Laser design system, laser design method, electronic equipment and medium

Abstract

The invention discloses a laser design system, a laser design method, electronic equipment and a medium, and belongs to the technical field of integrated optics. The laser design system comprises a first model, input parameters of which comprise set grating parameters of a grating structure of a first to-be-designed laser, and which is used for predicting and outputting a predicted output characteristic curve of the first to-be-designed laser based on the set grating parameters; the input parameters of the second model comprise a set output characteristic curve of a second laser to be designed, and the second model is used for predicting and outputting predicted grating parameters of a grating structure of the second laser to be designed based on the set output characteristic curve; the first model and the second model share the same training data set, the input training data of the first model corresponds to the output training data of the second model, and the output training data of the first model corresponds to the input training data of the second model; the training data set includes pairing data between grating parameters of a grating structure of the sample laser and an output characteristic curve of the sample laser.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of integrated optics, and particularly relates to a laser design system, a laser design method, an electronic device and a medium. BACKGROUND

[0002] Laser design is a core link of the research and development of optoelectronic devices, and the performance of a laser directly determines the efficiency, precision and stability of a related system. There is a complex nonlinear relationship between the structural parameters and optical performance of a laser. At present, the design of a laser relies on experience and trial-and-error and time-consuming numerical simulation, and the design cycle is long and the precision of the finished product is low. SUMMARY

[0003] The present application aims to at least solve one of the technical problems existing in the prior art. To this end, the present application provides a laser design system, a laser design method, an electronic device and a medium, which can significantly improve the design efficiency and precision of a laser.

[0004] In a first aspect, the present application provides a laser design system, comprising: a first model, an input parameter of the first model comprising a set grating parameter of a grating structure of a first to-be-designed laser, the first model being configured to predict and output a predicted output characteristic curve of the first to-be-designed laser based on the set grating parameter; a second model, an input parameter of the second model comprising a set output characteristic curve of a second to-be-designed laser, the second model being configured to predict and output a predicted grating parameter of a grating structure of the second to-be-designed laser based on the set output characteristic curve; the first model and the second model share a same training data set, input training data of the first model in the training data set corresponding to output training data of the second model, output training data of the first model in the training data set corresponding to input training data of the second model, the training data set comprising paired data between grating parameters of a grating structure of a sample laser and an output characteristic curve of the sample laser.

[0005] According to the laser design system of the present application, by constructing two first models and second models which are inverse mappings of each other, and based on a same training data set for joint training, bidirectional closed-loop optimization from structural parameters to performance prediction and target performance to structural reverse design can be achieved. The first model can quickly predict the output characteristic curve corresponding to a given grating parameter, accelerating performance evaluation, and the second model can inversely deduce the optimal grating parameter according to the expected output characteristic curve, realizing reverse design of high-performance devices, which can significantly improve the design efficiency and precision of a laser.

[0006] According to one embodiment of this application, the training dataset includes at least one of a first subset and a second subset, wherein the first subset is generated by a transfer matrix algorithm and the second subset is generated by a finite-difference time-domain method.

[0007] According to one embodiment of this application, the type of grating structure of the sample laser includes at least one of a uniform grating structure, an apodized grating structure, and a phase-shifting grating structure.

[0008] According to one embodiment of this application, the first model is constructed based on a cross-modal network framework, and the second model is constructed based on a contrastive language image pre-trained model.

[0009] According to one embodiment of this application, the set grating parameters and the predicted grating parameters include at least one of coupling coefficient, reflectivity, refractive index, period, length, and duty cycle.

[0010] According to one embodiment of this application, the predicted output characteristic curve and the set output characteristic curve include at least one of a power curve and a slope efficiency curve.

[0011] According to one embodiment of this application, the first model is further used to simulate and fit the simulated grating structure corresponding to the first laser to be designed based on the set grating parameters.

[0012] Secondly, this application provides a laser design method for the laser design system described in the first aspect above, comprising: Based on the set grating parameters, the first model predicts and outputs the predicted output characteristic curve of the first laser to be designed; And / or predict and output the predicted grating parameters of the grating structure of the second laser to be designed based on the set output characteristic curve using a second model; The first model and the second model share the same training dataset. The input training data of the first model in the training dataset corresponds to the output training data of the second model, and the output training data of the first model in the training dataset corresponds to the input training data of the second model. The training dataset includes pairing data between the grating parameters of the grating structure of the sample laser and the output characteristic curve of the sample laser.

[0013] According to the laser design method of this application, by constructing two inverse mapping models, a first model and a second model, and jointly training them based on the same training dataset, a two-way closed-loop optimization from structural parameters to performance prediction and from target performance to structural reverse design can be achieved. The first model can quickly predict the output characteristic curve corresponding to a given grating parameter, accelerating performance evaluation. The second model can deduce the optimal grating parameter based on the expected output characteristic curve, realizing the reverse design of high-performance devices, which can significantly improve the design efficiency and accuracy of lasers.

[0014] Thirdly, this application provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the laser design method as described in the second aspect above.

[0015] Fourthly, this application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the laser design method as described in the second aspect above.

[0016] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the laser design method as described in the second aspect above.

[0017] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0018] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a schematic diagram of the laser design system provided in the embodiments of this application; Figure 2 This is a schematic diagram of the structure of the laser provided in the embodiments of this application; Figure 3 This is one of the schematic diagrams of the grating reflection spectrum provided in the embodiments of this application; Figure 4 This is a second schematic diagram of the grating reflection spectrum provided in the embodiments of this application; Figure 5 This is the third schematic diagram of the grating reflection spectrum provided in the embodiments of this application; Figure 6 This is the fourth schematic diagram of the grating reflection spectrum provided in the embodiments of this application; Figure 7 This is the fifth schematic diagram of the grating reflection spectrum provided in the embodiments of this application; Figure 8 This is the sixth schematic diagram of the grating reflection spectrum provided in the embodiments of this application; Figure 9 This is the seventh schematic diagram of the grating reflection spectrum provided in the embodiments of this application; Figure 10 This is a schematic diagram illustrating the training and operation of the first model provided in the embodiments of this application; Figure 11 This is a schematic diagram of the laser design system provided in this application performing laser design; Figure 12 This is a flowchart illustrating the laser design method provided in an embodiment of this application; Figure 13 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application.

[0019] Figure Labels Laser design system 100, first model 110, second model 120. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0021] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0022] The laser design system 100, laser design method, electronic device, and readable storage medium provided in this application will be described in detail below with reference to the accompanying drawings and through specific embodiments and application scenarios.

[0023] This application provides a laser design system 100.

[0024] like Figure 1 As shown, the laser design system 100 includes a first model 110 and a second model 120.

[0025] The input parameters of the first model 110 include the set grating parameters of the grating structure of the first laser to be designed. The first model 110 is used to predict and output the predicted output characteristic curve of the first laser to be designed based on the set grating parameters.

[0026] The second model 120 has input parameters including a set output characteristic curve of the second laser to be designed. The second model 120 is used to predict and output the predicted grating parameters of the grating structure of the second laser to be designed based on the set output characteristic curve.

[0027] In this embodiment, the first laser to be designed is a laser whose performance needs to be predicted in the forward direction based on given grating structure parameters. The grating parameters are the geometric and physical characteristics of the grating set by humans, such as period and duty cycle. The predicted output characteristic curve is the optical performance response of the first laser to be designed estimated and predicted by the first model 110 based on the set grating parameters.

[0028] The second laser to be designed is a laser whose grating structure needs to be designed in reverse according to the desired performance. The output characteristic curve is set as the curve corresponding to the optical performance of the laser, and the predicted grating parameters are the optimal grating structure parameters derived by the second model 120 based on the set output characteristic curve.

[0029] In this embodiment, the first model 110 is a forward design model that can simulate the physical behavior of the laser. It receives the set grating parameters as input, and then predicts the output characteristics of the laser using the set grating parameters by calculating or learning the rules, and outputs the predicted output characteristic curve, that is, it realizes the prediction of the laser performance from the grating structure of the laser.

[0030] The second model 120 is a reverse design model used for optimization and design. The second model 120 receives the user's desired set output characteristic curve as input, and then reverse calculates to predict what grating parameters the laser's grating structure should have to achieve this performance, and outputs the predicted grating parameters, that is, realizing the reverse deduction of the laser's grating structure from the laser's performance.

[0031] The laser design system 100 can predict the performance of the laser based on the designed grating structure, and can also deduce the required grating parameters based on the desired laser performance, thus forming a complete and intelligent laser design closed loop.

[0032] In this embodiment, the first model 110 and the second model 120 are used together to form an efficient design loop. First, the second model 120 is used to inversely derive the predicted grating parameters based on the set output characteristic curve. Then, the predicted grating parameters are used as the set grating parameters and input into the first model 110 for positive verification to obtain the predicted output characteristic curve. It is then determined whether the predicted output characteristic curve can meet the set output characteristic curve.

[0033] If the positive verification results do not meet the requirements, the output of the first model 110 can be used as feedback to adjust the input constraints or optimization objectives of the second model 120, and multiple iterations can be performed until satisfactory grating parameters are obtained.

[0034] In this embodiment, the first model 110 and the second model 120 share the same training dataset. The input training data of the first model 110 in the training dataset corresponds to the output training data of the second model 120, and the output training data of the first model 110 in the training dataset corresponds to the input training data of the second model 120. The training dataset includes pairing data between the grating parameters of the grating structure of the sample laser and the output characteristic curve of the sample laser.

[0035] The sample laser is an actual or simulated laser with known grating parameters and corresponding output characteristic curves used to collect training data.

[0036] The first model 110 and the second model 120 use the same training dataset, which includes paired data of grating parameters of multiple sample lasers and their corresponding output characteristic curves. During training, the first model 110 learns with grating parameters as input and output characteristic curves as output, while the second model 120 learns with output characteristic curves as input and grating parameters as output.

[0037] The laser design system 100 provided in this application embodiment can achieve bidirectional closed-loop optimization from structural parameters to performance prediction and from target performance to structural reverse design by constructing two mutually inverse mapping first models 110 and second models 120 and jointly training them based on the same training dataset. The first model 110 can quickly predict the output characteristic curve corresponding to a given grating parameter, accelerating performance evaluation. The second model 120 can deduce the optimal grating parameter based on the expected output characteristic curve, realizing the reverse design of high-performance devices, which can significantly improve the design efficiency and accuracy of lasers. The first model 110 and the second model 120 share training data and complement each other, which can enhance the model's generalization ability, reduce computational costs, and effectively solve the optimization problem caused by multi-parameter nonlinear coupling in laser design.

[0038] According to the laser design system 100 provided in the embodiments of this application, by constructing two inverse mappings, a first model 110 and a second model 120, and jointly training them based on the same training dataset, a two-way closed-loop optimization from structural parameters to performance prediction and from target performance to structural reverse design can be achieved. The first model 110 can quickly predict the output characteristic curve corresponding to a given grating parameter, accelerating performance evaluation. The second model 120 can deduce the optimal grating parameter based on the expected output characteristic curve, realizing the reverse design of high-performance devices, which can significantly improve the design efficiency and accuracy of lasers.

[0039] In some embodiments, the training dataset includes at least one of a first subset and a second subset, wherein the first subset is generated by a transfer matrix algorithm and the second subset is generated by a finite-difference time-domain method.

[0040] In this embodiment, the training dataset used to train the two models in the laser design system 100 may include simulation data from one or two sources: the first subset of the dataset is generated by simulation using the transfer matrix method (TMM), and the second subset of the dataset is generated by simulation using the finite-difference time-domain method. These two datasets can provide paired data between grating parameters and output characteristic curves from different levels of accuracy and efficiency, thereby enhancing the generalization ability and accuracy of the model.

[0041] The following describes a specific implementation of generating the training dataset.

[0042] The grating in the laser is a Bragg grating. Mode coupling in a Bragg grating is essentially an energy conversion between modes, and simulation modeling is mainly based on coupled-mode theory. The basic principle of the transfer matrix method is as follows: In a non-uniform grating, the amplitude, phase, and periodicity factors of the grating are no longer constants, but vary with the transmission distance z. The grating can be regarded as a cascaded structure of multiple uniform grating segments. Assuming the entire non-uniform grating is divided into M uniform segments, the incident and emitted light fields after passing through the k-th grating segment are respectively... and Both are connected via a transmission matrix. In relation to each other, it can be represented as:

[0043]

[0044] in, Let k represent the length of the k-th grating segment. For the entire non-uniform grating segment, the transmission form can be expressed as:

[0045] If there is a phase shift between the two gratings Then in and Multiply by a phase shift matrix between them.

[0046]

[0047] Using the transmission matrix method described above, the reflection spectrum of any single waveguide grating structure can be calculated. The Bragg wavelength can be expressed as:

[0048] in, Let Λ be the effective refractive index of the Bragg grating, Λ be the grating period, and the bandwidth of the grating reflection spectrum can be expressed as:

[0049] Where L is the length of the Bragg grating, and κ represents the coupling coefficient of the Bragg grating, the magnitude of which depends on the refractive index modulation amount Δn, and can be expressed by the following formula:

[0050] Where η is the overlap factor between the forward and backward optical field propagation modes and the grating, typically taking a value between 0 and 1; for a uniform Bragg grating, η can be 1. The reflectivity of the Bragg grating can be expressed as:

[0051] The coupling coefficient κ and the grating length L determine the reflectivity of the Bragg grating. Increasing κ and L can improve the reflectivity, thereby enhancing the reflection of the Bragg grating. Therefore, the coupling coefficient and grating length can be calculated by inversely from the grating output reflectivity.

[0052] The laser's grating parameters (coupling coefficient, reflectivity, etc.) are calculated using the transfer matrix method, and then the laser's power, slant efficiency, and other parameters are solved based on these calculated grating parameters. A modified transfer matrix method allows for precise control of the grating's coupling coefficient.

[0053] Two sets of simulation data were generated for the same grating structure using the Finite-Difference Time-Domain (FDTD) algorithm and the Time-Matching Model (TMM) algorithm, respectively. In the simulation using the FDTD algorithm, its sweep function was used to scan specific structural parameters, efficiently obtaining the second subset of data under different grating configurations. The TMM algorithm part was automated through scripts to complete the calculations and data storage, obtaining the first subset of data.

[0054] In some embodiments, the grating structure of the sample laser includes at least one of a uniform grating structure, an apodized grating structure, and a phase-shifting grating structure.

[0055] Among them, the uniform grating structure is a regular structure in which the grating period and etching depth are the same everywhere; the apodized grating structure is a structure in which the coupling coefficient of the grating gradually changes along the length direction according to a specific function (such as Gaussian, Hanning window, etc.) to suppress end face reflection and side mode; and the phase shift grating structure is a structure in which a phase shift is introduced in the middle or a specific position of the grating to break mode degeneracy and realize single-mode output.

[0056] In this embodiment, the grating structure of the sample laser includes at least one of a uniform grating structure, an apodized grating structure, and a phase-shifting grating structure, which can cover laser performance data with different design features, thereby improving the diversity and representativeness of the training data, and enabling the model to more accurately learn and predict the mapping relationship between grating parameters and output characteristic curves under various grating structures.

[0057] In some embodiments, the first model 110 is constructed based on a cross-modal network framework, and the second model 120 is constructed based on a contrastive language image pre-trained model.

[0058] Among them, the Any-to-Any cross-modal network framework is a deep learning architecture that can process and associate data from different modalities (such as images, text, audio, etc.), aiming to achieve semantic alignment and interaction between modalities; the Contrastive Language–Image Pre-training (CLIP) model is a model that is pre-trained on large-scale image-text pairs through contrastive learning, so that the corresponding content of text and images are close in a common semantic space, thereby achieving cross-modal understanding and zero-shot classification.

[0059] In this embodiment, based on a cross-modal network framework, grating parameters are used as text input, and the output characteristic curve of the laser is used as the output image. A stable diffusion (SD) model is used to stably train the dataset. The specific process is as follows: CLIP Text Encoder is used as a pre-module of the SD model to encode the input text information into corresponding text embeddings feature matrices. These feature matrices are then used to guide and control the generated content during the subsequent image generation process. After the text encoding is completed, the feature matrix is ​​optimized by an image optimization module consisting of a U-net and a scheduling algorithm. At the same time, the input grating temporal finite difference simulation image is encoded to generate latent space features, which are then input into the optimization module.

[0060] After the optimized latent space features are decoded by the image decoder, the target image is reconstructed, which is the output characteristic curve of the laser under specific grating parameter input. After training, the text-generated image model can realize the automated simulation and fitting of the grating structure and directly calculate the corresponding laser output characteristics. Since the model is trained with the TMM simulation results as the true value, its output grating reflection spectrum curve and laser output characteristic image have high accuracy. At the same time, the generative laser forward modeling can significantly reduce the computation time.

[0061] In this embodiment, the core idea of ​​CLIP is to create a shared, high-dimensional embedding space where semantically related images and text are mapped to positions very close to each other. The CLIP model consists of two parts: an image encoder and a text encoder. The image encoder is typically a vision transformer tasked with converting an input image into a high-dimensional vector representation, while the text encoder is a transformer tasked with converting the input text into a high-dimensional vector representation. Training CLIP is a contrastive learning process. During training, the model learns how to adjust the encoder to make the embedding vector distance of positive sample pairs as close as possible, and the embedding vector distance of negative sample pairs as far as possible. In this way, the CLIP model learns that specific grating parameter text is semantically matched with the image of the laser's output characteristic curve. The grating parameter text is paired one-to-one with the laser's PIV curve to train the CLIP model. Finally, an inverse design model for extracting grating parameters from the laser's PIV curve can be implemented.

[0062] In this embodiment, by combining a cross-modal network framework and a contrastive language image pre-trained model, image-text semantic alignment and deep interaction are achieved, which can improve the model's semantic understanding and generalization ability in tasks such as cross-modal retrieval.

[0063] In some embodiments, setting grating parameters and predicting grating parameters include at least one of coupling coefficient, reflectivity, refractive index, period, length, and duty cycle.

[0064] In this embodiment, the grating parameters include at least one of coupling coefficient, reflectivity, refractive index, period, length and duty cycle, which can accurately model the optical properties of the grating and predict its behavior, thereby supporting the optimization and simulation of the grating structure of the laser.

[0065] In some embodiments, the predicted output characteristic curve and the set output characteristic curve include at least one of a power curve and a slope efficiency curve.

[0066] In this embodiment, by limiting the predicted output characteristic curve and the set output characteristic curve to include at least one of the power curve and the slope efficiency curve, the optical performance of the laser under different grating parameters can be quantitatively evaluated, and the output power distribution or energy conversion efficiency can be accurately predicted and optimized, thereby guiding the grating structure design of the laser, improving device performance, and ensuring the consistency between simulation and actual needs.

[0067] In some embodiments, the first model 110 is also used to simulate and fit the simulated grating structure corresponding to the first laser to be designed based on the set grating parameters.

[0068] Among them, the simulated grating structure is a digital model constructed based on the set grating parameters to simulate the optical behavior of the grating structure of a real laser.

[0069] In this embodiment, by simulating and fitting the simulated grating structure based on the set grating parameters using the first model 110, the optical characteristics of the grating structure and the performance of the laser can be predicted quickly and accurately. This enables the automated mapping of design parameters to the structural model, improves the efficiency and accuracy of grating design, and reduces the dependence on iterative physical experiments.

[0070] like Figure 2 The diagram shows a schematic of a DFB laser, where a uniform grating is located above the active region. Based on simulation software, the uniform grating is modeled and simulated, and the resulting reflectance spectrum curve is shown below. Figure 3 As shown, in the reflection spectrum, the center wavelength λ is 1550 nm. The coupling coefficient κ of the grating can be calculated from the reflectivity of the main peak and the set grating length L. Therefore, the main characterizing parameters of the grating's reflection spectrum are κ, L, and λ.

[0071] like Figures 4-9 As shown, the reflection spectra and their fitting curves of a uniform grating (grating period of 238 nm, Bragg wavelength of 1542.9 nm, grating length L of 70 μm, coupling coefficient κ of 42171) obtained by the finite-difference time-domain algorithm and the TMM algorithm are presented. Figures 4-6 The grating reflection spectrum is given when κL = 2.952. Figures 7-9 For the grating reflection spectrum at κL=4.9517, due to the TMM algorithm not considering the dispersion of the coupling coefficient and using a first-order approximation for the effective refractive index, there is a certain deviation between the results of the two methods, but it is within an acceptable range. At smaller κL values ​​(e.g., κL=2.952), Figure 6 When the time-domain finite-difference algorithm and the TMM simulation curve are basically coincident, the deviation between the two increases significantly when the coupling coefficient and the grating length are increased to κL=4.9517. Figure 9 ).

[0072] like Figure 10As shown, based on the Any-to-Any first model 110, during the training process, the grating reflection spectrum is converted into image encoding, and the image is used as the output data of the model. Grating parameters such as period, length, duty cycle, and refractive index modulation are converted into structural information encoding, and the structural information is used as the input data of the model to train the first model 110. During the operation of the first model 110, the grating parameters are input into the first model 110 to obtain the image output of the first model 110. The image is diffused to obtain the laser PIV curve and slope efficiency output image. In addition, structural information is output, and the corrected parameter text is decoded.

[0073] like Figure 11 The diagram shows the flowchart of the laser design system 100 executing laser design. For the simulation of semiconductor lasers, the active region, grating region, and PN region of the laser are first structurally modeled, and parameters (doping, cavity length, grating period, etc.) and simulation settings are configured. After calculation by the carrier transport module, the system outputs the carrier concentration distribution under different bias voltages. Combined with the carrier and photon interaction module, the gain response at different wavelengths is calculated. Finally, the direct data obtained from the model is post-processed to solve for the output characteristic curves of the laser, such as power and slope efficiency.

[0074] With the rapid development of big data and artificial intelligence (AI) technologies, the demand for energy efficiency and speed in information processing is increasing. Low-power, large-scale, multi-channel photonic integrated chips have become core devices in the optoelectronic information field. Photonic integration technology integrates multiple optical components (such as lasers, modulators, waveguides, and filters) onto a single chip. Semiconductor lasers, as the core device of photonic integrated chips, are widely used in lidar, sensors, and fiber optic communications due to their advantages of small size, long lifespan, and high integration. The manufacturing process of semiconductor lasers is highly complex, requiring precise control of multiple physical processes. Their design heavily relies on complete physical models and accurate numerical simulations. Typical devices, such as distributed feedback lasers (DFBs) and vertical-cavity surface-emitting lasers, integrate multiple functional regions, including active gain, grating mode selection, and PN doping. The structures are tightly coupled, and any change in local parameters will affect the overall performance. Therefore, it is necessary to coordinate global integration based on regional optimization to achieve optimal output characteristics.

[0075] Semiconductor lasers are the core light source of photonic integrated chips, complementing and evolving alongside them. DFB lasers, a commonly used semiconductor laser, integrate a Bragg grating within the laser itself, achieving highly stable single-mode output. Due to their superior monochromaticity and stable single-mode characteristics, DFB lasers have become a cornerstone light source in modern high-speed fiber optic communication systems and precision optical sensing. Although their manufacturing is complex and costly, they are an irreplaceable choice for applications requiring high performance.

[0076] As a key component of DFB lasers, the Bragg grating is a special grating structure whose core principles are mode coupling and the Bragg reflection condition (phase matching condition). The structure of the Bragg grating directly determines the characteristics of the laser's resonant cavity, which is the core of the laser and therefore has a decisive influence on key performance characteristics such as output power, slant efficiency, wavelength, and linewidth. If the grating's reflectivity is too high, although the photon density inside the cavity is high, the laser is difficult to output, leading to increased internal losses and a decrease in actual output power. If the reflectivity is too low, there will be insufficient feedback, and the cavity cannot establish sufficient oscillation, increasing the laser threshold and also reducing output power. Simultaneously, optimizing the output coupling ratio can maximize the laser's slant efficiency. Furthermore, the reflection bandwidth of the grating is inversely proportional to its length; the longer the grating, the narrower the reflection spectrum.

[0077] The laser design system 100 provided in this application is a generative system that automatically optimizes the grating structure of a DFB laser to achieve high-performance laser characteristic output. It utilizes the powerful reasoning capability of large model algorithms to build a set of forward modeling models for lasers of different grating types, including a one-to-one mapping of the grating structure to the laser's slope efficiency and PIV curve, thereby improving the efficiency and accuracy of DFB laser design.

[0078] In related technologies, single-mode yield and mode stability are key issues. Currently, standard λ / 4 phase-shifted DFB lasers theoretically have two modes with the same threshold gain on either side of the Bragg wavelength, i.e., degenerate modes. Although introducing a phase-shifting grating (such as a λ / 4 phase shifter) is crucial to solving this problem, in actual manufacturing, small errors in the phase shift position, material gain inhomogeneities, and end-face reflections can all disrupt perfect symmetry, potentially leading to mode competition or mode hopping, thus reducing single-mode yield. On the other hand, even if single-mode lasing is achieved, a sufficiently high gain difference between the primary mode and the side modes is required to obtain a high SMSR (typically >45dB). Optimizing the product of the grating coupling coefficient κ and the cavity length L is central to the design, but excessively large or small κL values ​​can introduce other problems.

[0079] The precise control and trade-off of the coupling coefficient: The coupling coefficient κ is extremely sensitive to the grating's etching depth, duty cycle (DC), and shape (sinusoidal, rectangular, trapezoidal, etc.). Any minute deviation during manufacturing can cause a significant discrepancy between the κ value and the design value, thus affecting the final performance. With a strong grating (larger κL), although high mode stability and SMSR are provided, it leads to increased threshold current (high optical loss) and decreased output power (light is confined within the cavity and difficult to output). With a weak grating (smaller κL), although the threshold is low and the output power is high, mode stability is poor, SMSR is low, it is more sensitive to end-face reflections, and its anti-interference capability is weak. Therefore, the design needs to find the optimal balance between high single-mode stability and low threshold / high power.

[0080] The laser design system 100 provided in this application mainly solves the technical problems of single-mode yield and mode stability, and the problem of precise control of coupling coefficient.

[0081] The laser design system 100 provided in this application embodiment can achieve precise control of the grating coupling coefficient, improving the stability of laser output modes and the yield of single-mode products. First, an end-to-end physical model of the output characteristics of semiconductor lasers with different grating types and structures is established, focusing on indicators such as the laser's output center wavelength spectrum and slope efficiency. Commonly used grating types and structures for DFB lasers are determined, and three gratings (uniform grating, phase-shifting grating, and apodized grating) are selected as input options for the model. The coupling coefficients corresponding to the three grating structures are calculated based on coupled-mode theory, and then the laser's power, slope efficiency, and other performance characteristics are calculated based on the grating's output reflectivity and coupling coefficient. Next, an Any-to-Any algorithm network framework is built, using different grating structure parameters as the text input framework, the grating's reflectivity and coupling coefficient as the output text, and the laser's slope efficiency and power curves as the output images. A large dataset is collected for training based on the laser grating simulation model, using the SD model as the training model for the text-to-text + image paradigm. The trained model enables automated forward modeling from grating structure to laser performance, improving the efficiency and accuracy of forward design while enhancing control over coupling coefficients. Finally, based on the CLIP model, textual information of grating structure parameters is extracted from the laser's output characteristic curves to complete the reverse design of the DFB laser. Corresponding grating structure parameters can be provided for different DFB laser performances, simplifying the complexity of laser design and making it suitable for the design and fabrication of specialized high-power lasers.

[0082] The laser design system 100 provided in this application mainly focuses on grating type and structure as design parameters, and laser output performance as optimization index.

[0083] The laser design system 100 provided in this application improves the computational efficiency of solving the laser output characteristics from the grating structure. Compared with the method of calculating the characteristics of DFB lasers through the transfer matrix algorithm and coupled-mode theory, the proposed generative laser design system 100 can achieve efficient prediction of laser structure performance. The model can directly generate the corresponding laser PIV curve through several key grating structure parameters, which greatly reduces the simulation calculation time while ensuring calculation accuracy.

[0084] Enhanced control over the coupling coefficient: The structural parameters of the grating, such as the grating period, refractive index modulation, and grating length, determine the grating coupling coefficient. By correcting the accuracy of the TMM simulation grating results through a multi-mode algorithm, precise control of the coupling coefficient can be achieved.

[0085] To improve the yield of single-mode products, the CLIP model is used as a reverse design method. The grating structure can be deduced from the expected laser power, thereby improving the yield of single-mode products in the experimental process.

[0086] This application also provides a laser design method based on the laser design system 100 described above.

[0087] The laser design method can be applied to the terminal, specifically executed by the hardware or software within the terminal.

[0088] The laser design method provided in this application embodiment can be executed by an electronic device or a functional module or entity in an electronic device that can implement the laser design method. The electronic devices mentioned in this application embodiment include, but are not limited to, mobile phones, tablets, computers, cameras, and wearable devices. The laser design method provided in this application embodiment will be described below using an electronic device as the execution subject.

[0089] The laser design methodology includes: Based on the set grating parameters, the first model 110 predicts and outputs the predicted output characteristic curve of the first laser to be designed.

[0090] And / or predict and output the predicted grating parameters of the grating structure of the second laser to be designed based on the set output characteristic curve through the second model 120; The first model 110 and the second model 120 share the same training dataset. The input training data of the first model 110 in the training dataset corresponds to the output training data of the second model 120, and the output training data of the first model 110 in the training dataset corresponds to the input training data of the second model 120. The training dataset includes paired data between the grating parameters of the grating structure of the sample laser and the output characteristic curve of the sample laser.

[0091] like Figure 12The image shows an embodiment of a laser design method, including: Step 1210: Based on the set grating parameters, the first model 110 predicts and outputs the predicted output characteristic curve of the first laser to be designed.

[0092] Step 1220: Based on the set output characteristic curve, the second model 120 predicts and outputs the predicted grating parameters of the grating structure of the second laser to be designed.

[0093] According to the laser design method provided in the embodiments of this application, by constructing two inverse mappings, a first model 110 and a second model 120, and jointly training them based on the same training dataset, a two-way closed-loop optimization from structural parameters to performance prediction and from target performance to structural reverse design can be achieved. The first model 110 can quickly predict the output characteristic curve corresponding to a given grating parameter, accelerating performance evaluation. The second model 120 can deduce the optimal grating parameter based on the expected output characteristic curve, realizing the reverse design of high-performance devices, which can significantly improve the design efficiency and accuracy of lasers.

[0094] The laser design method provided in this application can be executed by a laser design device. This application uses a laser design device executing the laser design method as an example to illustrate the laser design device provided in this application.

[0095] This application also provides a laser design apparatus.

[0096] The laser design device includes: The first processing module is used to predict and output the predicted output characteristic curve of the first laser to be designed based on the set grating parameters through the first model 110.

[0097] The second processing module is used to predict and output the predicted grating parameters of the grating structure of the second laser to be designed based on the set output characteristic curve through the second model 120. The first model 110 and the second model 120 share the same training dataset. The input training data of the first model 110 in the training dataset corresponds to the output training data of the second model 120, and the output training data of the first model 110 in the training dataset corresponds to the input training data of the second model 120. The training dataset includes paired data between the grating parameters of the grating structure of the sample laser and the output characteristic curve of the sample laser.

[0098] According to the laser design apparatus provided in the embodiments of this application, by constructing two inversely mapped first models 110 and second models 120, and jointly training them based on the same training dataset, a two-way closed-loop optimization from structural parameters to performance prediction and from target performance to structural reverse design can be achieved. The first model 110 can quickly predict the output characteristic curve corresponding to a given grating parameter, accelerating performance evaluation. The second model 120 can deduce the optimal grating parameter based on the expected output characteristic curve, realizing the reverse design of high-performance devices, which can significantly improve the design efficiency and accuracy of lasers.

[0099] The laser design device in this application embodiment can be an electronic device or a component within an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal or other devices besides a terminal. For example, the electronic device can be a mobile phone, tablet computer, laptop computer, PDA, in-vehicle electronic device, mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc. It can also be a server, network attached storage (NAS), personal computer (PC), television (TV), ATM, or self-service machine, etc. This application embodiment does not specifically limit the specific device.

[0100] The laser design device in this application embodiment can be a device with an operating system. This operating system can be a Microsoft (Windows) operating system, an Android operating system, an iOS operating system, or other possible operating systems; this application embodiment does not specifically limit the specific operating system.

[0101] The laser design apparatus provided in this application embodiment can achieve Figure 1 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0102] In some embodiments, such as Figure 13As shown, this application embodiment also provides an electronic device 1300, including a processor 1301, a memory 1302, and a computer program stored in the memory 1302 and executable on the processor 1301. When the program is executed by the processor 1301, it implements the various processes of the above-described laser design method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0103] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0104] This application also provides a non-transitory computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described laser design method embodiments and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0105] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0106] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the laser design method described above.

[0107] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0108] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described laser design method embodiments and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0109] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0110] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0111] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the related technology, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0112] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

[0113] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0114] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

Claims

1. A laser design system, characterized in that, include: The first model, whose input parameters include the set grating parameters of the grating structure of the first laser to be designed, is used to predict and output the predicted output characteristic curve of the first laser to be designed based on the set grating parameters. The second model, whose input parameters include a set output characteristic curve of the second laser to be designed, is used to predict and output the predicted grating parameters of the grating structure of the second laser to be designed based on the set output characteristic curve. The first model and the second model share the same training dataset. The input training data of the first model in the training dataset corresponds to the output training data of the second model, and the output training data of the first model in the training dataset corresponds to the input training data of the second model. The training dataset includes pairing data between the grating parameters of the grating structure of the sample laser and the output characteristic curve of the sample laser.

2. The laser design system according to claim 1, characterized in that, The training dataset includes at least one of a first subset and a second subset, wherein the first subset is generated by the transfer matrix algorithm and the second subset is generated by the finite-difference time-domain method.

3. The laser design system according to claim 1, characterized in that, The grating structure of the sample laser includes at least one of a uniform grating structure, an apodized grating structure, and a phase-shifting grating structure.

4. The laser design system according to claim 1, characterized in that, The first model is built on a cross-modal network framework, and the second model is built on a contrastive language image pre-trained model.

5. The laser design system according to any one of claims 1-4, characterized in that, The set grating parameters and the predicted grating parameters include at least one of the following: coupling coefficient, reflectivity, refractive index, period, length, and duty cycle.

6. The laser design system according to any one of claims 1-4, characterized in that, The predicted output characteristic curve and the set output characteristic curve include at least one of a power curve and a slope efficiency curve.

7. The laser design system according to any one of claims 1-4, characterized in that, The first model is also used to simulate and fit the simulated grating structure corresponding to the first laser to be designed based on the set grating parameters.

8. A laser design method based on the laser design system according to any one of claims 1-7, characterized in that, include: Based on the set grating parameters, the first model predicts and outputs the predicted output characteristic curve of the first laser to be designed; And / or predict and output the predicted grating parameters of the grating structure of the second laser to be designed based on the set output characteristic curve using a second model; The first model and the second model share the same training dataset. The input training data of the first model in the training dataset corresponds to the output training data of the second model, and the output training data of the first model in the training dataset corresponds to the input training data of the second model. The training dataset includes pairing data between the grating parameters of the grating structure of the sample laser and the output characteristic curve of the sample laser.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the laser design method as described in claim 8.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the laser design method as described in claim 8.

Images