An adaptive optical fiber communication rate intelligent detection method

Abstract

The application discloses a kind of self-adapting optical fiber communication rate intelligent detection methods, belong to optical fiber communication field, it is characterized by: transmitter and receiver use SFP optical module, through multimode optical fiber transmission, utilize lens to collect guide laser signal and divide into two ways, one is received and parsed by receiver and then connected to display, another is converted into electrical signal containing rate key information by photoelectric detector;The electrical signal is preprocessed, such as filtering, noise reduction, sampling, and adjusted to the format adapted to DenseNet121 model, after multiple steps such as convolution layer, batch normalization layer, the output predicted optical fiber communication rate is inputted into the model.Compared with the traditional method, the method can solve the problem of long time consumption, low precision and high misjudgment of traditional measurement method, improve the adaptability and reliability of the detection device in the optical fiber communication environment.

Description

Technical Field

[0001] This invention belongs to the field of optical fiber communication, and in particular relates to an adaptive intelligent detection method for optical fiber communication rate. Background Technology

[0002] In fiber optic communication systems, accurately detecting fiber optic communication rates is crucial for system optimization, maintenance, and fault diagnosis. Precisely measuring the actual transmission rate of the fiber optic cable verifies network performance indicators and ensures the stable operation of the entire network system under design standards. In complex network environments, network bottlenecks and faults frequently occur. When network issues such as lag and latency arise, measuring the fiber optic communication rate is a powerful tool for network maintenance, optimization, and troubleshooting.

[0003] In complex network environments, existing diagnostic methods are time-consuming in locating network bottlenecks and faults. Furthermore, traditional measurement methods may lack accuracy, and misjudgments of network performance due to measurement errors can affect overall network deployment and optimization. These shortcomings severely restrict the effective operation and development of networks. Therefore, a new method is urgently needed to improve the situation. A novel fiber optic communication rate detection method has emerged, capable of rapidly detecting anomalies in fiber optic communication rates. It is expected to overcome the deficiencies of traditional methods and provide strong support for network performance improvement and optimization.

[0004] Deep learning algorithms are used in the data processing stage. By training a convolutional neural network (CNN) model, the electrical signals are preprocessed and features are extracted, ultimately outputting the fiber optic communication rate. Summary of the Invention

[0005] The purpose of this invention is to provide an adaptive intelligent detection method for optical fiber communication rate, which solves the problems of long time consumption, low accuracy and high misjudgment in traditional optical fiber communication rate measurement methods, thereby improving the adaptability and reliability of detection devices in optical fiber communication environments.

[0006] To address the aforementioned technical problems, this invention employs the following technical solution: an adaptive intelligent detection method for optical fiber communication rates, wherein the laser uses an SFP optical module as the transmitter of the entire system. A lens is selected to collect the laser signal, and a receiver is used to receive and analyze the optical signal. Data preprocessing involves filtering, noise reduction, and sampling of the electrical signal passing through the photodetector to improve data quality, remove noise interference, unify feature scales, and enhance data diversity. A CNN neural network model is designed and trained: the rate is estimated through regression output to improve the performance and intelligence level of the optical communication system.

[0007] Further, training data is acquired. This invention utilizes an SFP optical module as the laser output, with multimode optical fiber as the transmission medium. Multimode optical fiber has multiple transmission modes, capable of carrying a large capacity of optical signals. When the laser propagates in the multimode optical fiber, the principle of total internal reflection ensures stable signal transmission. During propagation, the material and structural characteristics of the optical fiber minimize signal attenuation, ensuring the integrity of the optical signal. A lens is used to collect and guide the laser signal; this lens has high transmittance and precise optical focusing performance, effectively ensuring the stability and directionality of laser emission. The propagated optical signal is split into two paths: one path reaches the receiver, which receives and analyzes the optical signal, converting it into a recognizable electrical signal, and then connects to the display. The display visually presents the transmission status of the optical signal in the optical fiber. By observing parameters such as signal strength and waveform on the display, the smoothness of the optical fiber transmission can be verified, and potential signal interruptions or abnormal attenuation during transmission can be detected promptly. The other path connects to a photodetector. Photodetectors are based on the photoelectric effect and can efficiently convert optical signals into electrical signals. The converted electrical signals contain key information such as the speed in optical fiber communication.

[0008] Furthermore, batch preprocessing of the data is performed. Python is used to filter, denoise, and sample the raw electrical signals converted by the photodetector, enhancing data diversity. Since the DenseNet121 model was originally designed for image data, the converted electrical signals need to be adjusted to a format suitable for the DenseNet121 model's input. The electrical signals are organized into a two-dimensional or three-dimensional data matrix according to the time series, similar to the pixel matrix structure of an image, ensuring that its dimensions match the model's expected input dimensions and guaranteeing that the model can correctly process the electrical signal data. Attached Figure Description

[0009] Figure 1 This is an experimental system diagram of an adaptive optical fiber communication rate intelligent detection method provided in an embodiment of the present invention;

[0010] Figure 2 This is a flowchart of an intelligent detection method for adaptive optical fiber communication rate provided in an embodiment of the present invention;

[0011] Attached labels: 1 for lens; 2 for transmitter; 3 for optical fiber; 4 for receiver; 5 for display; 6 for photodetector; 7 for deep learning model. Detailed Implementation

[0012] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The steps in the following embodiments are merely for illustrative purposes, and the order of the steps is not limited. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0013] In the actual training process of this invention, various SFP optical modules with different fiber optic transmission rates were selected, namely 1.25Gbps, 2.125Gbps, 4.25Gbps, and 8Gbps. By using these optical modules with different rates for training, the model can better learn the characteristics of optical signals at different rates, thereby improving the accuracy and adaptability of fiber optic communication rate prediction.

[0014] See Figure 1 This invention provides an experimental system diagram of an intelligent detection method for adaptive optical fiber communication rates, including a lens 1, a transmitter 2, an optical fiber 3, a receiver 4, a display 5, a photodetector 6, and a deep learning model 7. The specific workflow is as follows: Laser signals from SFP optical modules with different optical fiber transmission rates (1.25Gbps, 2.125Gbps, 4.25Gbps, and 8Gbps) are used as transmitters and transmitted stably via multimode optical fiber using the principle of total internal reflection. A lens with high transmittance and precise focusing performance collects and guides the laser signal, splitting it into two paths. One path is received and analyzed by the receiver and connected to the display to visually present the optical signal transmission status. The other path is a photodetector that converts the optical signal into an electrical signal containing key rate information based on the photoelectric effect. Subsequently, the original electrical signals are preprocessed in batches using Python, including filtering, noise reduction, and sampling. Finally, the processed data is input into a specified CNN model. This model undergoes batch normalization of the input layer, feature extraction of the convolutional layer, dimensionality reduction of the pooling layer, and integration and classification prediction of the fully connected layer. Finally, the output layer linearly activates and outputs the optical fiber communication rate.

[0015] Specifically, SFP optical modules with different transmission rates (1.25Gbps, 2.125Gbps, 4.25Gbps, and 8Gbps) act as transmitters in the system, emitting infrared light. This infrared light is transmitted through multimode optical fiber, which, due to its multiple transmission modes, can carry large-capacity optical signals. During transmission, the principle of total internal reflection maintains stable signal propagation, while its material and structural characteristics minimize signal attenuation. A collecting lens with high transmittance and precise optical focusing performance collects and guides the laser signal. The propagated optical signal, guided by the lens, is then split into two paths. One path is received and analyzed by a receiver, which converts it into an electrical signal and connects it to a display. Through the display, operators can visually observe the transmission status of the optical signal in the fiber, thereby verifying the smoothness of fiber transmission and promptly detecting signal interruptions, abnormal attenuation, and other issues. The other path is transmitted to a photodetector, which, based on the photoelectric effect, efficiently converts the optical signal into an electrical signal containing key information such as the fiber optic communication rate.

[0016] See Figure 2 This invention provides a flowchart of an adaptive intelligent detection method for optical fiber communication rates. First, electrical signals are acquired. Then, the acquired images are preprocessed, mainly including filtering, noise reduction, and sampling operations on the electrical signals passing through the photodetector, thereby improving data quality. The model uses a DenseNet121 training model. Before model training, the electrical signals are adjusted to a format suitable for the DenseNet121 model input, enabling the model to directly process the electrical signal data. Then, the model is trained using a training set to accurately predict the optical fiber communication rate based on the new input electrical signals, thereby achieving performance evaluation of the optical fiber communication system.

[0017] Specifically, the electrical signal converted by the photodetector is adjusted to a format suitable for the input of the DenseNet121 model, so that its dimensions match the model's expected input dimensions (height, width, channels). The processed electrical signal can be input into the DenseNet121 model. First, it passes through convolutional layers, batch normalization layers, ReLU activation function layers, and max pooling layers for preliminary feature extraction and filtering. Then, it enters multiple Dense Blocks and Transition Layers to continuously mine the complex features in the electrical signal. Next, it enters a global average pooling layer to obtain a one-dimensional feature vector containing global feature information. Then, it enters a fully connected layer to map the one-dimensional feature vector to a new dimensional space. Finally, after simple post-processing operations, a scalar is output, namely the predicted fiber optic communication rate.

[0018] Furthermore, in the DenseNet121 model used in this system, the convolutional layers use 7×7 kernels with a stride of 2 for convolution operations, generating 64 feature maps to extract preliminary features of the electrical signal. The batch normalization layer normalizes the feature maps output by the convolutional layers, accelerating model convergence. The ReLU activation function introduces nonlinear factors, increasing the model's expressive power. The max pooling layer uses a 3×3 window with a stride of 2 for downsampling, reducing the feature map size while retaining significant features. The Dense Block consists of multiple convolutional layers, employing dense connections to promote feature reuse and alleviate the gradient vanishing problem. The Transition Layer contains 1×1 convolutional layers and 2×2 average pooling layers to reduce the number of channels and shrink the feature map size. The global average pooling layer performs average pooling on the feature map output by the last Dense Block, obtaining a one-dimensional feature vector containing global feature information. The fully connected layer converts this one-dimensional feature vector into a single value through weight and bias operations, and after post-processing, the final predicted fiber optic communication rate is obtained.

[0019] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. An adaptive intelligent detection method for fiber optic communication rate belongs to the field of fiber optic communication, and its characteristics are as follows: The lens (1), transmitter (2), optical fiber (3), receiver (4), display (5), photodetector (6), and deep learning model (7) are described below. The specific steps of the adaptive optical fiber communication rate intelligent detection method are as follows: 1) The laser uses an SFP optical module with an optical fiber transmission rate of 1.25Gbps, 2.125Gbps, 4.25Gbps, or 8Gbps to emit laser light. 2) Laser transmission is achieved using multimode fiber. Multimode fiber transmits signals stably due to the principle of total internal reflection, and its material and structural characteristics reduce signal attenuation. 3) Use a lens with high light transmittance and precise optical focusing performance to collect and guide the laser signal, so that the optical signal is split into two paths: one path is received and analyzed by the receiver and connected to the display to observe the optical signal transmission status and verify the smoothness of optical fiber transmission; the other path is a photodetector that converts the optical signal into an electrical signal containing key information about the optical fiber communication rate based on the photoelectric effect. 4) The raw electrical signal converted by the photodetector is filtered, denoised, and sampled using Python. The processed electrical signal is then adjusted to a format suitable for the input of the DenseNet121 model. After processing by the DenseNet121 model, the predicted fiber optic communication rate is output.

2. The adaptive optical fiber communication rate intelligent detection method according to claim 1, characterized in that, The lens (1) has high light transmittance and precise optical focusing performance, which can effectively ensure the stability and directionality of laser emission.

3. The adaptive optical fiber communication rate intelligent detection method according to claim 1, characterized in that, The photodetector (6) has high sensitivity and fast response characteristics, can accurately capture optical signals and quickly convert them into electrical signals, and has strong resistance to environmental interference, ensuring the accuracy and reliability of the converted electrical signals, thus providing a guarantee for the accurate detection of fiber optic communication rates.

4. The adaptive optical fiber communication rate intelligent detection method according to claim 1, characterized in that, In the DenseNet121 model, the convolutional layers use 7×7 kernels with a stride of 2 to perform convolution operations, and 64 kernels generate 64 feature maps to extract preliminary features of the electrical signal. Batch normalization layers accelerate convergence, ReLU activation functions increase nonlinearity, and max pooling layers downsample to preserve significant features. Dense blocks employ dense connections to promote feature reuse, while transition layers reduce the number of channels and shrink feature map size. Global average pooling yields a one-dimensional feature vector, and fully connected layers output the predicted fiber communication rate.

Images