preliminary area of ​​the light signal, and obtain spectral data;

[0117] S2. Determine whether the optical signal conforms to a preset distribution by fitting the spectral shape. If so, determine the optimal standard deviation through iterative optimization and further calculate the optical power of the optical signal. If not, the optical signal deviates and an early warning is issued.

[0118] S3, dividing the optical signal into regions and calculating the optical power of the optical signal in each region;

[0119] Scan the spectral data to find the wavelength with the maximum optical power value. The power value corresponding to this point is the maximum power value point. With the maximum wavelength point as the center, expand a certain wavelength range in both directions (for example, based on the expected signal light bandwidth and spectral resolution of the system, expand an initial range first). This range is used as the preliminary area where the signal light may exist.

[0120] Furthermore, within the initially determined signal light area, assuming that the spectrum shape of the signal light conforms to a Gaussian distribution, the optical power of the optical signal in each area is calculated to obtain the optical power of the optical signal in each area;

[0121] The calculation expression for the total power density of each region spectrum is as follows:

[0122]

[0123] Where P(γ) is the optical power density of the optical signal at wavelength γ in the initial region, P max is the optical power density corresponding to the spectral peak in the region, γ is the wavelength of the optical signal in the input region, in nm, γ max is the maximum wavelength of the optical signal in the region, in nm, σ is the standard deviation of the spectrum in the region, in nm, and e is the exponential constant;

[0124] S4. Obtain the final optical power of all areas and add them up;

[0125] The optical power of the optical signal in each area is calculated using the obtained standard deviation. The calculation expression is as follows:

[0126]

[0127] Among them, P z is the total optical power of all optical signal areas, σ i is the standard deviation of the optical signal power in the i-th region, i is the light preference segmentation region, and n is the total number of regions;

[0128] S5. Calculate the noise optical power and deduct the noise;

[0129] The noise optical power calculation method includes:

[0130] S101. In the spectral data, select a wavelength range far from the signal light region (for example, a range with a width of λ at the edge of the spectrum at both ends), and calculate the average and standard deviation of the optical power in this region. The average value is used as an estimate of the background noise.

[0131] S102. After excluding the identified signal light region within the entire spectral range, the remaining region is the noise light region. However, this region needs to be further screened and processed to remove any outliers that may be caused by instrument errors or other interference. For example, a threshold value (a constant determined based on system reliability requirements, generally between 3 and 5) can be set based on the statistical characteristics of the noise. Points above this threshold value are considered outliers and corrected or excluded.

[0132] S103 , for the determined noise light area, calculate the noise light power, divide the noise light area into multiple small wavelength intervals, and approximate the light power in each interval to the light power in the jth small wavelength interval of the noise light area, and calculate the noise light power.

[0133] Furthermore, the calculation expression of noise optical power is as follows:

[0134]

[0135] Among them, Pnoi is the noise optical power, P j is the optical power in the jth small wavelength interval of the noise light area, Δλ j is the width of the jth small wavelength interval divided by the noise light area, j is the light signal area, and m is the total number of areas;

[0136] The width of the spectral edge region of the estimated background noise, the average optical power of the background noise region, and the standard deviation of the optical power of the background noise region are selected to calculate the noise optical power in the region;

[0137] Data acquisition segmentation: The data unit includes an optical power acquisition subunit and a spectrum acquisition subunit, which can perform specialized data acquisition on the two key optical signal elements of optical power and spectrum, respectively, to obtain more comprehensive and detailed original information of the optical signal, laying the foundation for subsequent accurate analysis.

[0138] Improved processing flow: The processing unit includes filtering and noise reduction modules as well as data calibration and normalization. Filtering and noise reduction can effectively remove noise interference in optical signals and improve signal quality. Data calibration and normalization operations can bring the collected data into a more appropriate standard range, facilitating subsequent unified analysis and comparison, and ensuring the accuracy and usability of the processed optical signal data.

[0139] Scientific optical power calculation: The optical analysis unit has a rigorous procedure for calculating the optical power of optical signals. From detecting the signal light peak, judging whether the spectral shape conforms to the preset distribution, regional segmentation calculation, to noise subtraction and other multi-step operations, it can accurately determine the actual optical power of the optical signal. Even in the face of complex optical signal conditions, it can obtain the effective power value as accurately as possible, reducing the deviation of optical signal performance evaluation caused by calculation errors.

[0140] Reasonable noise calculation: The noise optical power calculation method can more accurately determine the noise optical power by selecting an appropriate wavelength range to estimate background noise, correcting and eliminating abnormal points, and reasonably dividing the interval calculation. Then, the noise effect can be more accurately deducted in the final optical power calculation, further improving the accuracy of optical power calculation, making the detection and analysis of optical signals more in line with actual conditions.

[0141] Feedback module, used to detect light performance indicators and provide feedback control compensation;

[0142] A high-resolution spectrometer and an optical power meter are used to monitor the optical signal in real time, collecting data series of optical power changes over time and spectral data, including the light intensity distribution at different wavelengths.

[0143] Preprocess the collected data, such as removing outliers (which may be caused by measurement noise or transient interference) and smoothing the data (using moving average or Savitzky-Golay filtering) to reduce the impact of high-frequency noise on subsequent analysis;

[0144] Features are extracted from the optical power data, such as the average value, variance, peak value, minimum value, and power change rate (by calculating the ratio of the power difference between adjacent time points to the time interval). These features can reflect the intensity stability and fluctuation of the optical signal.

[0145] For spectral data, calculate the spectral bandwidth (by determining the wavelength range corresponding to the decrease of spectral intensity from the peak to a certain proportion), the central wavelength (calculated based on the weighted average of spectral intensity), the spectral flatness (measures the uniformity of intensity at different wavelengths of the spectrum, such as calculating the ratio of the standard deviation of spectral intensity to the average value) and other characteristics. In addition, the spectrum can also be analyzed in the frequency domain to extract spectral characteristics, such as the energy distribution of different frequency components.

[0146] Construct an optical performance indicator prediction model. Use the extracted features as input and known optical performance indicators (such as optical signal-to-noise ratio (OSNR), dispersion, nonlinear effect intensity, etc.) as output labels for model training.

[0147] During training, cross-validation techniques (such as k-fold cross-validation) are used to optimize the model's hyperparameters to improve the model's generalization ability and prediction accuracy. For example, for neural networks, hyperparameters such as the number of hidden layers, number of neurons, learning rate, and activation function can be adjusted; for support vector machines, the kernel function type and parameters can be adjusted.

[0148] After the training is completed, the optical signal data collected in real time and extracted with features is input into the trained model, and the model can predict various performance indicators of the optical signal;

[0149] The input optical signal data feature vector is X=[x1,x2,...,x n ], where X is the feature of the optical signal data, x1 is the first feature of the optical signal data, x2 is the second feature of the optical signal data, and x n is the nth feature of the optical signal data;

[0150] The output light performance index vector is Y=[y1,y2,...,y m ], where Y is the comprehensive optical performance index, y1 is the optical signal-to-noise ratio (OSNR), the first performance index of the optical signal, y2 is the dispersion of the second performance index of the optical signal, and y m is the mth performance index of the optical signal;

[0151] The optical performance index prediction model consists of an input layer, several hidden layers, and an output layer. The number of neurons in the input layer is equal to the number of input optical signal data features, n, and the number of neurons in the output layer is equal to the number of output optical performance indicators, m. Assume there are L hidden layers, and the number of neurons in the Zth layer (Z = 1, 2, 3, ..., L) is h.

[0152] Furthermore, for the connection weights between neurons, let the weight matrix from the input layer to the first hidden layer be W1, with a dimension of h1×(n+1), then the weight matrix from the Z-1th hidden layer to the lth hidden layer is W Z , dimension is h Z ×(h Z-1 +1);

[0153] Forward propagation includes inputting optical signal data features into an input layer, calculating an optical performance index vector to a first hidden layer through an activation function, performing propagation calculations on the optical performance index vector between multiple hidden layers, receiving the optical performance index vector in a last hidden layer, propagating the optical performance index vector received by the last hidden layer to an output layer, and calculating and outputting the optical performance index vector through the output layer;

[0154] The optical performance index vector is passed from the first hidden layer to the Lth hidden layer. The calculation expression is as follows:

[0155]

[0156] in, is the predicted value of the optical performance index output by the output layer, is the weight of the kth group of optical performance index vector output by the Z+1th hidden layer, X k Z The optical performance index vector input from the Zth hidden layer to the Z+1th hidden layer;

[0157] The difference between the predicted value of the optical performance index vector and the true value of the optical performance index vector is quantified by the loss function. The calculation expression is as follows:

[0158]

[0159] Among them, T(W) is the loss value, is the input optical performance index vector prediction value, y i1 is the true value of the input optical performance index vector, I is the total number of data, and i1 is the number of input optical performance index vector groups;

[0160] Furthermore, feedback compensation includes compensation strategy generation and compensation execution and feedback adjustment;

[0161] Furthermore, the predicted optical performance indicators are compared and analyzed with the preset target performance indicators. For example, if the predicted OSNR is lower than the target value, the cause is determined to be insufficient optical power or excessive noise. If the problem is optical power, further analysis is conducted to determine whether it is caused by attenuation during transmission or low optical transmit power itself.

[0162] Construct an optical performance index prediction model, train the model based on the collected data series of optical power changes over time and the features extracted from spectral data. With a reasonable structure of input layer, hidden layer and output layer, the intrinsic relationship between optical signal characteristics and optical performance indicators can be learned based on a large amount of data, thereby achieving effective prediction of optical performance indicators and grasping the performance trend of optical signals in advance.

[0163] Based on the analysis results, a compensation strategy is generated. If it is an attenuation problem, the amount of gain to be added at the appropriate location (such as the optical amplifier) ​​is determined; if it is a transmission power problem, the driving current or voltage of the optical transmitter can be adjusted. For dispersion compensation, the length of the dispersion-compensating fiber or the parameter adjustment in the digital dispersion compensation algorithm can be determined based on the predicted dispersion amount. For compensation of nonlinear effects, the pre-compensation parameters of the optical signal can be adjusted (such as pre-chirping the signal at the transmitting end) or the iterative step size and compensation coefficient in the digital backpropagation algorithm can be used.

[0164] Furthermore, the generated compensation strategy is converted into a specific control signal and sent to the corresponding optical communication equipment components, such as multiple MEMS lenses, vertical cavity surface emitting lasers, multiple detectors (for dispersion and nonlinear compensation), etc., to perform compensation operations.

[0165] After compensation is executed, the optical signal is continuously monitored using the aforementioned data collection and performance indicator detection methods to evaluate the compensation effect. If the performance indicator after compensation still does not meet the target requirement, the new performance indicator data and compensation operation information are fed back to the compensation strategy generation module. The compensation strategy is then adjusted and optimized using the reward mechanism in the reinforcement learning algorithm. For example, positive rewards are given when the compensated performance indicator approaches the target value, and negative rewards are given when it does not. This process is repeated until the optical signal performance indicator meets the requirements or reaches the optimal state.

[0166] The feedback compensation process covers compensation strategy generation, compensation execution, and feedback adjustment. It can generate targeted compensation strategies based on the comparative analysis results of the predicted optical performance indicators and the preset target performance indicators. Whether it is dealing with attenuation problems, transmission power problems, dispersion, nonlinear effects, etc., corresponding adjustment measures can be taken to ensure that the optical signal can always maintain a good performance state, improve the stability and communication quality of the entire multi-channel optical routing system, and reduce the risks of communication errors and interruptions caused by poor optical signal performance.

[0167] It is important to note that the construction and arrangement of the present application shown in a plurality of different exemplary embodiments are merely illustrative. Although only two embodiments are described in detail in this disclosure, it should be readily understood by those who refer to this disclosure that, without departing substantially from the novel teachings and advantages of the subject matter described in this application, many modifications are possible, for example, the size, scale, structure, shape and ratio of various elements, as well as parameter values ​​(e.g., temperature, pressure, etc.), mounting arrangements, use of materials, color, directional changes, etc. For example, an element shown as region-shaped can be composed of multiple parts or elements, the position of the element can be inverted or otherwise changed, and the nature or number or position of the discrete elements can be altered or changed. Therefore, all such modifications are intended to be included within the scope of the present invention. The order or sequence of any process or method steps can be changed or reordered according to alternative embodiments. Any "device plus function" clause is intended to cover the structure of the execution function described herein, and is not only structurally equivalent but also equivalent structures. Without departing from the scope of the present invention, other replacements, modifications, changes and omissions can be made in the design, operating conditions and arrangement of the exemplary embodiments. Therefore, the invention is not limited to the specific embodiments, but extends to various modifications that still fall within the scope of the appended claims.

[0168] Additionally, in order to provide a concise description of exemplary embodiments, all features of an actual embodiment (ie, those features that are not relevant to the best mode presently contemplated for carrying out the invention or those that are not relevant to implementing the invention) may not be described.

[0169] It should be understood that in the development of any actual embodiment, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but for those of ordinary skill having the benefit of this disclosure, the development effort will be a routine task of design, fabrication, and production without undue experimentation.

[0170] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A multi-channel optical routing system module, characterized in that: include: A transmitting module, comprising a plurality of vertical cavity surface emitting lasers; The MEMS lens module includes a plurality of MEMS lenses, wherein the first MEMS lens is used to transfer and reflect the laser beam emitted by the transmitting laser and reflect it to the transfer reflector, and the second MEMS lens reflects the laser beam from the transfer reflector to the detector; A transfer reflector, comprising an upper reflector and a lower reflector, wherein the upper reflector is used to receive the laser beam reflected by one MEMS lens, and the lower reflector is used to reflect the laser beam reflected by the upper reflector to the second MEMS lens; a detector module, configured to receive the laser beam reflected by the second MEMS lens; An optical signal detection module includes a data unit, a processing unit, and an optical analysis unit, wherein the data unit is used to collect optical power and spectrum, the processing unit is used to pre-process the collected optical power and spectrum optical signals, and the optical analysis unit is used to calculate the performance indicators of the optical signal; The feedback module is used to obtain optical performance indicators and provide feedback control compensation.

2. The multi-channel optical routing system module according to claim 1, wherein: Multiple vertical cavity surface emitting lasers are integrated into one chip, and the detector module is integrated into another chip; A vertical cavity surface emitting laser and a detector form a transceiver channel; Multiple MEMS lenses are integrated into one MEMS lens module; The MEMS mirror can reflect the laser beam emitted by the vertical cavity surface emitting laser into the transfer mirror or reflect the laser beam from the transfer mirror to the detector.

3. The multi-channel optical routing system module according to claim 2, wherein: In the transfer reflectors, every two reflectors form a group of transfer reflectors to realize the transfer transmission of laser beams in different channels, including a lower reflector and an upper reflector; Furthermore, two opposite transmitting and receiving channels share a set of transfer reflectors; The transfer reflector simultaneously transfers the light beam information emitted in the channel and the opposite channel, and transfers the light beam information received by the two channels.

4. The multi-channel optical routing system module according to claim 3, wherein: The data unit includes an optical power acquisition subunit and a spectrum acquisition subunit; The processing unit includes a filtering and noise reduction module and a data calibration and normalization module; The optical analysis unit calculates the signal optical power by: S1. Detect the signal light peak, determine the preliminary area of ​​the light signal, and obtain spectral data; S2. Determine whether the optical signal conforms to a preset distribution by fitting the spectral shape. If so, determine the optimal standard deviation through iterative optimization and further calculate the optical power of the optical signal. If not, the optical signal deviates and an early warning is issued. S3, dividing the optical signal into regions and calculating the optical power of the optical signal in each region; S4. Obtain the final optical power of all areas and add them up; S5. Calculate the noise optical power and deduct the noise.

5. The multi-channel optical routing system module according to claim 4, characterized in that: If the spectral shape of the signal light in the initially determined signal light area conforms to the preset distribution, the optical power of the optical signal in each area is calculated by first obtaining the standard deviation of the spectrum in each area and then calculating the optical power of the optical signal in each area based on the obtained standard deviation. The calculation expression for the total power density of each region spectrum is as follows: Where P(γ) is the optical power density of the optical signal at wavelength γ in the initial region, P max is the optical power density corresponding to the spectral peak in the region, γ is the wavelength of the optical signal in the input region, in nm, γ max is the maximum wavelength of the optical signal in the region, in nm, σ is the standard deviation of the spectrum in the region, in nm, and e is the exponential constant; The optical power of the optical signal in each area is calculated using the obtained standard deviation. The calculation expression is as follows: Among them, P z is the total optical power of all optical signal areas, σ i is the standard deviation of the optical signal power in the i-th region, i is the optical preference segmentation region, and n is the total number of regions.

6. The multi-channel optical routing system module according to claim 5, characterized in that: The noise optical power calculation method includes: S101, selecting a wavelength range far from the signal light region, calculating the average value and standard deviation of the optical power in the region, and using the average value as an estimated value of the background noise; S102: After excluding the determined signal light area within the spectral range, the remaining area is the noise light area. A noise threshold is set, and points above the noise threshold are regarded as abnormal points, which are corrected and excluded. S103. Calculate the noise light power for the determined noise light area. Divide the noise light area into multiple small wavelength intervals. The light power in each interval is approximately equal to the light power in the jth small wavelength interval of the noise light area. Calculate the noise light power. The calculation expression of noise optical power is as follows: Among them, P noi is the noise optical power, P j is the optical power in the jth small wavelength interval of the noise light area, Δλ j is the width of the jth small wavelength interval divided by the noise light area, j is the optical signal area, and m is the total number of areas.

7. The multi-channel optical routing system module according to claim 6, characterized in that: Collect the data series of optical power changes over time and spectral data, and extract features from the optical power data; Build a light performance index prediction model, use the extracted features as input, and use the known light performance indicators as output labels for model training; The optical performance index prediction model consists of an input layer, several hidden layers, and an output layer. The number of neurons in the input layer is equal to the number of input optical signal data features n, and the number of neurons in the output layer is equal to the number of output optical performance indicators m. Suppose there are L hidden layers and the number of neurons in the Zth layer (Z = 1, 2, 3, ..., L) is h; For the connection weights between neurons, let the weight matrix from the input layer to the first hidden layer be W1, with a dimension of h1×(n+1), then the weight matrix from the Z-1th hidden layer to the lth hidden layer is W Z , dimension is h Z ×(h Z-1 +1); The forward propagation includes inputting the optical signal data feature into the input layer, calculating the optical performance index vector to the first hidden layer through the activation function, the optical performance index vector performing propagation calculation between multiple hidden layers, the last hidden layer receiving the optical performance index vector, and propagating the optical performance index vector received through the last hidden layer to the output layer, and calculating and outputting the optical performance index vector through the output layer.

8. The multi-channel optical routing system module according to claim 7, characterized in that: The difference between the predicted value of the optical performance index vector and the true value of the optical performance index vector is quantified by the loss function. The calculation expression is as follows: Among them, T(W) is the loss value, is the input optical performance index vector prediction value, y i1 is the true value of the input optical performance index vector, I is the total number of data, and i1 is the number of input optical performance index vector groups.

9. The multi-channel optical routing system module according to claim 8, characterized in that: Feedback compensation includes compensation strategy generation, compensation execution and feedback adjustment; Compare and analyze the predicted optical performance indicators with the preset target performance indicators; Based on the analysis results, a compensation strategy is generated. If it is an attenuation problem, the amount of gain to increase is determined; if it is a transmission power problem, the drive current or voltage of the VCSEL is adjusted. Determine the length of the dispersion-compensating fiber and the parameter adjustment amount in the digital dispersion compensation algorithm based on the dispersion amount of the optical performance index prediction model; For nonlinear effect compensation, pre-compensation parameters of the optical signal are adjusted.

10. A computer device, characterized in that: include: a memory for storing instructions; The processor is configured to execute the instructions so that the device implements a multi-channel optical routing system module as claimed in any one of claims 1 to 9.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed, a multi-channel optical routing system module as claimed in any one of claims 1 to 9 is implemented.

Images